GENDER, CULTURE CHANGE, AND FERTILITY DECLINE IN HONDURAS: AN

Document technical information

Format pdf
Size 822.5 kB
First found Jun 9, 2017

Document content analysis

Language
English
Type
not defined
Concepts
no text concepts found

Persons

Carl Friedrich Gauss
Carl Friedrich Gauss

wikipedia, lookup

Franz Joseph I of Austria
Franz Joseph I of Austria

wikipedia, lookup

Brad Cox
Brad Cox

wikipedia, lookup

Elisabeth of Valois
Elisabeth of Valois

wikipedia, lookup

David Strassman
David Strassman

wikipedia, lookup

David Kertzer
David Kertzer

wikipedia, lookup

Hank Green
Hank Green

wikipedia, lookup

Organizations

Places

Transcript

GENDER, CULTURE CHANGE, AND FERTILITY DECLINE IN HONDURAS: AN
INVESTIGATION IN ANTHROPOLOGICAL DEMOGRAPHY
By
DAVID PATRICK KENNEDY
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
2002
Copyright 2002
by
David P. Kennedy
This work is dedicated to my wife and my parents. I could not have completed this
project without their many sacrifices and years of unconditional support.
ACKNOWLEDGMENTS
A project like this goes beyond the effort of one person. There are many people
who have helped me, in one way or another, over the years. It is my pleasure to thank all
those who have contributed to the successful completion of this dissertation.
First, I would like to thank Martin F. Murphy. When I first started thinking about
going to graduate school for anthropology, he made sure I knew what a bad idea it was.
Once that sunk in, he then told me why it was also a good idea. His straightforward
assessment of academic life is something I have always remembered. It allowed me to
know that I was getting into it for the right reasons.
Next, I would like to thank my advisor, H. Russell Bernard. He has always shown
honesty, respect and genuine concern for his students. My appreciation of his sense of
humor is matched only by my appreciation for his editing skills. I feel fortunate having
received years of tutelage from someone with such a passion for the scientific study of
humanity. The enthusiasm I have for testing ideas with empirical evidence is in large
part due to his inspiration.
The other members of my dissertation committee have also been instrumental in
the completion of this project. Maxine Margolis, Jim Stansbury, Larry Severy, and
Chuck Wood have given me encouragement and support during various stages of my
graduate training.
I would also like to mention the encouragement and inspiration I received from
Marvin Harris. I remember the exact moment I realized I would become an
iv
anthropologist, while I was reading his book, Our Kind. I was struck by his ability to
take explanations for human behavior to their ultimate causes. His writings are some of
the most intellectually satisfying works I have ever read. I use his forcefulness and
clarity of argument as a model for my writing. I am fortunate to have participated in his
last seminar in 1997 during my first year of graduate training. His encouragement and
interest in my work were major milestones in my development as an independent scholar.
My graduate training would not have been complete without the encouragement,
inspiration, debates, and good times I had with my fellow University of Florida graduate
students. I would like to thank my fellow SORs, Nanette Barkey, Lance Gravlee, Hank
Green, Ken Sturrock, and Elli Sugita, for helping me through the rich tapestry that is
proposal writing. I would also like to thank the second generation SORs, Stacey Giroux,
Mark House, Rosalyn Negron, Fatma Soud, and Amber Yoder Wutich, for allowing us to
pass along our passion for data sucking. The weekly SOR meetings helped me in my
professional development while allowing me to sort through many data collection, data
analysis, and theoretical issues.
Many professionals in the field helped me in one way or another during my
graduate training. I would like to thank Gene Hammel, Penn Handwerker, Stephen Perz,
Warren Miller and Gery Ryan for making important contributions towards my
professional development. I would like to thank Jeff Johnson, Susan Weller, Stephen
Borgatti, and all those who participated in the 1999 National Science Foundation
Summer Institute for Research Design in Cultural Anthropology. Participating in the socalled Methods Camp was a key moment in my scientific training. I would also like to
thank David Kertzer, Monica Das Gupta, and all of the participants of the session on
v
anthropological demography at the 2001 annual meeting of the Population Association of
America.
I would not have been able to complete my training if not for the assistance of the
Bureau of Economic and Business Research. I would like to thank Stan Smith and Clint
Collins for their support of my employment at BEBR. I would especially like to thank
Chris McCarty, who provided me with more than just a steady paycheck. In addition to
direct assistance and mentorship during various stages of my graduate training, he
allowed me to develop my skills as a researcher in a stimulating and flexible
environment.
Completion of a dissertation in anthropology typically depends on the assistance of
many people during fieldwork. My project is no exception. I would not have been able
to complete my graduate training without the assistance of Jesús Salinas Pedraza, Josefa
Leonarda Gonzalez Ventura, CELIAC, the people of Ampliación Dolores, Martha W.
Rees, Arthur D. Murphy, and the Welte Institute while I conducted fieldwork in Oaxaca.
Also, while I conducted research in Honduras, many people helped make my data
collection efforts as successful as possible. I especially would like to thank Amanda
Madrid for her facilitation of my research and her friendship while my wife and I grew
accustomed to living in a strange environment. I would also like to thank Amanda’s
family, Doris Clark, Celia Lett, and all the other members of PREDISAN for their
generous support and friendship. I would like to thank Marlene Ruiz Mejia and her
family for their friendship and neighborly support. I would also like to thank Jose
Manuel García for many stimulating conversations about life in Catacamas.
vi
I was able to collect high quality data mainly because of the efforts of my research
assistants. Rudy Ordoñez, Rina Maradiaga, and Carmen Rojas inspired me with their
professionalism, hard work, and dedication towards making my project successful. We
are also grateful for the hospitality and friendship they and their families showed us
during our stay in Catacamas.
My dissertation fieldwork was made possible through a grant from the Andrew W.
Mellon Foundation through the RAND Small Grants Program for Research on Central
America. I would like to thank the Andrew W. Mellon Foundation, RAND, and Anne
Pebley for funding dissertation fieldwork in Central America. I would also like to thank
them for their support of innovative research techniques, especially the combination of
qualitative and quantitative methods.
I also received emotional support from many sources during my graduate training.
I would like to thank my parents, my brothers and sisters, my nephews and nieces and my
in-law family. Everyone has always been supportive of my decision to pursue my
doctorate. Although pursuing a Ph.D. in anthropology is not a common career choice, I
never heard one word of discouragement from anyone in my family. For that, I am
particularly grateful. Also, I would like to thank my parents for letting me know that I
always had their financial support if necessary. That peace of mind removed much of the
uncertainty of completing many additional years of schooling.
Most importantly, I would like to thank my partner, Marie. Through the years she
has braved a lack of teaching jobs, grade school photography, finals-taking law students,
after school bus rides, steamers, malaria medication, solitaire, re-booting houses, chunky
style water/air, hot boiled cutlery, encroaching termites, ladrones, public health sprayers,
vii
crazed taxi drivers, the itch, University of Florida parking, state employee work ethic,
north central Florida fine arts support, multiple moving days, much less attention from
me than she deserved, and many, many other horrors that someone not receiving a degree
should not have to endure. It is due to her love and support that I have been able to
accomplish this project.
Finally, I would like to thank Mavis, Mystere, Schmelvis, and Milka for reminding
me that, sometimes, naps, going outside, sock tug-a-war, and food are the most important
things in the world.
viii
TABLE OF CONTENTS
page
ACKNOWLEDGMENTS.................................................................................................. iv
LIST OF TABLES ...........................................................................................................xiii
LIST OF FIGURES........................................................................................................... xv
ABSTRACT ..................................................................................................................... xvi
CHAPTER
1 INTRODUCTION........................................................................................................... 1
The Essential Elements to an Interdisciplinary Theory of Fertility Decline................... 2
Human Reproductive Biology.................................................................................. 2
Culture and Human Reproduction............................................................................ 3
Gender and Human Reproduction............................................................................ 4
Falsification of Fertility Decline Theories ............................................................... 5
Anthropology and Theories of Fertility Decline ............................................................. 5
Organization of the Manuscript....................................................................................... 6
2 DEMOGRAPHIC TRANSITION THEORY AND ITS AFTERMATH ....................... 8
Introduction ..................................................................................................................... 8
Demographic Transition Theory ..................................................................................... 9
Criticism of Demographic Transition Theory............................................................... 10
Responses to the Failure of Demographic Transition Theory....................................... 14
Wealth Flows Theory: Caldwell............................................................................ 14
Modification of Wealth Flows Theory: Handwerker ............................................ 15
Evaluation of Wealth Flows Theories .................................................................... 18
Empirical Problems with Wealth Flows Theories.................................................. 20
Institutional Determinants of Fertility Decline....................................................... 22
Evaluation of the Institutional Approach to Fertility Decline ................................ 23
Critique of the Institutional Approach.................................................................... 24
Evolutionary Models of Fertility Decline............................................................... 29
3 EVOLUTIONARY THEORY, FERTILITY, AND HUMAN BEHAVIOR ............... 37
Introduction ................................................................................................................... 37
Two Models of Gene/Culture Interaction ..................................................................... 37
ix
Common Social Science Model ............................................................................. 39
Alternate Model...................................................................................................... 41
Model Discussion ................................................................................................... 43
The Evolutionary Mechanism of Fertility Decline ....................................................... 45
The Black Box of Evolutionary Theories of Human Fertility Decline .................. 45
EEA Variability Selection and Human Plasticity................................................... 47
EEA Selection for Reproductive Motivation ......................................................... 48
Pan-Human Reproductive Drives, the EEA and Culture ....................................... 49
The Analysis of Variance in Reproductive Behavior Caused by Genes and
Environment ............................................................................................... 51
Social Science Resistance to Evolutionary Explanations for Human Behavior ........... 52
The Use and Misuse of the Word Biology.................................................................... 59
Conclusion..................................................................................................................... 60
4 CULTURE AND REPRODUCTIVE BEHAVIOR...................................................... 62
Introduction ................................................................................................................... 62
Validity and Reliability in Demography and Anthropology......................................... 63
Disciplinary Differences between Anthropology and Demography ...................... 63
Disciplinary Congruence Between Anthropology and Demography..................... 65
The Uses of Culture in Demography and Anthropology .............................................. 68
Anthropological Critiques of Demographic Uses of Culture ................................. 68
Anthropological Conceptualizations of Culture..................................................... 71
Cognitive Theory of Culture .................................................................................. 72
Culture as Evolutionary Adaptation ....................................................................... 75
Economic and Cultural Explanations of Reproductive Behavior .......................... 78
A Method of Culture for Demography and Anthropology ........................................... 80
The Measurement of Culture: Consensus Analysis .............................................. 80
Qualitative Methods, Traditional Statistics, and Multivariate Statistical Methods 85
Debates of Qualitative and Quantitative Methods in Anthropology and
Demography ............................................................................................... 87
The Use of Quantification in Anthropological Demography................................. 90
Conclusion..................................................................................................................... 94
5 REPRODUCTIVE ATTITUDES AND BEHAVIOR IN HONDURAS...................... 95
Introduction ................................................................................................................... 95
Honduras as a Case Study in Theories of Fertility Decline .......................................... 96
History of Fertility Decline in Honduras and Central America ............................. 96
Causes of Fertility Rate Variation in Central America .......................................... 97
Unmet Need for Family Planning .......................................................................... 99
Social Interaction as a Cause of Fertility Decline ................................................ 101
Critique of Demographic Analyses of Fertility Decline ............................................. 102
Critique of Social Interaction Theory................................................................... 104
National Level Studies of Fertility Decline.......................................................... 106
Culture Theory and Fertility Decline ................................................................... 108
Gender and Fertility Decline ................................................................................ 109
x
Reproductive Attitudes of Honduran Men and Women.............................................. 112
Analysis of Dios Me Mande ................................................................................. 115
Re-analysis of National Data................................................................................ 117
Discussion ................................................................................................................... 121
6 ADAPTATION AND ADMINISTRATION OF THE CHILDBEARING
QUESTIONNAIRE .................................................................................................... 122
Introduction: Natality Culture .................................................................................... 122
The Childbearing Questionnaire ................................................................................. 123
The Childbearing Questionnaire and Evolutionary Theory ................................. 123
The Motivational Model Behind the CBQ ........................................................... 124
The Potential Cross-Cultural Application of the CBQ......................................... 127
Empirical Evidence Supporting the T-D-I-B-E Model ........................................ 128
Genes and Childbearing Motivation..................................................................... 130
Adaptation of the Childbearing Questionnaire in Central Honduras .......................... 132
Research Location ................................................................................................ 132
Instrument Translation ......................................................................................... 134
Ethnography and Scale Adaptation: Instrument Pre-Test ................................... 135
Determining Cultural Validity of Scale Items............................................... 135
Modification of the CBQ............................................................................... 138
Pre-Test Sample ............................................................................................ 142
Scale Administration ............................................................................................ 146
Building a Research Team ................................................................................... 146
Research Team Training ...................................................................................... 148
Random and Non-random Sampling.................................................................... 150
Quota Sample ....................................................................................................... 151
Combining Exploration with Questionnaire Administration ............................... 155
Scale Administration Phase Two: Random Sample............................................ 159
Urban Sampling Frame ........................................................................................ 160
Rural Sample ........................................................................................................ 162
Phase Two Questionnaire Modifications.............................................................. 165
7 ANALYSIS ................................................................................................................. 167
The Development of the Hypothesis of One Culture of Pronatality........................... 167
Construct Validity of Pronatality Culture ................................................................... 169
Elimination of Sex-Bias in the CBQ ........................................................................... 174
Description of Culture via Analysis of Factors and Ethnographic Interviewing–Phase
One ................................................................................................................ 177
Description of Culture via Analysis of Factors and Ethnographic Interviewing–Phase
Two................................................................................................................ 185
The Threat of the Introduction of Interviewer Error ............................................ 186
Phase Two Factor Analysis .................................................................................. 188
Factor Analysis of Sub-Scales: PCM .................................................................. 195
Factor Analysis of Sub-Scales: NCM ................................................................. 197
Examination of Intra-Cultural Variation Using OLS Regression Analysis ................ 200
xi
Discussion of Analysis ................................................................................................ 207
8 CONCLUSION AND DIRECTIONS FOR FUTURE RESEARCH ......................... 212
Findings....................................................................................................................... 212
Future Research........................................................................................................... 215
APPENDIX
A THE CHILDBEARING QUESTIONNAIRE (CBQ) ................................................ 219
B SPANISH VERSION OF QUESTIONNAIRE FOR PHASE ONE .......................... 222
C ENGLISH VERSION OF QUESTIONNAIRE FOR PHASE ONE.......................... 235
D SPANISH VERSION OF QUESTIONNAIRE FOR PHASE TWO ......................... 244
E ENGLISH VERSION OF QUESTIONNAIRE FOR PHASE TWO ......................... 261
F PHASE ONE SAMPLE SELECTION ....................................................................... 274
LIST OF REFERENCES ................................................................................................ 276
BIOGRAPHICAL SKETCH .......................................................................................... 291
xii
LIST OF TABLES
Table
page
5-1 Median Ideal Number of Children by Sex and Location.......................................... 113
5-2 Planned/Not Planned Status of Last Birth by Sex (in percentages).......................... 114
5-3 Has Ideal Number vs. “Dios Me Mande” ................................................................. 117
5-4 Dichotomized Ideal Number of Children. (High >= 5 or “Dios me mande”) .......... 118
5-5 Has Ideal Number vs. “Dios Me Mande” by Sex and Location ............................... 118
5-6 Dichotomized Ideal Number of Children by Sex and Location ............................... 119
7-1 Factor Loadings for PCA of All CBQ Items Sorted by Factor 1–Phase One........... 172
7-2 Factor Loadings for PCA of All CBQ Items Sorted by Factor 2–Phase One........... 173
7-3 Factor Loadings for PCA of All Non-Sex Biased CBQ Items Sorted by Factor 1–
Phase One............................................................................................................ 176
7-4 Factor Loadings for PCA of All Non-Sex Biased CBQ Items Sorted by Factor 2–
Phase One............................................................................................................ 177
7-5 Factor Loadings for PCA of All Non-Sex Biased CBQ Items Sorted by Factor 1–
Phase Two ........................................................................................................... 189
7-6 Factor Loadings for PCA of All Non-Sex Biased CBQ Items Sorted by Factor 2–
Phase Two ........................................................................................................... 191
7-7 Factor Loadings for PCA of Positive Non-Sex Biased CBQ Items Sorted by Factor 1–
Phase Two ........................................................................................................... 198
7-8 Factor Loadings for PCA of Positive Non-Sex Biased CBQ Items Sorted by Factor 2–
Phase Two ........................................................................................................... 199
7-9 Factor Loadings for PCA of Negative Non-Sex Biased CBQ Items Sorted by Factor
1–Phase Two ....................................................................................................... 202
7-10 Factor Loadings for PCA of Negative Non-Sex Biased CBQ Items Sorted by Factor
2–Phase Two ....................................................................................................... 203
xiii
7-11 Sources of Intra-cultural Variation for Pronatality ................................................ 204
7-12 Sources of Intracultural Variation for Manageable Family Size Culture .............. 204
7-13 Sources of Intra-cultural Variation for Fears About Quality of Children Culture. 205
xiv
LIST OF FIGURES
Figure
page
3-1 Two Models of the Effect of Gene/Culture Interaction. A) Common Social Science
Model; B) Alternate Model ................................................................................... 39
5-1 Latin American TFR by country 1950-55 to 1985-90. .............................................. 97
5-2 Degree of Urbanization expressed in percentage of total population residing in urban
areas. . ........................................................................................................................ 98
5-3 Tempo of urbanization expressed in difference between urban population growth
rate and the population growth rate of whole population. ....................................... 100
6-1 T-D-I-B Psychological Model .................................................................................. 125
6-2 T-D-I-B Psychological Model with Partner Interaction ........................................... 127
7-1 Scatter Plot of First Two Factors of PCA for all CBQ Items–Phase One ................ 169
7-2 Scree Plot of PCA Eigenvalues for All CBQ Questions–Phase One........................ 170
7-3 Scatter Plot of First Two Factors of PCA for all non-sex biased CBQ Items–Phase
One ...................................................................................................................... 175
7-4 Scree Plot of PCA Eigenvalues for Non-Sex Biased CBQ Questions–Phase One... 176
7-5 Scatter Plot of First Two Factors of PCA for all non-sex biased CBQ Items with
descriptions of cultural patterns– Phase One ...................................................... 180
7-6 Scatter Plot of First Two Factors of PCA for all non-sex biased CBQ Items–Phase
Two...................................................................................................................... 186
7-7 Scree Plot PCA Eigenvalues for All non-sex biased CBQ Questions–Phase Two .. 187
7-8 Scatter Plot of First Two Factors of PCA for all non-sex biased PCM Items–Phase
Two...................................................................................................................... 196
7-9 Scree Plot PCA Eigenvalues for non-sex biased PCM Questions–Phase Two ........ 197
7-10 Scatter Plot of First Two Factors of PCA for all non-sex biased NCM Items–Phase
Two...................................................................................................................... 200
7-11 Scree Plot PCA Eigenvalues for non-sex biased NCM Questions–Phase Two...... 201
xv
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
GENDER, CULTURE CHANGE, AND FERTILITY DECLINE IN HONDURAS: AN
INVESTIGATION IN ANTHROPOLOGICAL DEMOGRAPHY
By
David Patrick Kennedy
December 2002
Chair: H. Russell Bernard
Major Department: Anthropology
In this study, I investigate the context and causes of global human fertility decline.
The global total fertility rate–the average number of children born to women over their
lifetimes–has fallen for the past two hundred years. This process, which began in Europe
and continues today in the developing world, has been described but not sufficiently
explained by demographers. This study is directed towards understanding the root causes
and context of the trend in falling birth rates for populations around the world. I present
an analysis of this phenomenon with an investigation of Honduras as a case study in
fertility decline.
I present a critique of social science theories of fertility decline. I argue for a
theory that crosses disciplinary boundaries in the social sciences, addresses gender, and
includes a component that acknowledges the role of evolutionary forces on human
reproductive motivation. I also argue for recognition of the role of culture in human
xvi
behavior and an understanding of culture as an end product of human evolution. I stress
the need for empirical studies to operationalize and measure cultural variables.
I discuss my choice to adapt Warren Miller’s Childbearing Questionnaire as a
means of operationalizing and measuring what I call natality culture. I describe my data
collection technique for adapting and administering this scale using a variety of sampling
procedures on 400 people living in and around the city of Catacamas, Olancho,
Honduras.
I present an analysis of these data and data collected nationally on reproductive
attitudes. My analysis demonstrates that there is considerable evidence for cultural
agreement about positive and negative motivations for childbearing. However, there is
also evidence for important intra-cultural diversity. This diversity is primarily associated
with sex and urban or rural location. I argue that changing evaluations of the appropriate
quality and quantity of childbearing drives this cultural diversity of natality culture. I
also argue that these changing evaluations are related to shifting notions of appropriate
masculine and feminine roles. Changing economic opportunities that affect parents’
evaluations of their contributions towards their children’s futures cause this shifting of
gender roles.
xvii
CHAPTER 1
INTRODUCTION
Total fertility rates around the world–the average number of children born to
women over their lifetimes (TFR) (Yaukey 1985, p. 148)–have been dropping for the past
century and a half. It once was common for populations to average seven or eight births
per woman. Today the overall world average is under three, with some regions having
fertility rates edging toward one birth per woman (Potts 2000). The process that began in
the industrialized nations of Europe in the 18th century has now spread to parts of the
developing world. Explaining this phenomenon, called the modern demographic
transition, was once the sole domain of the demographers who developed demographic
transition theory (DTT). They postulated that the same economic changes that began in
Europe, called modernization, would eventually spread around the world, first causing
mortality rates to drop from very high to very low levels, leading to population growth
and then population pressure, causing a rationally inspired increase in the use of
contraceptives (Notestein 1945).
As demographers gathered evidence in an attempt to support and refine the theory,
contradictions to its postulates began to emerge. In order to resolve these contradictions,
DTT has been repeatedly modified in ways that have violated the scientific goal of
parsimony. The list of variables that have been theorized as causing the drop in fertility
rates continues to grow. This situation has prompted voices from both inside and outside
of demography to question the validity of DTT. In the past thirty years, many attempts to
replace the theory have emerged from researchers across the social sciences.
1
2
The question of why the reproductive behavior of populations around the world
would undergo this drastic and sometimes sudden shift is such an important and basic
question to all of social science that it rightly stands outside of disciplinary boundaries.
There is a wide array of interacting influences on human reproduction, including social,
psychological, cultural, economic, and biological influences. This pushes the explanation
of reproductive behavioral outcomes outside of the expertise of any one discipline.
Researchers in disciplines such as psychology, sociology, anthropology, evolutionary
biology, economics, and history are working to enhance the work done by demographers
to resolve the conflicts and fill in the gaps of DTT.
Unfortunately, this multi-disciplinary activity has not necessarily led to the
development of an interdisciplinary theory of fertility decline. Fertility decline
researchers have often developed new theories that concentrate on the aspect of fertility
decline that is pertinent to their disciplinary expertise. This concentration often comes at
the expense of theory that might be relevant in other disciplines. In order for an
interdisciplinary theory of fertility decline to develop, there has to be simultaneous
attention given to certain fundamental aspects of human reproduction.
The Essential Elements to an Interdisciplinary Theory of Fertility Decline
Human Reproductive Biology
First, a truly interdisciplinary theory of fertility decline must take into account how
the human species reproduces biologically. The understanding of this reproductive
system must be consistent with the evolutionary history that has shaped human biology.
The species Homo sapiens evolved under specific environmental conditions in the recent
past (relative to evolutionary time). A theory of human reproduction that does not take
this recent evolutionary past into account is incomplete. An understanding of the
3
evolutionary roots of human reproductive behavior should include not only an
understanding of the physiological needs and limitations of human reproduction–such as
coitus or gestation–but it also should include an understanding of the evolved structure of
the human central nervous system. This is important in order to understand how human
beings process the environmental information necessary to make choices among
behavioral options.
Culture and Human Reproduction
An understanding of the human central nervous system should include an
understanding of how human beings process cultural information and the connection
between culture and reproductive behavior. Human beings’ capacity for culture should
be understood as the end product of a unique evolutionary history. The degree to which
human beings rely on cultural learning to influence their behavior is unmatched by any
other species. Human genotypes allow for (but do not determine) a variety of human
behaviors that vary within and across populations. This behavioral trait is referred to as
plasticity, which implies that there is limited flexibility on a range of possible behaviors.
The limitation in behavior is in part due to a limitation on what and how human beings
are able to learn (Tooby and Cosmides 1992). Although there are limits, this range is
much wider than any other species (Knight et al. 1999).
A complete understanding of culture includes the culturally specific ways in which
populations access resources, the institutions that develop to influence continuity in
attitudes and behavior, and the shared meanings that groups of people give to the events
and things in the world around them (Harris 2001a). Also, there should be a recognition
that certain reproductive behaviors are culturally variable while other behaviors are
4
human universals. In addition, an account of culture must include not only the function
of culture but also the determinants and process of culture change.
Gender and Human Reproduction
Attention to the development of gender roles, including how they are maintained
and how they change, must be central to any theory of fertility decline. This is because,
by definition, fertility decline represents changing gender roles. In all societies, the two
human sexes display varying patterns of behaviors, especially with regard to reproductive
behaviors (i.e., mating effort, childcare effort, etc.). However, these patterns differ
among populations and over time. This spatial and temporal variety in patterns of sextyped behavior is what is meant by gender. Fertility decline represents a shifting pattern
of gendered behavior over time with regard to reproductive effort. Research on fertility
decline must have a place for defining gender institutions, the historical circumstances
under which they developed, and the forces that cause them to change.
A complete understanding of gender institutions acknowledges that they are not
created in a vacuum and are often shaped by the ways in which human populations access
resources from the environment. Although reproduction is an important human behavior,
human beings need to secure resources in order to reproduce. Depending on the material
constraints of particular environments, human beings develop a variety of ways of
structuring how they access environmental resources. Gender systems influence
reproductive behavior, but they also influence resource-seeking behavior. An important
human universal is the division of labor by sex. This division of labor is gendered
because it varies across populations. The shifting definition of the sexual division of
labor has an important impact on gendered reproductive behavior. With changes in the
cultural conception of what types of work are “masculine” or “feminine” there is an
5
effect on the reproductive choices made by individuals. An approach to fertility decline
that does not consider the impact that environmental resources have on gendered
behavior is incomplete.
Falsification of Fertility Decline Theories
Any successful theory of fertility decline must encompass a means of testing ideas
about the change process. This includes ways to operationalize all of the above factors,
including culture, and the ability to collect data in order to test the theory in multiple
populations worldwide. Any theory of a complex process cannot be tested directly.
Instead, various hypotheses result from theories, stating what should be found according
to the theory. If what is expected is not found, the theory must be modified to account for
the new evidence. Without a process of data gathering and analysis that allows for
falsification and comparison of competing theories, the acceptance or rejection of a
theory relies more on persuasion and politics than scientific merit (Johnson 1998).
Anthropology and Theories of Fertility Decline
As researchers in a variety of disciplines have challenged the assumptions of DTT,
new theories of fertility decline have emerged. These theories have been largely
unsuccessful in crossing disciplinary boundaries because they do not address
reproductive biology, culture, gender, access to resources, or they are not testable. The
work I present in this manuscript is guided by an attention to all of the above elements.
The empirical contribution this work makes to the effort to understand human fertility
decline is primarily in defining and measuring the cultural aspects of fertility. Among the
above criteria, culture is probably the most neglected aspect in theories of fertility
decline. Although many theorists have discussed culture’s role in fertility decline, culture
is generally neither defined nor directly measured. Words such as “pronatal” and
6
“antinatal” have been used to describe culture’s positive or negative effect on fertility
rates, but culture’s effect on fertility is typically assumed or asserted without empirical
support. The primary contribution of this work is the definition and measurement of
natality culture, which I use to test hypotheses about culture change and fertility decline.
In order to develop this definition, I relied on theory and methods developed in cognitive
anthropology.
Although global fertility decline is a subject that deserves the attention of multiple
disciplines, anthropologists should have a particular interest in resolving questions about
causes of fertility decline. Anthropology’s traditional four field approach, which draws
on long term evolutionary trends as well as particular historical, ecological, economic,
social and cultural environments for explanations of change and continuity in human
behaviors, makes it particularly well suited to address the myriad of factors that influence
changes in human reproductive behavior (Hill and Kaplan 1999, p. 145). In addition,
many of the debates about the causes of global fertility decline represent the same debates
that have been ongoing in anthropological theory for over a century. As I will
demonstrate in Chapter 2, the explanations of fertility decline involve aspects of idealism
vs. materialism, biological vs. cultural determinism, and historical particularlism vs. the
comparative method. These are issues that anthropologists have debated for over onehundred years (Harris 2001b). The study of fertility decline as a case study in human
behavior and culture change may help to better define the arguments in these historical
debates.
Organization of the Manuscript
This manuscript is divided into eight parts. In Chapter 1, I introduce the theoretical
issues involved in developing a theory of fertility decline that is accessible across
7
disciplinary lines. In Chapter 2, I review existing theories of fertility decline. The main
goal of this chapter is to assess various theories of fertility in light of the interdisciplinary
criteria established above, and in light of their attention to changing gender roles. The
subsequent chapters build off of the theoretical evaluations I make in Chapter 2. Chapter
3 is an elaboration of the place that evolutionary theorizing has in social scientific
evaluation of human behavior. I discuss the debates within social science regarding the
possible evolutionary roots of human behavior. I discuss these debates in conjunction
with a discussion of evolutionary roots of human fertility decline. In Chapter 4, I give an
expanded definition of culture, including its place in evolutionary theory and theories of
human reproduction. I also discuss how culture may be operationalized and measured in
order to test its force on the culture change process. In Chapter 5, I introduce Honduras
as an area of the world that offers an opportunity to test theories of fertility decline,
culture change, and changing gender roles. I use Honduras as an example of the
difficulties inherent with traditional demographic attempts to explain fertility decline. In
Chapter 6, I introduce the Childbearing Questionnaire and I discuss its potential use in
measuring natality culture and testing theories of fertility decline. I describe the process
of adapting this instrument in a field site located in central Honduras. In Chapter 7, I
discuss my findings from the administration of the Childbearing Questionnaire on two
samples of 200 people living in and around Catacamas, Olancho, Honduras. In Chapter
8, I discuss directions for future studies that can build on this research.
CHAPTER 2
DEMOGRAPHIC TRANSITION THEORY AND ITS AFTERMATH
Introduction
In this chapter, I look at various theories that have developed to explain global
fertility decline. The theories include demographic transition theory, John Caldwell’s
wealth flows theory, W. Penn Handwerker’s resource access theory, Gregory McMichael
and Susan Greenhalgh’s theories of institutional determinants, and Hillard Kaplan’s
theory of embodied capital. These theories are not an exhaustive list of fertility decline
theories. Rather they represent chronological changes in fertility theories during the past
three decades. Also, they represent theories that give varying attention to economic,
structural, ideational and biological explanations of fertility decline. I review each theory
and assess how well it addresses changing gender roles and how well it contributes to an
interdisciplinary theory of fertility decline. I argue that a combination of Kaplan’s
embodied capital theory and Handwerker’s resource access theory is the most profitable
approach to understanding human fertility. I conclude by stating my working hypothesis
for the remaining sections of this manuscript: changes in resource production create a
novel dynamic for women in their evaluation of the cost-benefits to spending energy on
reproduction vs. spending energy on pursuing resource acquisition opportunities. This
drives a culture change process that has subsequent motive force for men.
8
9
Demographic Transition Theory
Demographic transition theory (DTT) was formulated in the early part of this
century by demographers who were trying to understand the process of third world
population change (Davis 1963; Notestein 1945; Thompson 1929). The developers of this
theory used the Western European fertility decline that occurred during the industrial
revolution as a model for how the process would work around the world. Early
formulators of DTT divided up the Western European demographic transition into a
three-phase process. The first phase started with a balance between high levels of
fertility and mortality. As European countries underwent industrialization, their mortality
levels dropped. This created much higher population growth rates than had previously
been experienced by European populations. Next, once mortality levels decreased,
fertility rates were out of line with mortality, causing populations to grow at
unprecedented rates. Demographers theorized that population pressure influenced
individual households to reduce fertility in order to reduce economic strain. They
assumed that people rationally began to use contraceptives to lower their number of
children, causing a move away from so-called natural fertility. As this happened
throughout the population fertility dropped and returned to its previous balance with
mortality.
Demographic transition theory postulated that this same process would be repeated
around the world. Once developing nations had enough industrialization and
improvements in health care and nutrition, they would experience a drop in mortality that
would eventually lead to a drop in fertility. However, demographers noticed that at times
populations in developing nations were not conforming to these theoretical assumptions.
Often, populations in the developing world did not reduce fertility even though it seemed
10
as though they were experiencing what seemed to be severe overpopulation.
Demographers attributed this to the existence of traditional cultures that valued a high
number of births and resisted contraceptives. Demographers supported family planning
programs that educated people about contraceptives and attempted to replace traditional
beliefs about family size with economic rationality.
DTT remained the unchallenged explanation of fertility decline during the decades
after WWII even though many of its theoretical assumptions had not been tested. In the
atmosphere of post-WWII, demographic theorizing about the causes of fertility decline
were linked to policies that sought to westernize third world populations living in the
battlegrounds of the cold war (Greenhalgh 1996). Thus, those fighting against
communism had an interest in the veracity of DTT.
Criticism of Demographic Transition Theory
DTT was challenged in the 1970s and 1980s when the Princeton European fertility
project and other studies examined historical data on the fertility decline in Europe
(Coale and Watkins 1986; Greenhalgh 1995a). The goal was to test the hypotheses of
DTT on historical data from Europe. It was clear from these studies that industrialization
was the primary cause of fertility decline and that socioeconomic conditions were linked
to specific fertility declines. This association, however, was found to be weakly
predictive. The onset and pace of each case of fertility decline varied greatly (Bongaarts
and Watkins 1996). Regions that had low levels of development tended to have higher
fertility than did regions with higher levels of development. However, regions with
intermediate levels had sometimes high and sometimes low fertility rates. One finding
indicated that rural regions adjacent to urban areas that had experienced recent declines in
11
fertility often experienced their own declines without accompanying economic
development (Bongaarts and Watkins 1996; Watkins 1986).
DTT began to receive criticism after these studies. Critics pointed out that DTT
was not a theory because it accommodated every causal variable (Hirshman 1994). The
main problem with DTT is that demographers have approached the explanation of
fertility decline as an exercise in determining the elements of a regression equation, with
fertility rates acting as the dependent variable. Their goal is to determine the independent
variables from a collection of aggregate socio-economic variables. Because there are no
obviously strong associations with a few variables and because the operationalization of
indicators of modernization can be tweaked and re-tweaked ad infinitum, demographic
explanations of fertility decline have mostly amounted to descriptions of associations
between fertility and aggregate economic indicators.
Although this strategy has not yielded much progress in the parsimonious
determination of underlying causal mechanisms, it seems to be alive and well in current
demographic analyses of fertility decline. In a recent synthesis of the state of the art in
demographic explanations of fertility decline, Bulatao lists eight so-called explanations of
fertility decline: 1. mortality reduction, 2. reduced economic contributions from children,
3. opportunity costs of childbearing, 4. family transformations, 5. vanishing cultural
props for childbearing, 6. improved access to effective fertility regulation, 7. marriage
delay, and 8. diffusion (Bulatao 2001, p. 2-3). Bulatao describes these eight forces as
reasons for fertility decline and comments that they show up in one demographic theory
after another. He admits that these forces can be seen as either separate, independent
forces, or effects of another higher order force, such as changing education.
12
This approach does little to advance an overall theory of fertility decline. Although
each cause is associated empirically with fertility decline in one case or another, there is
no theorizing about what caused each association in the first place, or why all the
conditions may be advancing simultaneously today. Without a theory of fertility decline
that has a deeper connection with a theory of human behavior in general, any aspect of
modernization that can be operationalized into a variable and can be demonstrated as
having even a weak connection with a change in fertility rates can qualify as another socalled explanation of fertility decline.
Certainly demographic transition theory does not meet many of the criteria that I
set out in the previous chapter for a successful interdisciplinary theory. DTT did
correctly identify the primary economic change (industrialization) that set in motion the
European fertility decline. This same primary cause continues to affect the developing
world today (Bongaarts and Watkins 1996). However, there are major problems with
DTT as a theory. Greenhalgh (1996) has discussed the historical roots of the
underdevelopment of demographic theory of fertility. She argues that government
interest in stemming the tide of world population growth and the subsequent strong
support to family planning programs enabled demographers to legitimate demography as
an individual, scientific discipline. She argues that because DTT played such a big part
in legitimizing the nascent discipline of demography, there was a disincentive to
incorporate theoretical developments from other fields into demography’s primary
theoretical product.
This disciplinary isolation resulted in the unsophisticated use or outright neglect of
many possible influences of the timing of fertility change. Such influences include
13
culture, biology and gender. The concept of culture that is generally used by
demographers is one that assumes equivalence between social categories (such as
speakers of a language, members of an ethnic group or residents of a region or nation)
and shared beliefs and behaviors. As Gene Hammel wrote in 1990, “the use of ‘culture’
in demography seems mired in structural-functional concepts that are about 40 years old,
hardening rapidly, and showing every sign of fossilization” (Hammel 1990). Biology is
incorporated into demographic transition models in the form of changing rates of
proximate determinants of fertility (rates of marriage, abortion, contraceptive use, and
duration of postpartum infecundability) (Bongaarts 1978). There is generally no
framework of evolutionary influence on human reproductive behavior in most
demographic fertility models. Gender is almost completely ignored within the DTT
framework. Men and women are treated as the same rational actor with the same
incentives for birth limitation and couples are assumed to be in agreement about their
reproductive goals.
Ironically, one positive aspect of demographers’ use of transition theory, which
should be imitated by the developers of any new theory of fertility decline, was their
dedication towards operationalizing variables and testing hypotheses. Even though the
results of these tests gave critics of DTT the ammunition needed to challenge its status as
the explanation of fertility decline, the fact that some of its claims were refuted shows
that it had the quality of being falsifiable. If transition theory was based on persuasion
and unoperationalizable concepts, it would have been much more difficult for social
scientists to develop new ways of approaching fertility decline. The political influence
enjoyed by demographers would have carried DTT much further and would have
14
minimized theoretical competition. Once demographers began to show that some
assumptions of demographic transition theory did not stand up under the weight of
empirical evidence, other theories that addressed aspects of fertility once neglected by
demographers began to emerge.
Responses to the Failure of Demographic Transition Theory
Wealth Flows Theory: Caldwell
John Caldwell’s wealth flows theory was a response to the limitations of classic
demographic transition theory, especially the characterization of high birth rates as
irrational and non-economically reasonable. His theory argued that the model of large
numbers of children being nothing but an economic hardship was not relevant to parts of
the world that relied on agricultural production (Caldwell 1982). Wealth flows theory
stated that, for those populations in a high plateau of fertility, there were economic
benefits to having many children. In these settings, wealth, defined as “all the money,
goods, services and guarantees that one person provides to another” (Caldwell 1982, p.
333) flowed up from children to their parents. The more children parents had the
wealthier they were. This was because children worked on family land. Also, they were
a source of insurance once the parents were old and could not work for themselves.
Caldwell postulated that the direction of these wealth flows changed with the
importation of Westernization–especially Western style education. Once these different
ideas entered a population’s consciousness, people no longer thought of children as
wealth but rather as pure expense. Caldwell states that the fertility transition brings with
it “ultimately an emphasis on what parents owe children rather than what children owe
parents” (Caldwell 1982, p. 97). This ideational change in how parents viewed the
economic relationship with their children causes fertility decline.
15
Modification of Wealth Flows Theory: Handwerker
The anthropologist W. Penn Handwerker modified wealth flows theory into what
he calls “resource access theory” (Handwerker 1986c; Handwerker 1989). Handwerker’s
theory is a restatement of Caldwell’s wealth flows theory with an emphasis on economic
changes and gender roles. Handwerker accepts Caldwell’s hypothesis of changing wealth
flows as the main force behind fertility decline but he disagrees with Caldwell on the
source of the change in wealth flows. Rather than pinpointing education as the root cause
of change, Handwerker cites changing competition for resources.
Handwerker’s interpretation of the change in wealth flows is that it is a culture
change process. He credits economic changes, in particular changes to the means of
accessing resources from the environment, as the primary force behind culture change.
In structuring his argument, Handwerker starts with the basic necessities of life. He
describes life as “an open energy system controlled by nucleic acids” (Handwerker 1989,
p. 20). This requires all life forms, including human beings, to receive regular inputs of
energy in order to survive as life forms. He uses the term resources broadly to refer to all
energy inputs and the means of accessing them. Channels, or gatekeepers (Handwerker
1989, p. 32), are terms he uses to describe a particular means of accessing a resource.
The ratio of gatekeepers to people seeking the resource is an indication of the power
individual gatekeepers will have. As this ratio changes from place to place or time to
time, there are significant effects on the social structure as a result.
The primary systems of resource production in human history have been hunting
and gathering, intensive agriculture, and industrialization. As populations shift from one
of these forms to another, there are important consequences for fertility. A shift from
hunting and gathering to intensive agricultural production raises the amount of reliable
16
resources that can be harvested from the environment, in the form of land and livestock,
but it lowers the overall channels to access these resources (Handwerker 1989, p. 38). As
a result, there are fewer gatekeepers who have more power relative to those who seek
economic resources through them. This results in a cultural system that promotes the
exploitation of personal relationships for economic gain. In these cultural systems, in
which this exploitation is considered morally correct, the “moral economy” of parentchild relationships is altered through an emphasis on what what children owe their
parents rather than what parents owe children (Handwerker 1989, p. 17).
One type of relationship that typically develops under these circumstances is the
economic exploitation of women by men. Anthropologists have often characterized
relationships between men and women in hunting and gathering societies as (relatively,
not perfectly) sexually egalitarian (Fisher 1992; Harris 1989, p. 280; Leacock 1981;
Margolis 2000, p. 2). Certainly women in hunting and gathering societies have more
access to resources compared with agricultural societies. Also, women in hunting and
gathering societies often are more productive and reliable caloric producers than men
(Bird 1999). However, in intensive agricultural societies with sedentary lifestyles, there
is a distinct shift to subordination of women by men. Anthropologists have hypothesized
that this subordination results from the physiological difference of greater male strength
and size, on average, than women. This difference in size and strength is most evident in
warfare (Harris 1989, pp. 88-93) and plow agriculture (Fisher 1992, p. 279). Culture
systems highlighting the physiological differences between the sexes in these two
important areas caused the subordination of women by men forming systems of sexual
17
stratification. This resulted in the ability of men to control the means of resource
production, land and livestock, and exploit women economically.
Handwerker argues that, in this environment, women rely on their children as
resource gatekeepers to reduce the economic power men hold over them. This strategy
increases the ratio of resource access channels to people accessing them (Handwerker
1989, p. 49). Since children are also economically valuable to men for their role in
agricultural labor, women use their position as gatekeepers to the resource of children to
increase their power relative to men. As a result, in this environment, children are
culturally defined as wealth because of their position as resource gatekeepers. This has a
result in increasing fertility rates for intensive agricultural producers in comparison with
hunting and gathering populations, which is a well-documented phenomenon (Gillian et
al. 1993).
In the shift from intensive agricultural resource production to industrial production,
there is an opposite effect on both fertility and the ratio of resource channels to resource
seekers. The shift from intensive agriculture to industrialization, like the shift from
hunting and gathering to agriculture, represents a major improvement in the ability of
human populations to harvest and store energy supplies from the environment. However,
unlike the shift from hunting and gathering to intensive agriculture, industrialization
increases these channels in the form of competitive labor markets. Thus, the power of
individual gatekeepers is diminished. This creates a disincentive to rely on other people
as resource channels. Women’s power relative to men increases because industrial
economies minimize the importance of physical size and strength. This opens up
opportunities for women to access resources independently of men.
18
As a result, there is a diminished incentive for high fertility. As opportunities arise
for women to pursue income opportunities in which they are able to act as their own
gatekeepers they tend towards lower fertility. This is because children now represent
expense rather than wealth. Handwerker dismisses education as an independent cause of
the conceptualization of children as expense rather than wealth because education merely
acts as the means for women to gain access to resources (Handwerker 1989, p. 22).
Handwerker argues that the above scenario leads to a culture change process in the
economic evaluation of children. The reversal of wealth flows between children and their
parents is really a shift in the cultural construction of wealth. Wealth is a “conceptual
model of material reality” (Handwerker 1989, p. 29). It is the cultural representation of
resources and the means of accessing them. As people rely less on other people and more
on themselves to access resources, the moral economy of child-parent relationships shift
to emphasize what parents owe their children (Handwerker 1989, p. 17).
Evaluation of Wealth Flows Theories
These wealth flow theories of fertility decline are an improvement over the theories
developed in the name of DTT. Both theories involve changing household economics as
a factor influencing the decline, but wealth-flow theories move beyond the ethnocentric
notion that social changes that happen in Western nations are the most desirable and
beneficial and will eventually be copied by the rest of the world. Wealth flow models are
able to account for rationales of high plateau levels of fertility as well as the causes of
their decline. They do this without relying on the assumption of irrationality caused by
traditional cultural beliefs.
Handwerker’s emphasis on the influence of changing resources as the principal
driving force for fertility decline makes his a much more successful version of wealth
19
flows theory. Caldwell relies on an idealist-driven model of the cause of change, which
does little to account for the ultimate causes of change in attitudes about family size. His
theory of importation of Western-style attitudes never addresses the question of how
these attitudes developed in the first place. Handwerker’s theory successfully integrates
the underlying resource base with the systems that develop to acquire these resources and
the meanings given to this process by a population. His theory also accounts for changes
in systems and meanings by highlighting the changing availability of resources.
Handwerker also accounts for the differing motivations of women and men within
changing gender role systems. His theory accounts for the shifts in the sexual division of
labor, which lead to varying access to resources for men and women. Fertility decline is
cast as an indirect result of the changing sexual division of labor brought about by
shifting resource availability.
Resource access theory is also successful because its concepts are can be
operationalized and falsified. Besides Handwerker’s own tests of hypotheses generated
by resource access theory (Handwerker 1998a; Handwerker 1989; Handwerker 1991;
Handwerker 1992; Handwerker 1993; Handwerker 2001), there also have been
independent tests of resource access theory. Bradley (1997) tested the theory in Kenya.
She tested the hypothesis that contraceptive use would begin among Kenyan women once
education was combined with access to power. Independently, education was not
predictive, but when combined with power it predicted contraceptive use. Bradley has
characterized resource access theory as the best available theory to understand fertility
decline (in the fieldsite of Igunga sublocation) because it lays out the mechanism of
wealth flows (p. 247). Moore (1998) tested resource access theory on data from China.
20
He found that as women increased their education relative to men, there was an inverse
relationship with fertility. This relative increase was a change from previously
disproportionate amounts of men receiving higher education. As women narrowed this
gap, the total fertility rate went down. Moore considered this a tentative test of the
resource access corollary to wealth flows theory, but agrees that a more complete test
would determine if these educational gains by women were translating to a corresponding
increased level of power in the workplace.
Empirical Problems with Wealth Flows Theories
Although there have been successful tests of aspects of resource access theory, one
major drawbacks of the wealth-flow line of fertility decline research is its inability to
produce convincing empirical evidence that there is a net flow of resources from children
to parents in high fertility regimes. There is evidence that in situations where people
must rely on other people to access resources, fertility rates are high. As people are able
to rely more on themselves for income, fertility rates drop (Handwerker 1986a).
However, there is no evidence that children are actually net producers in any situation.
Studies that have measured the actual time and resources spent on children, children’s
actual production, and the continued production of adults into old age indicate that
children are unable to exceed what they have been given by their parents with their
reciprocal production (Kaplan 1994; Turke 1989; Turke 1991).
These studies indicate that children’s economic participation is often assumed to be
greater than it actually is and the economic participation of older adults is often underestimated. These studies cast some doubt on the wealth-flows models and raise the
possibility that an association between lower fertility rates and a reversal of the moral
economy of childbearing may be spurious. Some have argued that Caldwell never stated
21
that direct economic benefits given by children to their parents are their only source of
wealth. Rather, Caldwell argued that children’s wealth includes things such as
guarantees (Caldwell, 1982: 333), safety (p. 334), pleasures (p. 334), and promised
economic benefits. In a criticism of Turke’s (1989) assessment of Wealth-Flows Theory,
Fricke (1990) notes that insurance companies provide a form of wealth in security to their
customers yet they certainly do not have a negative balance in payments (p. 112).
However, this line of reasoning actually supports criticisms of wealth-flows theories
because it implies a more inclusive account of the services provided by parents to
children. If children provide the service of security and safety to their parents in old age,
parents provide these same services to their children for a longer period of time and at a
much higher intensity. No study has quantified the amount of safety, pleasures, and
promised benefits that pass from parents to their children, but I would guess that few
would predict that there exists a society where parents receive more of these items than
they give.
The above arguments may refute the idea that children can ever be a source of net
wealth to parents, but it is clear that there is a fundamental difference between the
economic relationship that parents have with their children in pre- and post-transition
societies. The aspect of resource access theory that needs to be emphasized is the need
for people to access resources through other people (not just children) in pre-transition
societies and the ability of people to access resources on their own in post-transition
societies (Handwerker 1986b, p. 15). Later in this chapter, I will argue that the missing
ingredient in resource access theory is a greater recognition of the part that biology plays
in influencing reproductive motivation.
22
Institutional Determinants of Fertility Decline
A different response to the inadequacies of DTT comes from those advocating an
institutional approach to fertility decline. The Institutional approach, championed most
notably by the demographer Gregory McNicoll and the anthropologist Susan Greenhalgh,
advocate greater attention to the historical forces that have shaped institutions such as
family systems, gender systems, political systems, as well as economic systems
(Greenhalgh 1988; Greenhalgh 1990; Greenhalgh 1995a; McNicoll 1980; McNicoll
1994). This approach was a response to the rational economic decision making models
inherent in DTT. The institutional approach argues that individual decisions can never be
made outside of the confines of the prevailing institutional setting. Institutions, as
McNicoll sees them, are, “clusters of behavioral rules governing…human actions and
relationships in recurrent situations” (McNicoll 1994, p.201). These institutions are
developed over time to “address recurrent problems about material realities (resources,
technologies, environmental conditions, human biology) and as manifestations of
ideational systems that give some degree of coherence in the cultural domain of symbol
and belief” (McNicoll 1994, p. 202). Human beings still exercise free will in areas such
as fertility behavior but only within limits imposed by this institutional structure.
The institutional approach is also a response to the unilineal approach to
demographic change often included within DTT. The predictions of multi-stage mimicry
of the European fertility transition revealed the ethnocentric bias of the early developers
of DTT. When these predictions (and post-dictions as it turned out in the European
Fertility Project) did not come true, advocates of the institutional approach argued that
greater attention needs to be paid to the particular circumstances of societies undergoing
(or not undergoing) fertility decline. They argue that a greater knowledge of the social
23
framework of decisions and of the historical processes that produce this framework is
necessary to fully understand the process of fertility decline. Greenhalgh suggests, “there
are many demographic transitions, each driven by a combination of forces that are, to
some extent, institutionally, culturally, and temporally specific” (Greenhalgh 1990, p.
88). Also, “because the transition from high to low fertility works itself out over time,
the explanation of fertility decline must be essentially a historical one” (Greenhalgh
1990). In the institutional approach, understanding the specific historical process of the
formation of systems that regulate behavior is the best way to account for the failure of
fertility declines to match the model established by DTT.
Evaluation of the Institutional Approach to Fertility Decline
The institutional approach is certainly an improvement over the mostly ahistorical
demographic theories of the global fertility transition. Because fertility decline represents
culture change and one of the essential elements of this change is shifting gender roles, it
is important to understand the context of gender relations before the decline began.
Gender roles represent the modal and expected behaviors of men and women for a
particular group of people in a particular time and place. Declining fertility represents a
change in reproductive behaviors for men and women. It also represents a change in the
behavioral expectations for men and women as spouses and parents. Understanding the
trajectory of these changes in gendered behaviors in response to industrialism requires an
understanding of the starting points.
Elements of social structure, such as gender roles, do not re-invent themselves from
scratch when culture changing forces exert pressure on them. Rather, they shift from
some starting point to accommodate new material realities. Because there are a variety of
gender roles throughout the world, there should be a variety of responses to the forces
24
behind global fertility decline even if the forces causing change are consistent.
Traditional demographic theorizing has often neglected the potential that diversity in
historical and cultural situations can potentially play in creating diverse demographic
responses to similar forces (Mason 1997; Mason 2001).
Critique of the Institutional Approach
The institutional approach offers an important, often neglected, view of how human
behavior is limited and regulated by a social structure that is the result of historical
processes. However, it has some major deficiencies in its ability to offer a new theory of
fertility. In fact, the institutional approach is more of an anti-theory in that its advocates’
main goal is to highlight the difficulty of generalizing about fertility decline. Greenhalgh
criticizes attempts to extract reproductive behavior from its historical context in order to
establish views of behavioral regularities that all people may share in common
(Greenhalgh 1995a, p. 20). McNicoll argues that the coherence and seeming logic of
institutional settings that influence human behavior is neither coherent nor logical and for
the most part the result of arbitrary historical forces (McNicoll 1994, p. 204).
Geenhalgh takes the argument further by implying that all theories that share
DTT’s attempt to develop a unifying theory of fertility decline are in essence repetitions
of 19th century evolutionary theory. These theorists assumed that all societies were
irreversibly headed towards the same, progressively better, more civilized society
(Greenhalgh 1995a, pp. 5-6, 15-16). The assumptions made by the 19th century
evolutionists, and the first demographers who formulated DTT, were certainly
Eurocentric for assuming an inevitable, progressive nature to human social change, with
Europeans in the lead. However, Greenhalgh’s analogy breaks down because the work of
these evolutionists was not flawed because of their attempt to explain the consistencies in
25
social change. Instead, their work was unsuccessful because of racist assumptions that
were not based on empirical data (Harris 2001b).
A focus on institutions and their effect on fertility decline can be consistent with a
hypothesis that there is some consistent force that can account for much of the variance in
the “many kinds of fertility patterns” (Greenhalgh 1995a, p. 17) around the world.
McNicoll characterizes the function of institutions as “ways for people to deal with
recurrent problems about material realities” (McNicoll 1994, p. 202). Because of our
specific evolved physiology and our varied but limited types of technology, there is a
limited number of ways in which human beings can harvest energy from an environment
for survival. There is also a varied yet limited number of environmental types.
Therefore, historically evolved institutions have had a limited number of material realities
to address. If institutions are a means of dealing with a limited supply of material
problems, it is reasonable to assume that there should be some consistency with how
these institutions have developed. To what extent and under what conditions this
consistency exists is an empirical question that needs to be explored. An a priori denial
of any consistency with which populations have undergone fertility change is harmful to
the pursuit of greater understanding of human reproductive behavior.
An analogy between theories of fertility decline and theories of the origins and
spread of agriculture is helpful at this point. It is one thing to explain the use of
agriculture as evidence of higher civilization and higher order on the cultural
evolutionary scale (Morgan 1985). It is another to explain it in terms of a result of
population pressure (Cohen 1977). The first theory is racist, ethnocentric, and assumes
inevitable progress. The other characterizes agriculture as the end result of a struggle for
26
survival. Both theorists tried to explain the same phenomena, which was that at some
point agriculture was invented and it subsequently spread throughout the world. Fertility
decline is a similar phenomenon because it began at one point in time and has spread to
populations in every region of the world. Greenhalgh herself admits that, “In every
region of the world, couples have been having fewer children, shrinking the basic unit of
social life” (Greenhalgh 1995a, p. 3).
Have each of these regions undergone a unique historical process that
coincidentally parallels the unique historical processes of all other regions around the
world producing the same result? Obviously this is a stretch of the imagination and does
not even conform to the stated goal of the institutional determinists. Greenhalgh allows
for some generalization of reproductive processes. She states that, “Once enough cases
are collected and understood, they might serve as the building blocks of more general
understandings of reproductive dynamics” (Greenhalgh 1995a, p. 17). This view echoes
the research agenda of Franz Boas and his students in the early part of the 20th century.
This placement of the particular ahead of the general has been a common theme in the
history of anthropological research (Harris 2001b). The idea that detailed information
about particular societies will eventually allow us to come to some general conclusions
about regularities in human behavior continues to be a strategy of over-correction against
those who generate theories devoid of historical context. However, this strategy taken to
the extreme is destructive to the scientific enterprise. As Marvin Harris stated about the
Boasian program: “The proposal to substitute concrete historical data for speculative
deductions about history needs no defense. On the other hand, to deprive science of
speculation altogether is to deprive it of its very life blood” (Harris 2001b, p. 286).
27
The approach of the institutional determinists is a purely inductive approach to
understanding reproductive behavior. However, the cycle of scientific investigation must
include both deduction as well as induction (Kuznar 1997). General theories of fertility
decline can be developed through an inductive process, but at some point testable,
empirical hypotheses must be deduced so that the general theories can either be falsified
or accepted. The type of data collected is directed by these hypotheses. Without
deduction, there is danger that the type of data collected will not be capable of being used
to test general theories. Therefore, whatever general theories that may develop from
multiple “whole demographies” (Bachrach 1999, p. 23; Greenhalgh 1995b, p. 112) that
have not been collected with testable hypotheses in mind are in danger of not being
capable of falsification, which is perhaps the most important stage of the scientific
process (Kuznar 1997, p. 42).
As a general approach to scientific understanding, the institutional-centric approach
is flawed. As an approach to explaining and predicting fertility decline (or the absence of
decline) with a multidisciplinary theory, it is almost useless. McNicoll admits that,
“Persuasive post hoc reasoning is essentially the present state of the art” (McNicoll
1980). In Greenhalgh’s analysis of fertility decline in China, she also acknowledges the
reliance of persuasion in her arguments instead of empirically falsifiable ideas
(Greenhalgh 1988, p. 632). Institutional approaches to fertility decline exhibit the same
problem as idealistic approaches, such as Caldwell’s Wealth-Flows Theory, and other
theories that treat diffusion of ideas as the cause of fertility decline.
Ideas influence behavior, but behavior also influences ideas. Explaining changes in
behavior through changes in ideas does not explain anything. One is left wondering why
28
anything changed in the first place (D'Andrade 1999). Much is the same with
institutional approaches because institutions affect behavior, but those acting within the
institutions are constantly renewing, changing, or disregarding aspects of institutions
(McNicoll 1994, p. 201). Because behavior affects institutions, we cannot be satisfied
with institutions as the explanation for the cause of behavior.
When a theory depends on persuasion, there is little prospect for acceptance across
disciplinary lines. Effectiveness of persuasion is often tied to political influence, which
rarely is transferable from one discipline to another (Johnson 1998, p. 133). The ability
to show that one theory explains a phenomenon better than competing theories is
necessary to avoid the persuasiveness alternative. Operationalization of key variables is
necessary to be clear about what is meant by theoretical constructs such as gender,
family, culture, marriage, etc. This allows for measurement, which allows for statistical
assessment of competing hypotheses generated by these competing theories. McNicoll
recognized the necessity for measurement of fertility related institutions (McNicoll 1980,
p. 459) and Greenhalgh has encouraged increased quantification for those more
qualitatively inclined (Greenhalgh 1990, p. 101). However, Greenhalgh has also
criticized demography’s insistence on quantifiability and its ignorance of
unoperationalized anthropological concepts as its “greatest weakness” (Greenhalgh 1996,
p. 49).
My argument is not that quantification should be the only pursuit for research into
fertility decline or that institutions are not an important part of fertility decline theory.
On the contrary, I think that the inclusion of institutional processes is a necessary (but not
sufficient) component of any theory of fertility behavior (or any human behavior for that
29
matter). I believe the most successful research strategy is the one set out by Marvin
Harris’ Cultural Materialism which considers social structure as an integral component of
culture and acts as a means of satisfying the material needs of a population (Harris
2001a). I agree with McNicoll when he describes the ideal research instrument as a blend
of traditional survey and anthropological methodology (McNicoll 1980, p. 457). But
quantification is an essential component of the process of theory acceptance and refusal,
especially when the scientific audience is interdisciplinary, because only with statistical
methods can hypotheses generated by competing theories be judged against one another.
After all, the quantification that was part of the European Fertility Project’s test of DTT
was the main evidence that was used to refute its claims and assumptions and cleared the
way for many new theoretical approaches to explain fertility decline.
Evolutionary Models of Fertility Decline
Among the above theories of fertility decline, Handwerker’s theory of Resource
Access meets most of my criteria for a successful theory. Unlike classic demographic
transition theory, the idealistically centered version of wealth flows theory and the
institutional theories, resource access theory explains the reasons for both high and low
fertility as well as the reasons for the change from one to another. However, it is clear
that resource access theory is not complete because it does not meet the challenge of
falsifiability when it comes to backing up the claim of reversals of wealth flows with
empirical evidence. There is no empirical evidence that supports the claim that children
represent a source of economic investment in high fertility societies. Another important
problem with resource access theory and other economically centered theories of fertility
decline is the difficulty they have in explaining why people continue to have children in
situations where they are very expensive. If low fertility societies have low fertility
30
because children are expensive then why do people in these societies continue to have
any children at all?
A solution to this problem is to augment economic theories of fertility decline with
biological variables in a “biosocial” or “biocultural” theory of fertility decline (Bock
1999; Foster 2000; Kaplan and Lancaster 1999; Udry 1996). In this section, I argue that
an understanding of human beings’ evolutionary history and an incorporation of genetic
factors into existing theories of fertility decline (in particular resource access theory) is a
way to overcome the above-mentioned problems. Rather than argue for the replacement
of a flawed theory, I argue that a consideration of biology can be incorporated into
existing economic models. The inclusion of biology explains another level of analysis
that is not fully taken into consideration in traditional theories of fertility decline. The
strength of models that include human biology is that they can identify causes of high
fertility, low fertility, and fertility change. Also, they can explain the mechanism through
which environmental forces cause fertility change. Thus, biological models of fertility
decline become more than the descriptive theorizing, which is state of the art among
many demographers.
The assertion that genes play a part in fertility decline may seem ironic in light of
Vining’s suggestion that falling fertility rates in industrialized nations offers the best
evidence against an evolutionary approach to the origins of human behavior (Vining
1986). An evolutionary approach to animal behavior stresses reproductive fitness as the
only force on long-term changes in the frequency of genes. Fitness is the differential
reproductive success of individuals, some of whom are better able to assure
representation of their genes in future generations than their “competitors” (Dawkins
31
1976; Williams 1966). The genetic qualities that enable individuals to reproduce
successfully and rear offspring (who are also able to reproduce successfully) in specific
environments are termed adaptations. These adaptations are the result of non-random
forces in the environment, which influence phenotypic development, survival, and mate
selection in addition to reproduction. It follows that those genetic and phenotypic
qualities that are directly related to successful reproductive behavior should be among the
most highly selected of any human trait. So how can it be that among current human
populations and among individuals within these populations there is a strong inverse
relationship between wealth, health, and general well being and fertility rates (Pérusse
1993, p. 269-270; Vining 1986)? Given the above description of an evolutionary
perspective of behavior, shouldn’t those who are in the best position to support excessive
children be the ones who are maximizing their fitness? Wouldn’t an inverse relationship
between health, intelligence, and wealth and fertility prove that there is not a genetic
influence on current human reproductive behavior?
The problem with this argument against genes having a role in human reproductive
behavior is that it is based on the misconception that if genes influence fertility, there
must be a gene for fertility (Haaga 2001, p. 54). Although there certainly are erroneous
uses of evolutionary theory in both the scientific literature and the popular press, there is
a fundamental dependence on environmental factors for the evolution of genes as well as
the expression of genes in behavior. In order to understand the effect of genes on
contemporary human reproductive behavior, one must first consider the characteristics of
the environment under which human beings evolved, called the environment of
evolutionary adaptiveness (EEA) (Bock 1999, p. 197). Next, one must then evaluate the
32
contemporary environments within which these genes are functioning in order to
understand their effect on behavior. Capitalism and industrialism have only existed for a
few hundred years, cities have existed for less than ten thousand years, and agriculture
has existed for less than fifteen thousand years. A million years before that, humans
lived in small groups as foragers. It is in this environment that evolutionary arguments
will have to place adapted traits (Bock 1999; Ridley 1993, p. 188-189).
Because environmental pressures on humans living in industrial societies vary
significantly from this hunting and gathering Pleistocene existence, current human
biological traits may not be particularly adaptive to current environments. Evolutionary
“progress” can only be described in terms of the past. As soon as a gene is selected
because of its adaptive quality, it is likely that the environment will have changed.
According to evolutionary theory, novel environments pose novel selective pressures on
genes. Depending on the congruence between the new and old environments, genes may
or may not be successful at producing phenotypes that lead to successful reproduction
and successful representation in future generations. Genes that are not successful
eventually will cease to exist (Ridley 1993).
But if successful reproduction is fundamentally important to the successful
representation of genes in future generations, how can there be a connection between
evolution and behaviors that limit (and sometimes eliminate altogether) reproduction?
The answer lies in the biological concept of life history theory that has been adapted by
evolutionary anthropologists to understand differential rates of human reproduction
(Bock 1999; Hill and Kaplan 1999; Kaplan 1996; Pérusse 1993). Life history theory
includes many of the same assumptions about life that are included in resource access
33
theory. However, while Handwerker acknowledges that a genetically influenced
biological drive for reproduction plays a part in cultural evolution (Handwerker 1998b),
the implications for this drive on fertility decline remain, for the most part, unexplored in
his development of resource access theory. Life history theory adds to resource access
theory’s definition of life as “an open energy system controlled by nucleic acids”
(Handwerker 1989, p. 20) requiring regular inputs of energy with the qualification that
these life forms, as part of their nature, seek to convert this energy into replicates (Hill
and Kaplan 1999, p. 399).
Like resource access theory, life history theory centers on competition to harvest
energy from the environment. This competition causes living things to devise many
strategies to improve on the efficient transfer of harvested energy into offspring. An
individual organism can only use harvested energy for one purpose at a time. A choice
that is always necessary is whether to use energy on reproduction in the present or to
spend energy on somatic effort (tissue building or maintenance, growth) that will allow
for greater potential reproduction in the future.
The evolutionary anthropologist Hillard Kaplan has augmented the Life History
theory approach to human fertility with the concept of “embodied capital” (Hill and
Kaplan 1999; Kaplan et al. 1995; Kaplan 1996). Embodied capital is an extension of
somatic effort and it includes, “strength, immune function, coordination, skill, and
knowledge” which allow for more efficient allocation of energy and time towards
“resource acquisition, defense from predators and parasites, mating competition,
parenting, and social dominance” (Hill and Kaplan 1999, p. 402). Investing in embodied
capital in the present can extend the lifespan, increase reproductive potential, and allow
34
for future energy payoff to be greater than the energy originally invested. Thus, wouldbe parents are challenged to decide between reproducing now leaving more time to
reproduce in the future, or reproducing later and having more resources to assure greater
offspring developmental success. In other words, they must decide between offspring
quality versus offspring quantity.
How does the quality versus quantity calculation differ between the EEA and
modern, industrial societies? How does this difference lead to fertility decline? In the
environment of our ancestral hunter-gatherers, a long period of training was necessary to
acquire enough skills to see payoff in terms of resource acquisition, much the same as the
present industrialized societies (Bock 1999, p. 211; Hill and Kaplan 1999, p. 402). The
difference in industrialized societies is the existence of extra-somatic forms of wealth.
This includes technological advances that have allowed immense ability to extract energy
from the environment (Hill and Kaplan 1999, p. 423). Other technological advances
include modern medicine, which lowers infant mortality rates and increases the
probability that individual children will live to adulthood, and contraceptive methods,
which allow parents to delay desired childbearing. These changes in technology have
added many unique factors into the quality versus quantity calculation for prospective
parents.
Wealth, which during the Pleistocene was strictly limited to somatic wealth (tissue
growth, tissue maintenance, learning, etc.), is now potentially distributed in seemingly
infinite proportions outside of human bodies. In addition, the economic system of
capitalism that developed along with the Industrial Revolution allowed for wealth to be
converted into more wealth with minimal human energy expenditure. Parents and
35
prospective parents in industrial societies are not only faced with a choice between
quantity and quality of children but they are also faced with a choice between expending
energy on embodied capital development (for themselves and/or their offspring) and
(further) reproduction. Given the unique possibilities for wealth accumulation, it should
not be surprising that the conflict between income seeking effort and reproductive effort
caused a drop in reproduction. Pleistocene foragers, never faced with the wealth
producing possibilities of current industrial economies, could not have evolved any form
of behavioral check that would have motivated them to eschew income investment when
it interfered with reproductive fitness.
An evolutionary view of fertility decline is also congruent with resource access
theory in that it assumes that there can be conflict between men and women with regard
to resource access and reproductive goals. The change that occurs in resource access
channels during a shift between intensive agricultural and industrial infrastructures, and
the effect this change has on gender roles and fertility, is understandable in evolutionary
terms. According to evolutionary theory, all individuals are in competition with each
other to pass on their genes to future generations. Men and women are no exception to
this rule, although they also have to cooperate with each other in order to pass on these
genes in the first place. When there is a shift from an intensive agricultural infrastructure
to an industrial infrastructure, men and women are influenced to invest more of their time
in income generation energy expenditure and less on reproductive energy expenditure.
However, the minimal effort men have to expend in order to successfully raise an
offspring is very small when compared with women, so they do not have as much conflict
over the choice between reproductive and income seeking energy expenditure.
36
Industrialized wealth production has a large effect on fertility decline because it gives
women increasing power to act as their own gatekeepers to resources. In order to take
advantage of industrially produced resource access channels, they consequently must
limit their reproductive energy expenditure.
It is my hypothesis that this conflict not only causes fertility decline as individual
women are prompted to limit their fertility, but it begins a culture change process that
effects men as well. As women become their own resource access gatekeepers, they are
able to select mates who express similar views on family size (i.e. similar views on the
distribution of energy between reproductive and income seeking behavior). Women’s
selection of men for mates based on their congruent attitudes has the effect of selecting
for men’s attitudes about family size and reproductive behaviors. I hypothesize that this
selection process leads to the acceptance of smaller families as morally correct, causing a
change from pro-natal culture to culture that can be characterized as anti-natal. As this
aspect of culture “hardens”, it becomes institutionalized, takes on inertial force and
contributes towards individual motivation (D'Andrade 1999, pp. 95-96; Strauss and
Quinn 1997, pp. 101-110). I will elaborate on this hypothesis further in the subsequent
chapters.
CHAPTER 3
EVOLUTIONARY THEORY, FERTILITY, AND HUMAN BEHAVIOR
Introduction
The incorporation of biological variables into studies of human behavior is often a
source of contention among social scientists. In this chapter, I will discuss the
ramifications of a theory of fertility decline rooted in evolutionary biology. I will address
some of the common themes of criticisms and rebuttals used by the opposing sides of
what is often erroneously called the “nature/nurture” debate. I will present an argument
that obviates the need to draw a line in the sand between those explaining human
behavior in genetic terms with those explaining human behavior in environmental terms.
I conclude by arguing that this debate should ultimately hinge on the acceptance of
theories that better explain empirical data. I argue that anthropologists have an important
role the developing and testing these theories.
Two Models of Gene/Culture Interaction
Before I discuss the use of evolutionary biology to explain human reproductive
behavior in particular, I want to make clear the model I am using to understand the
simultaneous influence of culture and genes on human behavior. In this section, I discuss
two general models of culture-gene-behavior interaction that are used by social scientists.
I am struck by the misunderstandings that develop when social scientists debate theories
that incorporate evolutionary forces on human behavior. Several recent debates make it
clear that social scientists are not always speaking the same language when they discuss
the potential evolutionary roots of human behavior (Barash 2002; Kennelly et al. 2001;
37
38
Miller and Costello 2001; Risman 2001; Udry 2000; Udry 2001; Weidman 2002). These
debates are often unproductive because, instead of discussing the advantages or
disadvantages of one theoretical model’s ability to explain empirical evidence, opponents
actually discuss two different models of the gene-culture-behavior interaction. This
difference goes unstated during the course of the debates, making the finding of a middle
ground impossible.
In this section, I will present two different models of gene/culture determination of
human behavior. I illustrate these models with two diagrams. The first diagram is a
simplification of a model used by many opponents of theories that emphasize genetic
influences on human behavior. The second diagram shows a more detailed model used
by theorists who promote attention to evolutionary influences on contemporary human
behavior. I present these models to be clear in my explication in anticipation of potential
misunderstandings. Many of the nature/nurture debates are unproductive because of a
priori positions against the possible relevance of evolutionary theory for understanding
human behavior. In my view, opponents of evolutionary explanations of human behavior
often do not realize that evolutionary theorists do not use the same model of gene/culture
interaction that they are criticizing. Defining these opposing theoretical positions clearly
should result in a more informed dialogue about the role of the evolutionary past in
understanding contemporary human behavior.
I am also presenting these models as a means of describing in detail some of the
concepts I have been using and will continue to use throughout the manuscript. The
concepts include culture, biology, genotype, phenotype, and environment. These
concepts are another source of contention in debates on evolution and human behavior. I
39
believe that at least some of the source of friction between proponents of evolutionary
theories of human behavior and those who oppose these theories is the inconsistency in
the use of these terms. These terms are used in various contexts to mean various things.
In anticipation of potential miscommunication, I will present these models as a vehicle
for providing definitions of the terms I am using in this manuscript.
A)
Genes /
Biology
B)
EEA
Behavior
Phenotype
Culture
Culture / Cultural
Environment
Environment
Environment
Genotype
Physical
Environment
Behavior
Figure 3-1 Two Models of the Effect of Gene/Culture Interaction. A) Common Social
Science Model; B) Alternate Model
Common Social Science Model
Figure 3-1 shows two models of gene/culture interaction on human behavior. The
first diagram (diagram A) is the model that I argue is most commonly used by those who
argue against genetic influences on human behavior. I’ve developed this diagram from
40
my reading of various debates on the nature/nurture question, my personal experience as
a social scientist, and from the critiques offered by D’Andrade (1999) and Tooby and
Cosmides (1992). Diagram A in Figure 3-1 shows the hypothetical relationship between
individual behavior and the interaction of two opposing forces, culture and
genes/biology. (The terms genes and biology are often used interchangeably for those
invoking this model.) In order to represent the model used by ecologically or
materialistically minded social scientists, I have included a box and an arrow to represent
environmental and behavioral forces on culture. (In diagram A, these culture shaping
forces are represented in dashed lines to indicate that not all social scientists agree with
explanations of culture in materialistic or behavioral terms.) Because this model
represents a variety of theoretical positions, it is intentionally simplistic and does not
match any one theory exactly.
For many social scientists, the incorporation of genes or biology into this model is
problematic because of human behavioral variability. Genes and culture are seen as
alternate, competing explanations of human behavior. Because the genetic/biological
evidence for the existence of races has been discredited (Montagu 1997), the genetic
influence on human behavior is considered equal across populations. On the other hand,
most human behaviors are known to vary across populations. The cultural influences on
these behaviors also vary. If cultures vary and behaviors vary but genes stay the same, it
makes sense–and is more parsimonious–for those who use this model to consider culture
as the explanation of most human behaviors while ignoring the additional influence of
genes.
41
Given this model, skepticism of genetic and evolutionary explanations of human
behavior follows. I anticipate that social scientists who use this model would be skeptical
of a theory of fertility decline that implied an evolutionary connection to this change in
human behavior. Even among social scientists who fully accept the biological influence
on childbearing motivation, the choice is clear:
While biologically determined sentimental satisfactions enter into the reproductive
calculus, we can assume that all human groups share virtually the same set of
relevant genes. We may therefore regard the potential sentimental rewards of
reproduction as exercising a constant pressure for birthing and rearing as many
children as human physiology permits. It is clear, however, that whatever the
precise strength of this pressure, it can be modified by culture so completely as to
accommodate extremely antinatal practices resulting in lifetime childlessness for a
significant minority if not the majority of men and women…Moreover, in view of
the constancy of species given impulses leading to parental and child-rearing
behavior, the focus of our inquiry necessarily must turn to those variable costs and
benefits which sometimes give full reign to human fecundity and sometimes totally
suppress it (Harris and Ross 1987, p. 11). [Emphasis added]
In this passage, Harris and Ross assess the cause of fertility decline using the model
depicted by diagram A in figure 3-1. They argue that even though genes influence
human reproductive behavior, these genetic forces are constant and pro-natal. Genes
cannot be invoked to understand fertility decline because of behavioral variability and
anti-natal behavior. They argue that it makes more sense to focus exclusively on the
costs and benefits that influence cultures to promote anti-natal behavior.
Alternate Model
Diagram B in figure 3-1 presents an illustration of an alternate model of the
relationship human behavior has with genes and culture. I’ve developed this model from
a variety of sources that describe the connection between genes, culture and human
behavior (e.g. (Barkow 1989; Barkow et al. 1992; Boyd and Richerson 1985; Cronk et al.
2000; Dawkins 1976; Dunbar et al. 1999; Durham 1991; Handwerker 1989; Low 2000)).
42
This model contains many of the same elements as diagram A in figure 3-1, with some
modifications and additions. The most important additional element in this model is the
box indicating phenotype. Instead of genes and culture directly influencing behavior, this
model depicts both of these elements influencing behavior through the phenotype.
Phenotype is the entirety of tangible, observable properties of an individual organism
(Dawkins 1976, p. 235; Durham 1991, p. 11-12; Tooby and Cosmides 1992, p. 45).
Thus, an organism’s phenotype includes all somatic tissue as well as the central nervous
system (including ideas, memories and emotions). Because somatic tissue is constantly
being built throughout the organism’s lifetime and because new experiences result in new
ideas, memories and emotions, the phenotype of an organism is a constantly changing
entity. The genotype, on the other hand, is an organism’s immutable set of coded
instructions inherited from its ancestors. It is the individual’s link to its evolutionary past
and the environment of evolutionary adaptiveness, represented on the figure as EEA.
A phenotype is typically described as the expression of an organism’s genotype,
but it also can be similarly described as an expression of an organism’s environment.
Genes produce phenotypes only after interaction with the environment. In diagram B of
figure 3-1, the environment variable has been expanded and subdivided into the cultural
environment and the physical environment. (In the next chapter, I will make a detailed
argument for why I think that culture belongs as an element of environment.) The arrows
between culture and phenotype, physical environment and phenotype, and between the
phenotype and behavior have heads on both ends indicating dual interaction. Phenotypes
develop as a result of the interaction among genotypes, the cultural environment, and the
physical environment.
43
The relationship between phenotype and behavior is somewhat complex. A
phenotype includes all observable expressions of a genotype, including behavior
(Durham 1991, p. 11). In this sense, behavior is a phenotypic trait and is meant to
indicate a category of patterned activity rather than a discrete behavioral event. These
patterned behavioral activities are assumed to be included within the “phenotype” box.
The element of diagram B of figure 3-1 labeled “behavior” is not the same as this
phenotypic behavior. I use an oval rather than a rectangle to represent this behavior to
distinguish it as a different type of analytical category. This behavior is meant to
represent a discrete, observable and measurable behavioral event. It is an outcome
variable rather than a theoretically constructed element of the genotype-phenotypeenvironment system represented by this diagram. My purpose in including this second
meaning of behavior is to highlight that the exact state of the phenotype is the immediate
cause of a behavioral event at the time of the behavior.
Although the cultural environment and the physical environment shape the
phenotype, they are also shaped by phenotypes. Phenotypes exhibit patterns of behavior.
These patterns, in conjunction with the patterned behaviors of other phenotypes in a
population or social network, create culture. (I discuss this idea further in Chapter 4).
The patterned phenotypic behavior also interacts with and changes the physical
environment to some degree. These dual interactions–phenotypes both shaping and being
shaped by environmental forces–are indicated by the arrows with heads on both ends.
Model Discussion
Using this model instead of the model represented by diagram A, a choice between
genes and culture as the primary cause of human behavior change is no longer necessary.
Genes and environments are both essential components in the creation of phenotypes and
44
phenotypic behavior. Harris and Ross’ conclusion that only cultural and environmental
forces can cause behavioral change tells only part of the story. This model shows that the
causes of behaviors cannot be partitioned into cultural or genetic. All phenotypes are
created as an interaction between environmental and genetic elements. This implies that
behaviors are always the joint result of genes and environmental forces.
As I discussed in the previous chapter in the context of life history theory, the
relationship between genes and reproduction is not as simple as Harris and Ross assume.
Genes are certainly selected for their contribution to reproductive success in past
environments. Therefore, influencing the creation of phenotypes that are successful at
reproduction is certainly a quality that genes should possess. However, this does not
mean that genes are only able to influence successful reproduction through quantity of
offspring alone. Quality of offspring is just as important to reproductive success.
Offspring quality is directly tied to environmental conditions, which complicates the
assumptions that can be made about genetic influence on reproductive motivation.
In the following sections of this chapter, I will discuss in more detail the
evolutionary mechanism for human fertility decline. In particular, I will discuss how the
EEA selected for human phenotypic plasticity, which resulted in greater reliance on
culture as a determinant of human behavior. I will discuss how this “variability
selection” affected human reproduction. Also, I will further discuss the ramifications of
this model for social scientific debates about human behavior in general and for a theory
of human fertility in particular.
45
The Evolutionary Mechanism of Fertility Decline
The Black Box of Evolutionary Theories of Human Fertility Decline
In the previous chapter, I argued that human beings evolved under environmental
circumstances that were very different than the circumstances they live under today, in
particular industrial settings. This is important because it explains how contemporary
human genetic traits, which may have been selected for in past environments because of
their positive effect on fitness, are now leading to deviations from fitness maximization
(Kaplan and Lancaster 1999, p. 315). Therefore, genes that were selected for during the
Pleistocene because they interacted with the environment to produce reproductively
“successful” humans are now interacting with a new environment to create
reproductively “unsuccessful” humans. (My use of successful and unsuccessful in this
instance is in purely genetic terms–greater or lesser representation of genes in future
generations–and implies no social or cultural construction of success.) However, this still
does not explain the specific mechanism at work that produces different reproductive
outcomes in response to various environmental stimuli. This is the “black box” of human
reproduction that many fertility decline theories do not specifically explain.
Scientists often use black boxes to act as holding places for future definition while
they proceed with the process of theory building. As Udry points out, the concept of a
gene was itself such as black box until the discovery of DNA (2001, p. 616). This does
not detract from the utility of these theories while new empirical data and greater
theoretical sophistication are being generated. Theories that explicitly recognize yet-tobe-explored gaps in theoretical completeness are more complete than theories that ignore
these gaps altogether. If reproductive behavior is a result of the interaction between
genes and environment, it is better to have an unspecified mechanism to explain the
46
effect of this interaction than to pretend that it does not exist. In my opinion, the reason
there are so many less-than-successful theories of fertility decline is because they tend to
deal with environmental causes exclusively and ignore the effect of endogenous genetic
characteristics. Our knowledge of behavioral genetics may not be to the point where we
can specifically detail the interaction between genes and reproductive behavior with
much sophistication. This makes it necessary to use theoretical placeholders instead of
detailed causal sequences. However, this is better than assuming that genetic factors do
not have any effect on reproductive behavior.
This being said, the “black box” of evolutionary models of fertility decline is not
entirely black. Evolutionary anthropologists have theorized that many different
physiological and psychological characteristics of modern humans are probably the result
of natural selection in the EEA for genes that would lead to fitness-maximizing
reproductive behaviors (Barkow et al. 1992; Cronk et al. 2000; Ellison 2001b; Hill and
Hurtado 1996). Throughout this chapter, I will discuss examples of theoretical work that
is shaping our understanding of the connection between the evolutionary past and
contemporary fertility. In Chapter 6, I will discuss in detail empirical work that is
illuminating the black box by improving our understanding of the connection between
genes and childbearing motivation (Miller 1995; Miller et al. 1999).
As our knowledge about the connection between molecular genetics and behavior
grows, hypotheses about specific mechanisms will become correspondingly
sophisticated. The theoretical placeholder of a general genetic mechanism influencing
fertility decline will eventually be replaced by theories of specific interactions between
genetic and environmental qualities.
47
EEA Variability Selection and Human Plasticity
One recent improvement in the theoretical understanding of the link between
human behavior and our genetic heritage is the recognition that the human EEA was not a
consistent selective environment. Potts (1998) has shown that the most notable
characteristic of the Pleistocene environment of our hominid ancestors is its
unpredictability and variability in selective pressures. Theories that hypothesize specific
connections between genetic characteristics and environmental qualities have to
demonstrate some type of consistency in selective pressure. The instability of EEA
conditions makes many evolutionary theories of human behavior difficult to defend.
However, there is one EEA environmental characteristic that can be considered
consistent enough to have selective force: inconsistency itself (Potts 1998, p. 94). The
variable conditions of the EEA favored behavioral mechanisms that promoted flexibility
in response to environmental variability (Bock 1999, p. 197). Mobility, genetic
polymorphism, phenotypic plasticity, and complex behavioral systems designed to
respond to novel environmental settings are all adaptive options to organisms living in
highly variable habitats and together are termed “variability selection” (Potts 1998, p. 8485). Based on Potts findings regarding the fluctuations in the human EEA, it is likely that
variability selection was at least one of the main evolutionary forces affecting early
human genetic composition. An important phenotypic characteristic of humans that
likely resulted from this variable selective pressure is the reliance on cultural inputs to
determine behavioral responses to environmental conditions. Thus, individual human
beings did not have to rely on their own direct perceptions of environmental stimuli but
could determine appropriate behavioral responses to fluctuating conditions based on a
pool of knowledge created by other human beings.
48
EEA Selection for Reproductive Motivation
Although EEA genetic selection likely contributed to plasticity in phenotypes and
behaviors, there are certain adaptive issues that are consistent across environments. One
of these is the necessity for organisms to reproduce. The specific conditions under which
this reproduction should take place fall under the category of variability selection. These
aspects of reproduction should be variable because fitness is not necessarily tied to
specific numbers of offspring. Depending on the environment, fewer offspring with a
higher level of parental nurturing effort per child may be the optimal strategy for
maximizing genetic representation in future generations. In other environments,
intensive parental investment in individual offspring may have little effect on the their
reproductive success. This makes increasing the number of offspring the optimal strategy
for fitness maximization. The timing of reproduction is also subject to variability
selection because, in terms of fitness, giving birth to offspring may be more profitable
later or earlier in an organism’s life.
Although the specifics of reproductive amount and timing are behaviors that are
likely candidates for variability selection, the fact of reproduction–whether or not to have
children at all–would not be conditional on the environment. There is no justification
within evolutionary theory for supposing that a species typical genotype could have
resulted from natural selection without some sort of a reproductive drive. Sexual
reproduction assures variation among individual organisms in the strength of this drive.
Also, there may be competing drives, resulting from the forces of natural selection, which
might lower or suppress altogether the basic drive to reproduce. This would likely
happen only under extreme environmental circumstances. However, to presume that
there could be a species with a characteristic genotype that did not influence some form
49
of phenotypic reproductive motivation is antithetical to the most basic assumptions of
evolutionary biology.
Pan-Human Reproductive Drives, the EEA and Culture
This suggests that there should exist some genetically influenced, general
reproductive drive that operates independent of environmental conditions. Drives
associated with the quantity and timing of children should be dependent on evaluation of
environmental information via cultural channels. Among the pan-human motivating
systems that have been hypothesized as being subject to evolutionary selection are
general bonding drives that promote the formation of sexual unions and nurturing (Foster
2000; Miller 1995; Miller et al. 1999). These general drives would have been sufficient
in the EEA to motivate humans to become parents. While much of this theorizing is in its
early stages, there is some empirical evidence to support these theories. As I will discuss
in detail in Chapter 6, Miller et al. (1999) have demonstrated a genetic influence on
motivation to become a parent. This genetic influence acts on parental motivation
through its influence on the development of neurotransmitters associated with nurturing
behavior. Although environmental variability would preclude a specific mechanism
dictating human parity or childbearing timing, a drive to have sex and a drive to nurture
the resulting offspring would have guaranteed a consistent level of reproduction for EEA
hominids. This is primarily because they did not have access to reliable methods of
contraception (Potts 1997).
This does not mean that these drives would act independently of cultural or other
environmental factors. These drives trigger neurotransmitters and the release of
hormones with the exposure to environmental stimuli (Miller et al. 1999, p. 58). As
shown in diagram B of Figure 3-1, a significant amount of human environmental stimuli
50
is cultural. Pan-human reproductive and sexual drives lead to the universal existence of
sexual and parental roles throughout all human populations (Brown 1991). Although
genes create these drives and they ultimately make the creation of these roles necessary,
they do not control the specific qualities of these roles (Handwerker 1989).
Culturally defined sexual and parental roles develop as a consequence not only of
reproductive and sexual drives, but also other drives. Other generalized drives to meet
the panhuman needs for survival and access to resources are also important, and may be
more important (at least in short run) in the definition of sexual and parental roles.
Depending on environmental conditions, including things such as the availability of
resources, demographic factors, health risks, etc., sexual and parental roles are modified
in response to external realities. For example, sexual roles may change when faced with
the fact of widespread sexually transmitted disease. Also, parental roles and expectations
for parental nurturing behaviors may shift under conditions of high infant mortality
(Scheper-Hughes 1992). Environmental costs and benefits may even lead to the cultural
construction of roles and behavioral expectations that lead to the direct regulation of
fertility. Examples of this include sexual taboos, including regulation of celibacy and
incest, and systems of kinship (Harris and Ross 1987, p. 11). In the extreme, fertility can
be regulated by the cultural acceptance of infanticide, especially in populations under
excessive environmental stress without access to effective and reliable contraception
(Harris and Ross 1987, p. 169; Scrimshaw 1983).
Although the existence of anti-natal cultural systems may seem to give evidence
against the conjecture that there is a pan-human, pro-natal drive to nurture, it does not.
Anti-natal cultural systems only demonstrate that genetically influenced nurturing drives
51
do not directly determine behavior outcomes independent of environmental conditions.
Variability selection in hominid evolution may have allowed for behavioral and
phenotypic plasticity, but it did not eliminate genetic factors altogether.
The Analysis of Variance in Reproductive Behavior Caused by Genes and
Environment
In general, it is best to describe the relationship among genes, environment, and
reproductive behavior as statistical, in which certain genetic and/or environmental
conditions should result in higher probability of certain types of behaviors. If the
variance of the reproductive behavior of all of the individual humans in the world was
partitioned, the variance would be determined by 1) individual genetic differences, 2)
economic and other social systems, 3) ideology, 4) the interaction among 1,2 and 3. The
remaining variation would be effectively randomly distributed or will be described by
some as of yet unknown factor (for example, the level of certain types of toxins in the
water supply). Genetically influenced motivation for parenting and sexuality are
phenotypic traits that result from the interaction of all of the above factors.
The non-genetic factors vary for a variety of reasons. Economic and social systems
vary based on how much economic and social value children are worth in a society at a
particular time. Also, they vary based on the degree that people are able to access
resources individually as opposed to through other people (as I discussed in Chapter 2).
As children have more economic and social value and as connections to social networks
increase in importance for income pursuits, motivation to become a parent will tend to be
greater. The economic value of children increases with their ability to produce income
for a family with minimum parental expense. The income that children produce
compared with parental expenses may never be a net positive for parents (see the
52
discussion of wealth flows theories in Chapter 2), but it certainly can be less expensive in
some situations than others.
Non-genetic and non-economic factors also influence parental motivation. Social
value is the degree that status is conferred upon people for having children. As social
value of children increases, parents will be motivated to have them, regardless of the
expenses associated in their upbringing (Schoen et al. 1997). Ideology, such as religious
convictions, can influence people’s reproductive behavior positively (such as thinking
contraceptives are wrong because you must accept all the children God sends) or
negatively (in the case of clergy or people who believe in doing their part to fight
overpopulation). In the long term, ideology may be more of a reflection of economic
cost-benefits rather than an independent force (Iverson 1992). However, in the short
term, ideology can be thought of as having an independent effect on parental motivation.
Genes interact with each of these factors in ways that make having children more or
less likely given a set of economic, social, or ideological factors. All of these factors that
interact with genes are not themselves uncorrelated. They are all ultimately shaped by
the means of harvesting energy from the environment. However, in the short run, a
change in one or more of these factors may lead to changes in reproductive motivation
without corresponding changes in the other factors. For example, a change in the
economic cost-benefits of childbearing might lead to lower parental motivation while the
ideology of childbearing might continue to be highly pro-natal.
Social Science Resistance to Evolutionary Explanations for Human Behavior
Many social scientists resist explanations for human behaviors involving biological
variables (Tooby and Cosmides 1992; Udry 1996). Among social scientists, cultural
anthropologists are possibly the most resistant to any linkage between genes and human
53
behavior (Irons and Cronk 2000, p. 6-12). Some of the resistance to the incorporation of
biological variables into human behavioral analyses comes from the desire to deny that
there is such a thing as human nature. This desire is often expressed in the emphasis on
behavioral or cultural variation among human groups and the de-emphasis of any
characteristic that can be thought of as a human universal (Brown 1991; Tooby and
Cosmides 1992). Geertz illustrates this disdain for generalization about human beings
with the statement, “Any sentence that begins, ‘All societies have…’ is either baseless or
banal” (Quoted in Udry 2001, p. 612). The state of the art in anthropology for disproving
the genetic connection to a human behavior is to find an example of a society that
demonstrates variability in the behavior (Tooby and Cosmides 1992, p. 43).
With the de-emphasis of human universal behaviors, many cultural anthropologists,
as well as other social scientists, conclude that any evolutionary explanations of human
behavior must imply that there are genetic differences among groups of people. This is
anathema to social scientists who see this as indicating racism. Because they value their
ability to fight against the social evil of racism with scientific data, they also fight against
evolutionary explanations of human behavior (Irons and Cronk 2000, p. 12). This
tendency is particularly strong in cultural anthropology possibly because of its history of
producing racist theories of cultural evolution in the 19th century. The fight against
generalizing about human behavior, biologically driven or otherwise, is part of a
corrective anti-generalizing tendency that has dominated cultural anthropology since the
time of Franz Boas (Harris 2001b). However, the necessary connection between genetic
influences on behavior and genetic differences among groups is a misconception (Haaga
2001, p. 54). Theorists of evolutionary causes of human behavior are more often looking
54
to explain universal human genetic characteristics that vary by individuals. They seek to
explain how these pan-human characteristics express themselves as patterns in behavioral
variation due to the interaction between genes and ecological variables (Ellison 2001a, p.
1).
Although, for the most part, evolutionary theorists are attempting to explain
contemporary human behavior through the interaction between genes and environment,
there are certainly examples of the misapplication of evolutionary theory. Some
evolutionary theorists make claims about behavioral differences among populations in
genetic terms. Two recent papers, one by MacDonald (1999) and the other by Coney and
Mackey (1998), exemplify theoretical approaches that imply that differences in fertility
are to some extent independent of environment. These two papers make the argument
that fertility rates are to some degree heritable, either genetically (MacDonald) or through
some form of bio-cultural ancestry (Coney and Mackey). They both depict a future
scenario in which low fertility societies (primarily European or populations with
European ancestry) will be replaced by high-fertility (immigrant, non-European)
populations from other areas of the world. They assume that future descendents of these
high and low fertility populations will inherit their ancestors’ childbearing strategies and,
hence, their fertility rates. Therefore, their arguments claim, populations with high
fertility rates will eventually replace populations with low fertility rates.
These arguments differ from the majority of evolutionary theorists of human
fertility and subsequently have theoretical deficiencies that other theories do not have.
The major problem with these arguments is that they falsely assume ethnicities and/or
nationalities are biological entities rather than social entities. If these groups are
55
considered as social creations, there is no justification for assuming that the genetic
factors that are influencing fertility on an individual level will be consistent throughout
an ethnic or national population. Assuming that a population’s fertility rate has a simple
correspondence with either its genetic composition or some inherited cultural tendencies,
there is no room for explaining how low fertility populations could have come to exist in
the first place. It is a historical fact that, at one point, the ancestors of these soon-to-bereplaced European populations had higher fertility rates. Any argument that assumes that
contemporary populations with high fertility rates will replace low fertility rate
populations cannot explain why low fertility rate populations evidently replaced high
fertility populations in Europe during the 20th Century.
Although many social scientists are in opposition to evolutionary explanations of
human behavior on the grounds that there is no human nature, other theorists of human
behavior are more favorable towards generalizability. However, many of these social
scientists still resist the use of biological variables in the analysis of human behavior.
One of the reasons for this is because there are theorists who produce work similar to the
above-mentioned studies, which assume meaningful genetic differences among human
social groups. Other reasons include the over-enthusiastic use of evolutionary hypotheses
without evidence that the EEA conditions could have selected for the particular trait
(Potts 1998, p. 94). Along with this criticism is the objection that evolutionary theorists
tend to exaggerate the universality of behavioral traits throughout human populations
(Harris 2001a, p. 126). Besides objecting to the excess zeal displayed by evolutionary
theorists in their attempts to explain human behavior in evolutionary terms, critical social
scientists also point out that these hypotheses are often non-falsifiable (Sahlins 1976).
56
Also, sociobiologists and evolutionary psychologists are criticized for reducing the
evolutionary process to the gene, assuming a one-to-one correspondence between genes
and phenotypes, ignoring the effect of unique histories on the development and behavior
of individual organisms, and using culturally biased metaphors in their analyses (such as
genetic success, competition etc.) (Rose 1998; Rose 1999).
The majority of evolutionary theorists of human behavior remain unconvinced by
the above criticisms primarily because these objections are often exaggerations. Also,
many objections to evolutionary theorizing about human behavior are directed at
intermediate steps in the process of scientific study, such as the development of
hypotheses. Evolutionary theorists often complain that the critics of sociobiology and
evolutionary psychology are really criticizing caricatures of these disciplines (Alcock
1999; Daly and Wilson 1986). They also object to the notion that debunking certain
excessive claims of fringe theories done in the name of sociobiology or evolutionary
theory can be used to falsify the entire core of evolutionary theorizing regarding human
behavior (Canton 1986).
The primary argument of those who defend evolutionary theorizing is that the
critics are not critiquing particular hypotheses. Rather, they are objecting to the scientific
study of human behavior itself. The scientific process depends on the falsification of
hypotheses generated from a theoretical perspective. Despite claims of non-falsifiable
theories, evolutionary scientists have a long history of testing hypotheses generated from
sociobiological and evolutionary psychology perspectives with empirical data (Holcomb
III 1998; Irons and Cronk 2000). Individual hypotheses, no matter how overly ambitious
their claims, are useful only if they are tested. This is not only in regards to evolutionary
57
theories but to the counter theories developed by the critics of evolutionary perspectives
on human behavior. Evolutionary theorists are justifiably annoyed when those who
object to so called biological reductionism dismiss their theories for imperfect
operationalizations and a neglect of the full range of mitigating factors and complexity in
human behavior. Rarely do these critics offer competing theories that can meet these
same standards (Irons and Cronk 2000, p. 9).
Critics of biological reductionism who claim that evolutionary hypotheses are too
simplistic and are not accounting for the full range of factors that determine human
behavior are only stating the obvious. Simplification of all the information in the
universe into a set of theories that we can understand is the objective of scientific
research. The theories that best fit the evidence we have are the preferred theories for the
moment. Simplicity is the objective of scientific theories because it facilitates testing and
understanding. Simplicity is only objectionable if it fails to explain and predict empirical
observations (Charlton 1999). The public nature of science allows theories that better
explain empirical observations to be entered into the scientific discourse and replace
theories that are too simplistic (Kuznar 1997). Therefore, arguments against evolutionary
theories of human behavior because they are reductionistic are being critical of science in
general, because all science is reduction in one form or another. However, this reduction
is necessary for us to better understand the world we live in.
Many social scientists are critical of evolutionary theories of human behavior
because they have a philosophical commitment against any use of biological variables in
the study of human behavior. They are concerned with the use of evolutionary theory in
human behavior because of political and policy implications. This explains why their
58
approach is to criticize the intermediate steps that evolutionary theorists take in the
process of theory building without continuing the process with their own testable
hypotheses. This also explains why researchers who test theories of genetic and
evolutionary influences on human behavior while explicitly citing the corresponding
influence of the environment are often attacked as “genetic determinists” who are looking
for the “immutable biological roots” of human behavior (see (Barash 2002; Kennelly et
al. 2001; Miller and Costello 2001; Risman 2001; Udry 2000; Udry 2001; Weidman
2002). The number of theorists who believe in simplistic genetic determinism is often
greatly exaggerated (Udry 1999). Even the most famous sociobiologist, E.O. Wilson,
takes great pains to refute the possibility that genes can cause behavior independent of
environment (Wilson 1998, p. 136-139). However, for many opponents of evolutionary
theorizing, any incorporation of biological variables into theoretical models of human
behavior is akin to genetic determinism and racism and should be avoided at all costs.
I do not intend to downplay the impact that biological theories of human behavior
have had and could still have on racist thinking and racists policies. However, I do not
think the solution is to combat scientific findings on philosophical grounds alone. If the
scientific study of the biological basis of human behavior is left in the hands of racists
alone, we will be left with only racist theories. True genetic determinism has earned its
negative reputation on the basis of its historical role in fueling genocide and justifications
of the perpetuation of structural inequalities. However, when genes are completely
eliminated as potential causes of human behavior we are left with environmental
determinism. There is no evidence that environmental determinism is any more immune
to being used to justify social inequality or horrible maltreatment of defenseless victims
59
than genetic determinism (see Colapinto (2000) for an example of consequences of
environmental determinism taken to the extreme).
The Use and Misuse of the Word Biology
I think that much of the problem with communicating findings about the genetic
influence on human behavior comes from an inconsistent use of the word “biology.” At
times it is used to signify genes, other times it is used to signify phenotype, and
sometimes it is used ambiguously as one or the other. I think that this inconsistency is a
major stumbling block to the conveyance of ideas about evolution and behavior and the
possibility of meaningful dialogue between theorists on both sides of the debate.
Genes are only one (although very important) aspect of biology. Genes provide the
code that is used to develop a phenotype, but the interpretation of the code is dependent
on environmental conditions. Every aspect of gene function, whether it is building tissue
or building complex central nervous systems that determine behavior, operates through
environmental conditions. Environment cannot be thought of as a static snap shot of
particular conditions at one point in time but must include the entire historical conditions
that have affected gene expression (Rose 1999; Wilson 1998, p. 139). Therefore, since
genes play a limited part in the development of an organism, it is better to use the term
biology to refer to a phenotype, which includes not only genes but also the historical
events that lead to gene expression in a certain way. As I have discussed earlier in this
chapter in relation to diagram B of figure 3-1 (and will continue to discuss in detail in the
next chapter), one of the environmental factors that leads to gene expression is the
cultural environment. With this extended definition of biology, which includes historical
and cultural environmental forces, the use of the phrase “biological determinism” loses
60
much of its divisive force. In this sense, biological determinism includes both genetic
and environmental determinism.
Conclusion
Ultimately, I believe that ideas about the genetic influences on specific human
behaviors and the evolutionary theories that explain these influences will be accepted
only if they explain empirical evidence. The amount of knowledge we have about the
genetic influence on animal behavior will soon dwarf our current knowledge. Social
scientists will be challenged to either explain how this new evidence fits into existing
social theory or to develop new theories that take account of the new data (Udry 1996, p.
326).
In the coming years, emerging data on the connection between genes and behavior
should position human reproduction as a major hotbed of empirical and theoretical
activity in both biological and social sciences. This is not only because of the centrality
of reproduction to theories of evolution, but because of the behavioral plasticity in human
reproductive behavior. Fertility decline in particular is an issue that draws on a
combination of behavior genetics, genetic foundations of motivational states, economic
cost benefits of behavior, the influence of culture on behavior, etc. Tests of theories of
fertility decline offer a unique opportunity to develop an understanding of the
biosocial/biocultural basis of human behavior. However, in order for this to happen,
traditional demographic approaches to fertility decline, will need to be augmented with
more studies of the influences on individual childbearing motivations. Because
evolutionary theory focuses on the interaction between genotypes and environment at the
individual level, aggregate statistics are not sufficient to test evolutionary theories (Bock
1999, p. 200). Tests of evolutionary theories of human fertility require careful
61
measurement of genetic, behavioral, environmental, psychological as well as cultural
variables. The use of evolutionary theory in conjunction with careful, systematic
measurement of cultural variables on the individual level is the area of fertility theory that
would benefit most from an increased anthropological presence. In the following
chapters, I will detail my attempt to integrate anthropological and demographic
perspectives to shed light on the process of global fertility decline.
CHAPTER 4
CULTURE AND REPRODUCTIVE BEHAVIOR
Introduction
In this chapter, I will present an extended definition of culture that can be used in
tests of theories of fertility decline. Before presenting this definition, I will discuss the
emerging sub-discipline of anthropological demography and discuss the potential for
integrating cultural approaches to the study of human demographic behavior. A
consistent theme of the anthropological demography literature is the critique of
demographers by anthropologists for their lack of sophistication in the use of culture.
However, I argue that anthropologists as well as demographers need to present a clearer
definition of what they mean by culture in order to test competing theories. Without
alternatives to outdated definitions of culture, the gulf between demography and
anthropology will not be bridged.
Next, I will present an extended definition of culture that includes a discussion of
the part that culture plays in human evolution as well as a discussion of the properties of
culture that contribute to human beings’ unique behavioral plasticity. I argue that human
culture amounts to a type of virtual environment with which individual human beings
must interact to meet their basic needs. In order to understand individual motivation, this
virtual environment must be accounted for and measured. I discuss how this view of
culture can be used to resolve the debate among fertility decline theorists who emphasize
either culture or economics as the primary force on reproductive behavior.
62
63
In the final section, I present a description of a method for measuring culture. I
present this method while arguing against drawing a distinction between qualitative and
quantitative methods because the measurement of culture requires both qualitative and
quantitative components. I discuss the development of the operationalization and
measurement of culture as consensus by cognitive anthropologists. This
operationalization leads to improvement in construct validity and improved ability to
measure inter- and intra- cultural diversity. I argue that this explicit operationalization
and measurement is necessary to test existing theories of fertility decline. I conclude this
chapter by arguing for a greater use of quantification by anthropologists who are
interested in studying population processes.
Validity and Reliability in Demography and Anthropology
Disciplinary Differences between Anthropology and Demography
A fruitful collaboration between anthropology and demography has been
characterized as improbable because of extreme differences in methodology and
epistemology (Greenhalgh 1997; Hill 1997; Scheper-Hughes 1997). Using
methodological tools such as participant observation, unstructured interviewing, and
extended fieldwork, anthropologists report what they have seen with their own eyes. By
contrast, demographers depend on surveys and quantitative modeling, and report findings
about general trends in populations. The principle difference between anthropology and
demography is therefore in the amount of emphasis they tend to place on validity and
reliability. An examination of these disciplinary cultures (Greenhalgh 1997) with regard
to validity and reliability is the key to recognizing the possibility of a successful
partnership between anthropology and demography.
64
Most anthropological researchers tend to devote themselves to the pursuit of the
collection of highly valid data. Validity in scientific research is defined as the extent to
which what is intended to be measured is actually measured. In social research where
what is measured is typically theory driven, validity refers to the ability of the
measurement to represent the concept being studied. (Carmines 1979, p. 11-13).
Anthropologists who want to measure the attitudes and behaviors of their subjects of
study believe the best way to do this is to conduct in-depth interviews with key
informants and to engage in participant-observation by living with their informants for an
extended period of time. The instruments of data collection in most anthropological
research are the researchers themselves, and they spend years calibrating themselves to
be (as much as possible) objective, culturally un-biased observers and recorders of human
thought and behavior. Anthropologists are critical of methods such as survey research,
which use a limited number of variables, because they do not capture the full experience
of peoples’ lives and can miss important information. Anthropologists point out that
variables used in demographic questionnaires, such as birth, marriage, parenthood, and
household, are socially constructed and can change between settings and populations.
The conceptualization of these variables for the researcher may be different from the
lived reality of the people who are the subjects of study (Townsend 1997). Thus,
anthropologists are critical of survey research because it has potential limitations in data
validity.
Demographers, on the other hand, tend to use data that are highly reliable. The
reason for this stems from their primary research goal, which is to come to conclusions
about the characteristics of large populations. It is often impossible to ask questions of
65
everyone in a population. Instead, demographers elicit information from samples of
people making up a small percentage of the population and use statistics to come to
probabilistic conclusions about the entire population. In order to make statements about a
population on the basis of evidence from a sample, the sample must be drawn using
reliable research methods. Reliability in scientific research is the ability to produce the
same result using the same instrument on the same object of measurement more than
once (Carmines 1979, p. 11-13). With the use of reliable methods such as random
sampling with standardized questionnaires, a sample will be representative of the
population from which it was drawn. This allows for generalizations because reliability
implies that the results from one sample will be the same as those from another sample
drawn from the same population. Standardized questionnaires are reliable because they
use operationalized definitions of variables and are structured to limit the variability in
how questions are asked of each informant. Random samples minimize whatever
systematic bias might exist in the selection of each informant by assuring that each
member of a studied population had equal chance of being selected. Standard
questionnaire design and random samples together minimize the amount of non-random
error in the resulting data, which allows for analysis using statistical techniques based on
assumptions regarding random sources of error. Demographers are therefore critical of
the key informant and participant observation approaches because they have potential for
bias and the introduction of non-random error, thus yielding data of doubtful reliability
and generalizability.
Disciplinary Congruence Between Anthropology and Demography
Although this description of disciplinary cultures has been described similarly
elsewhere (Basu 1998; Greenhalgh 1996; Greenhalgh 1997; Hammel and Friou 1997;
66
Hill 1997) and should be easily recognizable by members of both disciplines, the
dichotomy is an illustrative caricature that obscures disciplinary heterogeneity and selfcriticism evident within both anthropology and demography. There have been many calls
for a more reliable anthropology among anthropologists and, from demographers, a more
valid demography. These internal criticisms deny the need for a choice between validity
and reliability and attempt to maximize both of the two elements to form a more
complete research agenda.
The movement within anthropology towards more reliable data collection centers
on the call for the discovery of basic laws that describe and predict human thought and
behavior (Bernard 1979, p. 33; Murphy and Margolis 1995, p. 1). Discovery of these
laws requires research that produces generalizable and testable findings (Harris 2001a, p.
16-17; Kuznar 1997, p. 22-24). Studies that are generalizable attempt to link information
gleaned from the study of a single group with what is known about human beings in
general. Moreover, a study is testable (i.e. falsifiable) if it can be repeated with the
possibility of discovering that its hypotheses are not true. Anthropologists who are
proponents of the discovery of these laws have been critical of other anthropologists who
ignore the potential for generalization from social research and pay too much attention to
the particularistic, “story-telling” aspect of ethnographic research (D'Andrade 1995b, p.
405; Harris 2001b). Many anthropologists have specifically called for greater use of
survey methods and quantification as a way of informing ethnographic research
(Greenhalgh 1990, 101; Hammel and Friou 1997, p. 181; Lewis 1950). The
anthropologists who call for quantification, survey methods, or studies based on
uncovering basic laws of human behavior are proposing that the research findings
67
produced by anthropologists would become stronger and more relevant if they paid
attention to not only validity, but also reliability.
Conversely, some demographers have levied criticisms against their colleagues for
a lack of attention to questions concerning validity. These criticisms focus mainly on the
acknowledgement of culture as a force behind demographic behavior and change, and the
need to incorporate ethnographic fieldwork and focus groups into the methodological
toolkit of demography. In the Princeton European Fertility study, demographers realized
that their standard instruments lacked adequate validity due to the inattention given to
cultural context as a force behind demographic change, (Bongaarts and Watkins 1996;
Kertzer 1995, p. 31-33; Pollak and Watkins 1993; Watkins 1986). Many called for
demography to expand its standard set of variables to include cultural differences. In
addition, John Caldwell (e.g. 1982) and other demographers have recognized the need to
engage anthropological methods of field research as a complement to their knowledge of
population processes based on surveys. Many demographers now use ethnographic and
other qualitative field methods that produce more valid data as a means of
complementing reliable survey research.
These internal movements away from the extremes of sacrificing validity for
reliability or vice versa suggest the possibility of a convergence of the strengths of the
two disciplines towards an anthropological demography. The main similarity between
these disciplinary self-criticisms is the recognition that there need not be a choice
between validity and reliability in scientific research. This recognition is an indication
that there can be mutually profitable collaboration between anthropologists and
demographers to produce original findings that are both valid and reliable.
68
The Uses of Culture in Demography and Anthropology
Anthropological Critiques of Demographic Uses of Culture
Although there is now a greater acceptance of the effect of culture on demographic
processes, demographers have tended to use culture in a way that is not satisfactory to
most anthropologists. Several anthropologists, along with demographers, economists and
historians, have criticized the use of culture by demographers in several recent
publications (Basu and Aaby 1998; Fricke 1997a; Greenhalgh 1995b; Hammel 1990;
Handwerker 1986a; Kertzer and Fricke 1997; Pollak and Watkins 1993). The overall
criticism of the use of culture in demographic studies is that demographers are not
actually using a theory of culture.
Demographers who worked on the European fertility project found weak and
inconsistent results while searching for the economic variables that would predict fertility
decline. Many of them concluded that culture explained much more than economic
forces (Cleland 1987; Dyson 1983). The variables demographers used to identify culture
consisted of social organizational groupings such as common location, language,
ethnicity, religion, etc. Their models of fertility decline generally improved once they
controlled for culture and they noticed that underdeveloped areas adjacent to urban areas
which supposedly shared the same culture often experienced fertility decline without
experiencing the same economic changes (Watkins 1986). The assumption was that
people who spoke the same language, lived in the same area, had the same religion or
ethnicity had homogenous views about things such as contraception and would hold these
similar views regardless of their economic status. Taking this approach, demographers
argued that future research on fertility decline would need to take into account the
diffusion of ideas as the main force behind fertility decline and supplement their survey
69
research methods with the more qualitative methodologies of anthropology (Greenhalgh
1995a, p. 9).
The problem with this approach is that demographers generally failed to offer any
explanation of what they meant by culture. Anthropologists and others in the social
sciences have criticized this use of culture because it is being used as an a-theoretical
“black box” in which the variance in fertility that is not explained by economic forces is
explained by “everything else” (Handwerker 1986b, p. 10). There are many problems
from the perspective of anthropology with this use of culture. First, there is no
connection between the variables that are labeled as cultural to a theoretical conception of
culture in general (Kertzer 1995). Also, the assumption that static social groupings
necessarily imply homogeneity of ideas is a view of culture that is outdated (Hammel
1990). Demographers’ use of culture in this way reveals their ignorance of more recent
anthropological developments in culture theory (Greenhalgh 1995a, p. 9). A main target
of these debates is the criticism of the ethnocentric assumption that societies undergoing
fertility decline are moving from traditional (i.e. irrational) to modern (i.e. rational)
cultures (Greenhalgh 1995a, p. 16-17). This assumption, especially evident in family
planning programs, often denies human agency in fertility decisions for so-called
traditional people and implies passive acceptance of cultural norms (Greenhalgh 1995a,
p. 25).
Another problem with this use of culture is the assumption that culture is something
that stands outside of other domains of social life and is not itself subject to economic
forces (Kertzer 1995). Anthropologists have argued that culture is more than ideas that
are transferred by social interaction. Rather, economic, political, and other social factors
70
are a part of culture, not distinct from it (Handwerker 1986b, p. 11). Demographers who
view culture as an individually owned independent variable that affects fertility behavior
or limit their analytic domain to couples, kin groups or networks are ignoring the context
of macro-level processes of culture shaped by unique histories (Greenhalgh 1995a, p. 9).
The use of qualitative or so-called anthropological methods in demographic
research is also a source of criticism because of the lack of connection between these
methods and a theory of culture. Used in this way, critics argue that qualitative methods
end up being quick technical fixes for the inadequacies of surveys in measuring cultural
variables (Fricke 1997b, p. 826). Also, they argue that the assumption that qualitative
methods such as focus groups and open-ended questions can be used to inform
quantitative models ignores the epistemological differences between anthropology and
demography (Greenhalgh 1997; Hill 1997; Scheper-Hughes 1997). The piecemeal
incorporation of methods used by anthropologists for completely different
epistemological purposes shows that there has not been an adequate accounting of culture
(Greenhalgh 1997). This epistemological difference is revealed in the lack of acceptance
by demographers of other anthropological methods such as community studies and
narrative modes of explanation (Fricke 1997b, p. 827; Greenhalgh 1995a, p. 12).
Differences in epistemology have influenced some critics to distinguish between
anthropological and qualitative methods (Greenhalgh 1997; Knodel 1997). The
implication in this distinction is that demographers are not any closer to incorporating a
sophisticated view of culture by simply using qualitative methods. Instead, they are
using different tools for the same purposes.
71
Anthropological Conceptualizations of Culture
An anthropological critique of demography's use of culture notwithstanding, there
continues to be internal debates within anthropology regarding the meaning of culture.
The main thrust of the arguments in these debates is at the most basic level: what is
culture and is it of any use? (See the debates in the recent special supplement of Current
Anthropology entitled, “Culture–A Second Chance?” (Aunger 1999; Brumann 1999;
Romney 1999) and in the recent edited volume Assessing Cultural Anthropology
(Borofsky 1994)). The forcefulness of these debates and the seemingly insurmountable
epistemological gulfs between the participants raises the question of how demography
would be able to incorporate an anthropologically informed definition of culture if
anthropologists themselves cannot even agree on a definition.
In this section, I will argue that a conception of culture that is rooted in cognitive
psychology and explicitly describes the interaction between ideas and behavior is the best
direction for a demographically useful definition of culture. This conception of culture
should include an understanding of the neurobiology of human brains and the
evolutionary forces that shaped this neurobiology. This approach to culture allows for
specific operationalizations of culture, and the possibility of direct measurement.
Operationalization of culture is essential to a better understanding of its effect on
reproductive behavior. Without operationalization, reference to culture as an entity that
can have independent effect on the world is meaningless. Without an explicit definition
of how culture affects human behavior, theorists often resort to circular reasoning by
assuming that “culture is everything” and can explain behavior without itself needing to
be explained (D'Andrade 1999). Differences and similarities in patterns of thought and
behavior that are explained by reference to different cultures is a covert tautology
72
(Handwerker 1986b, p. 11-12). This line of explanation offers nothing more informative
than the assertion that differences exist because people are different.
Another problem is that researchers who use an unoperationalized definition of
culture will never know if they are discussing the same thing. This eliminates the
possibility for constructive debate and falsification. Use of unoperationalized concepts in
research allows for theoretical inconsistency, in particular when theoretical concepts are
used across disciplines. Measurement of culture allows for testing and comparison of
competing theories that attempt to explain the influence of cultural processes on
demographic behavior. This is an essential element in the development of an
interdisciplinary theory of fertility decline.
Cognitive Theory of Culture
Although there is disagreement among theorists about many aspects of culture,
there is agreement that culture refers to something learned rather than inherited (Brumann
1999). The concept of culture is usually invoked to understand the behavior and thought
patterns of groups. However, only individuals can learn and they are the only source of
cultural data (Handwerker 2001, p. 10). Therefore, any definition of culture must begin
with knowledge that human beings possess and how individual human beings learn and
processes information. Since culture is learned primarily through other people, it is also
the result of social interaction and is shared. This results in culture being something that
is both socially and individually constructed.
The way individuals construct cultures begins with the formation of cognitive
models of reality. Cognitive psychologists discovered a limit to humans’ ability to
remember discrete bits of information through experimental research (Miller 1956).
However, humans have an almost unlimited ability to “chunk” together bits of
73
information into schematized models of particular domains of information (D'Andrade
1995a). Cognitive science has found that there is not a one-to-one correspondence
between a particular model and a particular domain. Rather, multiple models are at work
in concert at any given time. Some models are more likely than others to be invoked at a
given moment because of a weighting process that develops over time after repeated
experiences with a domain (Strauss and Quinn 1997). This weighting process is
mediated by emotions that are evoked during these experiences. Models invoked during
experiences that are associated with positive emotional feedback are more likely to be
used in the future. The opposite can be said for negative emotional feedback (Strauss and
Quinn 1997). Cognitive anthropologists define the complete set of an individual’s
cognitive models, including the models’ associated emotional weights and behaviors, as
the raw material of culture (Handwerker 2001).
One common assumption about culture (now proved incorrect) is that culture is a
unitary, internally consistent “seamless web” that contains unambiguous rules for
behavior (DiMaggio 1997, p. 267). Instead, cognitive researchers see culture as
fragmented and inconsistent. At any point in time, individuals may have internalized
cognitive models that are contradictory. These models, while guides for behavior, can
never have a one-to-one correspondence with behavior outputs because of their
heterogeneity. Rather than acting as a blueprint for behavior, culture acts like a “toolkit”
of strategies which individuals use to choose among behavioral options depending on
momentary external circumstances (DiMaggio 1997).
The cognitive models that individuals have at their disposal at any point in time
develop as a result of past experiences and are constantly being modified with new
74
experiences. Cognitive models influence the behavioral choices that individuals are
forced to make in the context of external circumstance. These behavioral choices then
provide individuals with additional experiential information from which ideas and
emotions are subsequently generated and modified. Because no two individuals have
exactly the same experiences, no two individuals have the same set of cognitive models.
And no one person has the same set of cognitive models from one moment to the next
because individuals are constantly behaving and processing additional experiential
information (Handwerker 2001, p. 8).
Although individuals are the only source of cultural data and the raw materials of
culture pertain to individuals, culture is created through social interaction. Cultural
models refer to models that are to some extent shared by members of a population
(Dressler and Bindon 2000, p. 246). However, since cognitive models are the result of a
creative process within individual brains, culture is not a “thing” that can be transferred
from one person to another (Handwerker 1989). Because an individual’s set of cognitive
models is the end product of life experience and because members of populations often
have similar, if not identical, experiences, this produces patterning of cognitive,
emotional and behavioral traits. Also, as individuals interact with members of their
social networks, they experience the world vicariously through other network members.
This enables individuals to hold ideas and emotions about experiences and behaviors
without actually experiencing them directly. Therefore, although the raw materials of
culture can be accurately characterized as residing within individual human beings, the
cognitive models and emotions of each individual human being depend in some part on
the cognitive models, emotions and behaviors of other members of their social networks.
75
Individual human beings do not passively accept models from their social network.
Rather, they accept the models that work, modify those that do not, and “share” these
modifications back into their social network in a dynamic, continually evolving creative
process. When models developed through previous experience are unable to account
adequately for new stimuli, individuals switch from “automatic” to “deliberative”
cognition which they use to actively and innovatively restructure their own models to
better account for new stimuli (DiMaggio 1997).
Subsequent interaction with a social network leads to the spread of the innovation
throughout the network if the innovation is successful at resolving similar inadequacies in
the models held by other network members (Tomasello 1999). The “spread” of
innovations throughout a network is actually individual brains making similar cognitive
adjustments after interactions with members of their social networks. Thus, culture can
be shared, but only metaphorically and imperfectly (Handwerker 1989).
Culture as Evolutionary Adaptation
The ability of human beings to pool their cognitive resources represents an
evolutionary adaptation that helps individual human beings better meet their biological
and psychological needs (Tomasello 1999). Culture gives human beings the ability to
address practical problems they face on a daily basis in their efforts to survive, feed
themselves, and reproduce. Compared with other species that interact socially, human
beings are uniquely adapted to pool their cognitive resources not only over populations
but also through historical time (Tomasello 1999). This creates what is called
“distributed cognition” which acts as a cognitive division of labor among the members of
a social network (DiMaggio 1997). What’s more, human beings are uniquely able to
pool these resources among a socially cooperating group of strangers or partial strangers,
76
unlike other socially interacting species which almost exclusively interact with kin
(Chase 1999). Although there is sufficient evidence that non-human animals also use
culture for survival, the uniqueness of human culture lies in the degree to which it
pervades all aspects of human life (de Wall 2001, p. 29; Harris 2001a, p. 125)
In order to understand this unique evolution of culture in the human species, we
need to consider the environment under which human beings evolved and developed this
trait. We have limited information about the conditions of the environment of
evolutionary adaptation (EEA) that served as the backdrop for human evolution during
the Pleistocene (Bock 1999, p. 197). The evidence we have points to an incredibly
unstable environment which would not have had consistent enough selective pressure for
specific behavioral traits to evolve (Potts 1998). Instead of the evolution of specific
behavioral responses to specific environmental conditions, the EEA selected for plasticity
in behavioral responses to changing environmental conditions. Human capacity for
culture developed under these conditions and has been such a successful behavioral
strategy that human beings now inhabit almost every ecological niche on earth.
Culture creates a buffer between people and their physical world. It is a type of
virtual environment with which people interact. Behavioral responses that are directly
tied to physical environmental cues and not through culture need to be developed over
long periods of time under conditions of systematic selection of genetic variations. The
variations that end up being widespread are those that allow individuals a greater
probability of success at securing resources from the environment, which leads to greater
probability of success at survival and reproduction. For human beings, culture fulfills a
similar role as genetic evolution. Through simplified models of reality that filter out
77
information and constrain behavior within a limited set of options, culture provides
individuals with strategies for how to live their lives in a way that assures consistent
access to resources in order to survive and reproduce within a given environmental
circumstance.
In contrast with genetic evolution, cultural evolution is much more dynamic and
can lead to more rapid behavioral change. If cultural strategies for the fulfillment of
biological and psychological needs are not successful for individuals, or if individuals
develop innovations that improve on previous strategies, these strategies will spread
throughout social networks. Cultural innovations that allow for greater and more reliable
access to resources are more likely to be selected over existing cultural models
(Handwerker 1989). An evaluative feedback loop enables the culturally shared models to
be constantly reevaluated and modified or replaced if necessary. Unlike genetic
evolution, cultural evolution happens without the need for new generations to be born.
Although many cultural traditions continue seemingly unmodified, generation after
generation, significant cultural evolution can occur within one generation.
Because human beings are a species that depends on culture to access resources
from the environment, culture becomes an important environmental element with which
humans must interact in order to survive and reproduce. As cultural beings, humans must
interact with other members of their species in order to access the resources that allow
them to meet their biological and psychological needs. Human beings have both a
cooperative and competitive relationship with other members of their species. In order to
survive, humans must be able to understand and predict the behaviors of other human
78
beings and must learn how they themselves must behave in order to access resources
within a particular cultural milieu.
Learning about appropriate behavior within a cultural setting is not limited to
simple observation and imitation. Rather, humans must be able to develop an
understanding of the intentionality behind human behavior, which means that they must
understand the cultural environment within which others are operating (Tomasello 1999).
Therefore, in order to access resources which may allow an individual human being to
survive and reproduce, human beings must not only successfully interact with their
physical environment but they must interact with the cultural environment being used by
other human beings in their social networks.
Economic and Cultural Explanations of Reproductive Behavior
This extended definition of culture helps to clarify the roles of culture and
economic forces on fertility decline. Much of the current debate about the influences on
fertility behavior centers on the competition between cultural factors and economic
factors (e.g. (Hammel 1995; Kertzer 1995; Pollak and Watkins 1993)). The inclusion of
both behavioral and ideological components into a definition of culture helps to resolve
some of these debates by considering economies as part of culture. Perhaps the most
extensive operationalization of culture as ideas and behaviors comes from Marvin Harris’
cultural materialism research strategy (Harris 2001a). In a three-tiered conceptualization,
cultural materialism divides culture into mode of production and reproduction
(infrastructure), social and political organization (structure), and idealistic systems of
meaning (superstructure). All three components have the ability to influence the others
but the infrastructure is said to have probabilistic influence on structure and
superstructure. Likewise, structure has probabilistic influence on superstructure. This
79
inclusive division of culture into both mental and behavioral aspects with each
influencing each other avoids the tautology of which D’Andrade (1999) and Handwerker
(1986b) warn.
Considering the conceptualization of culture as inclusive of both ideas and
behaviors, the debates about the primacy of cultural or economic factors having the most
influence on fertility decline are rendered meaningless through their false distinction
between culture and economics. This is especially true for arguments that put social
interaction at the center of the process of fertility decline and over-emphasize the causal
force of networks and diffusion in the fertility decline process. For example, one
sociological line of reasoning attempts to explain the incidence of fertility decline in
under-developed areas that are adjacent to developed areas as a result of the diffusion of
information about novel contraceptive behavior (Bongaarts and Watkins 1996; Pollak and
Watkins 1993). This model accounts for the spread of fertility limiting behavior through
the interaction between individuals and “opinion leaders” or “trendsetters” (Pollak and
Watkins 1993, p. 479). What is assumed is that the behavior starts with one person who
innovates and then interacts with others who begin to adopt the innovation. Thus,
networks are important in understanding fertility decline because they are the means with
which new information is spread which leads to behavior change.
What is left open in this analysis is an explanation of why the innovative behavior
was adopted in the first place and why some adopt the behavior and others either lag
behind or never adopt the behavior. It also ignores the possibility of extensive
simultaneous invention of novel behaviors by individuals who have never interacted with
each other. Thus, it is more of a description of the process than an explanation.
80
The biggest problem with the social interaction explanation of fertility decline is
the neglect of the interdependence of economic processes and social interaction, causing
a problem with teasing out one’s influence from the other. This is especially true for
information about novel economic opportunities. This type of information has been
found to be more likely to spread through network connections among those network
alters who are loosely tied to one another (Granovetter 1973). Given the equation of
culture with systems of ideas and behaviors that exist in the intersection of individual
network members, how can economic forces and social networks be distinguished?
When novel economic opportunities, or even the perception of novel economic
opportunities, depend on social networks for their spread, how can changing economic
forces be separate from social interaction? The functional limitation of networks as a
conduit for attitudes about novel contraceptive information is the weakest point of the
social interaction model of fertility decline. The broadening of the conception of culture
to include behavior (including economic behavior) and structures of social interaction is
the solution to this weakness.
A Method of Culture for Demography and Anthropology
The Measurement of Culture: Consensus Analysis
The theory of culture presented above is in terms of conceptual variables (Bernard
1994b, p. 25-32). Culture used as a conceptual variable is a theory, which cannot be
tested directly. Although theories cannot be tested directly, they imply hypotheses,
which are predictions of what a researcher expects to find given the veracity of the
underlying, conceptual theory. Hypotheses, unlike theories, can be tested but only when
the conceptual variables from the theory have been given operational definitions.
Operational definitions of culture, along with hypotheses generated from the conceptual
81
definition of culture, can provide the basis for empirically evaluating culture’s influence
on fertility decline.
If cognitive models are shared among members of a group implying a cultural
pattern of thought and behavior, this sharing needs to be empirically demonstrated in
order to be used in a test of a hypothesis. Members of social groups often have similar
cognitive models because of similar life experience and tend to agree with one another
about the meaning of certain domains of knowledge and behavior. However, within
social groups there are sub-groups of members who have slightly or not so slightly
different life experiences, which causes disagreement. Age, gender, and socio-economic
status are examples of characteristics that can influence members of the same language
group, religion or geographic area to have varying cultural models of particular domains.
Sometimes this varying life experience results in intra-cultural variation in which
there is variation within an overall level of agreement. Sometimes it may result in intercultural variation in which there are groups within the population that have high internal
levels of agreement but do not agree significantly with other groups within the same
social group. And sometimes there may be very low levels of agreement among all
members of the population (Handwerker 2001, p. 21). Thus, in order to validate claims
of agreement on cognitive models enough to classify them as cultural, levels of
agreement must be measured and cultures must be demonstrated empirically. Since the
mid-1980’s, cognitive anthropologists have developed ethnographic techniques to collect
field data which can be analyzed statistically to demonstrate this agreement empirically
and can be used to test the internal validity of theoretical claims about culture.
82
The methodology of consensus analysis began as an effort to limit the effect of an
ethnographer’s personal biases on the final product of ethnography and to determine an
objective measure of informants’ cultural knowledge (Romney et al. 1986, p. 314).
Noting well known controversies over the conclusions of famous ethnographies as well
as the tendency of informants to disagree with each other, Romney, Weller and
Batchelder developed a formal method that would determine which informants were
more culturally knowledgeable than others. Experiments showed that informants who
knew the correct responses to questions tended to agree with each other more than
informants who didn’t know the answers (Boster 1985). Noting that, mathematically, the
reverse of this–that those who agreed with each other would be more likely to have
knowledge about some unspecified truth–would also be true, Romney, Weller and
Batchelder’s formal model provides a means of determining who knows more than whom
as well as the answers to the questions themselves.
Consensus analysis uses the multivariate technique of factor analysis to develop
these measures. Instead of using factor analysis in the way that it is usually intended– to
reduce the numbers of variables for a group of informants by analyzing the variance in
values of these variables (Neter 1996)–consensus analysis reverses the data matrix. This
allows for the analysis of variance in an informant by informant agreement matrix to
determine the variance in agreement among informants [Romney, 1986 #133, p.
322;Weller, 1988 #142, p. 75]. Cultural consensus analysis can be used when certain
assumptions are met, including the assumption that there is a specific correct answer for
each question asked of an informant, each set of answers given by an informant is
independent of other informants answers, and the questions are drawn from one domain
83
of knowledge which assumes one culture (Handwerker 1998, p. 570; Romney 1999, p.
S107).
Handwerker has developed a modification of the consensus analysis model in order
to measure cultural differences when the above assumptions are not met. His
modification is to use principal components analysis instead of factor analysis when
analyzing the intersection of agreement among informants (Handwerker 2002, p. 112;
2001, p. 185). Principal components analysis and factor analysis are very similar
procedures. They both linearly transform the original variables into a smaller set of
uncorrelated variables called factors. Principal components analysis differs from factor
analysis in that it uses the maximum variance rather than the common variance
(Bartholomew et al. 2002, p. 167-168; Dunteman 1989: 7-9). Using the maximum
variance is more important in cases where an investigator wishes to examine the total
variance of the agreement among informants in order to determine if there exists one,
multiple, or no cultures of agreement on a particular domain. The factor analysis method
used by consensus analysis is more appropriate when the goal is to determine cultural
competence on a single pre-defined culture. Instead of looking for the correct answers,
Handwerker’s method looks for the possibility of sub-cultural agreement on different sets
of “correct” answers.
In order to determine sub-groupings of agreement, Handwerker proposes using the
factor loadings on the first two unrotated factors of a principal components analysis in a
cluster analysis (Dunteman 1989, p. 78-79; Handwerker 2002, p. 115; Handwerker 2001,
p. 88). A cluster analysis can be performed on the loadings of the first two factors if they
account for a substantial amount of the total variance. If there is one culture whereby
84
each informant is linked to one another through a shared agreement on a domain, the first
factor will represent at least 50% of the variance, the difference between the eigenvalues
of the two factors will be large (the first factor is 3 times the second) and there will be
negligible negative loading on the first factor (Handwerker 2001, p. 186). If there is not
one culture, plotting the factor loadings for the informants on the two factors against each
other will show patterns of agreement. These patterns can be used as clues for finding
multiple sub-cultures as well as multiple cultures within a social group.
This method provides a means of empirically justifying the claim that a group of
people belongs to a particular culture. In other words, it provides an empirical means of
assessing construct validity (Handwerker 2002; Handwerker 2001, pp. 183-217). If a
group of informants is selected based on their ethnicity or national origin and there is a
particular domain of knowledge, attitude, or behavior that is the focus of study, this
method provides a means of determining empirically if in fact there is one culture for this
group on this domain. If there is not a big difference between the first two factors, this
method will help determine what kinds of culturally relevant groups there are. Future
groupings of informants may reveal that other life experiences (such as age or gender) are
more important factors in determining cultural groups for the theoretical domain in
question. Thus, once an operational definition is given to a culturally based construct
variable, empirical evidence can determine how well this construct holds together. This
method can be used to evaluate the consistency of findings with future research that uses
the same operational definitions of theoretical concepts.
Factor analytic techniques are important substitutes for classical statistical methods
because of several properties of cultural data. First, cultural data is by nature
85
multidimentional (Handwerker 2001, p. 190). The cognitive models that form the basis
for cultural models contain a multitude of elements, their relationship with each other,
and how all the elements function for a particular domain (D'Andrade 1995a). In order to
understand the whole of a cultural model, measurement must be made on the individual
elements that make up the model. This requires analysis of answers given to structured
questions from survey questionnaires, which form the basis for comparison of agreement
among informants (Weller 1998). The responses to the questions, or the “verbal
production”, are not the cultural model. Rather, consistency in answering the questions
gives evidence that the informants are all drawing from the same shared cultural model in
order to produce their responses (Dressler and Bindon 2000: 247).
Qualitative Methods, Traditional Statistics, and Multivariate Statistical Methods
Qualitative methods should be used to develop a theory of the cultural model and
the questions that will be used to measure aspects of this model. Qualitative research
strategies such as participant observation, case studies, or repeated unstructured
interviewing aimed at eliciting cognitive domains fulfill a central role in the ethnographic
process. But in order to determine the amount informants differ from each other on
complex variables with subtle differences, ethnographers often need more precision than
qualitative methods alone can provide. Developing quantitative measurements of
concepts developed qualitatively allows for precise comparisons of variables and testing
of theories using multivariate statistical techniques. Qualitative interpretation of data
(either quantitative or qualitative) implies quantitative comparisons. Operationalization
of conceptual variables and analysis using multivariate statistical techniques makes these
comparisons explicit (Handwerker 2001, pp. 11-21)
86
Multivariate statistical methods are also more powerful than traditional statistical
modeling techniques, such as multiple regression, because the nature of cultural data
implies that responses given by one informant are by definition spatially and temporally
autocorrelated with the answers given by other informants (Handwerker and Wozniak
1997). Although this is considered a problem needing to be corrected in regression
analysis, the goal of statistical analyses of cultural data is to understand, define and
measure this correlation [Handwerker, 2001 #185, p. 19]. Because multivariate statistical
techniques can be used to analyze variance in informants rather than variables, they
provide the means to measure this autocorrelation.
The socially constructed nature of cultural data also causes another problem with
classical statistical methods due to sample size requirements. In order to determine subtle
cultural differences that may lead to cultural boundaries, the qualitative techniques that
allow researchers to uncover cultural meanings are often too taxing and labor intensive to
undertake on large sample sizes. Multivariate techniques such as factor analysis and
principal components analysis can determine if data sets have structure or random
distribution using much smaller sample sizes than traditional demographic or
psychometric statistical techniques.
In order to determine external validity (the ability to generalize from the analysis of
results on one sample) using structured interview questions allows for greater
comparability than traditional qualitative ethnographic techniques used in isolation.
Instruments developed for a community study can be exported and used in other
communities in order to test hypotheses. Structured interviewing also allows for
incorporation of standard large-scale demographic instrument questions making local
87
data comparable with national data as well as development of national surveys from
proven instruments produced at the community level. Conversely, standardized questions
developed in local settings can be used in subsequent national or regional large-sample
surveys to test the prevalence of local findings in multiple settings.
This method has important ramifications for the development of an anthropological
demography. Once an operational definition is given to a culturally based construct
variable, empirical evidence can determine how well this construct holds together. First,
this method can be used to justify the claim that a theoretical construct has construct
validity for one population. This method can then be used to evaluate the consistency of
findings with future research that uses the same operational definitions of theoretical
concepts for other populations. Thus, this method can be the key to linking traditional
anthropological local studies with traditional demographic studies of large national or
regional populations.
Debates of Qualitative and Quantitative Methods in Anthropology and Demography
How does this method fit in with the debate on uses of qualitative methodology to
answer questions about demographic processes? At a recent meeting of the Population
Association of America, several presenters debated the merits of using qualitative
methods on population issues (Fricke 1997b; Greenhalgh 1997; Kertzer 1997a; Knodel
1997; Obermeyer 1997; Rao 1997). Two of the participants, Susan Greenhalgh and John
Knodel, argued that using so-called anthropological qualitative methods was
incompatible with demographic research. Knodel argued that qualitative methods were
useful in demographic research in a variety of capacities but so-called anthropological
methods were not. His characterization of anthropological methods included
ethnography and full contextualization of variables (Knodel 1997, pp. 848-849). He
88
advocated limiting the use of qualitative methods to non-anthropological qualitative
methods, such as focus groups and focused, in-depth interviews because there is no way
that demographers can be expected to turn themselves into anthropologists. Greenhalgh
argued that when demographers use qualitative methods that originate in anthropology,
they are altered significantly so that they are no longer “anthropological.” She states that
this is mostly because of the difference in disciplinary cultures in the objectives of
research. Greenhalgh characterizes demographers as supporting the status quo of
institutions and normal scientific practice while anthropologists are more concerned with
the reflexivity of their methods and the problems with representation of the “other”
(Greenhalgh 1997, p. 820-821). Thus, when a qualitative method is used by
demographers to augment their quantitative models, they are using the method differently
than originally intended. This renders it non-anthropological.
These arguments are not convincing in showing that there needs to be a disciplinary
division of various methods because they ignore distinctions in the three different
meanings of the word method: epistemological, strategic approach, and technique
(Bernard 1994a). The above arguments by Knodel and Greenhalgh are only in terms of
epistemological differences in methodologies. Epistemological difference is the only
reason for characterizing one method as either anthropological or non-anthropological.
The implication is that demographers will accept quantitative models as evidence while
anthropologists will not and anthropologists will accept narratives of participant
observation experience as evidence while demographers will not. However, as I argued
in the beginning of this chapter, there are certainly epistemological differences within
anthropology, so there should not be a characterization of a method as anthropological
89
based on epistemology alone. Many anthropologists share the dedication to science that
Greenhalgh characterizes as demographic.
Method as strategic approach or technique need not belong to a particular discipline
and need not even belong to a particular epistemological approach (Bernard 1994a, p.
176). There is no need to argue that a particular methodological approach belongs to one
discipline or another. As long as they are useful in answering questions that are asked by
researchers, methods should be used regardless of their traditional placement in one
discipline or another. I think that the questions being asked by anthropologists and
demographers about demographic processes, such as fertility decline, are often the same
questions. This makes it imperative that they work together, using whatever
methodological strategies and techniques are available, to come to agreement on the
answers to these questions.
The biggest obstacle to a greater understanding of the potential for anthropological
demography may be the neglect of the compartmentalization of qualitative and
quantitative methods into data and analysis components (Bernard 1996). A research
method can have either qualitative or quantitative data and either qualitative or
quantitative analysis. There can be a combination of quantitative data and quantitative
analysis, qualitative data and quantitative analysis, etc. When anthropologists and
demographers have debated the inclusion of qualitative methods in demographic
research, the specification of qualitative data or qualitative analysis is generally ignored.
This is probably because it is assumed that qualitative data cannot be analyzed with
anything other than qualitative methods, which is not true.
90
Any form of qualitative data, if collected systematically, can be analyzed with
quantitative methods. Content analysis can be performed on all qualitative data, such as
open-ended answers to survey questions, transcribed interviews, television programs,
historical policy documents, family histories, newspaper articles, photographs, etc.
(Bernard and Ryan 1998; Krippendorff 1980). Many methods in the cognitive
anthropological tradition involve systematic collection of qualitative data that can be
analyzed quantitatively (D'Andrade 1995a; Weller and Romney 1988). Systematic
collection of qualitative data allows the use of whatever statistical approach is appropriate
to answer theoretical questions about populations. The method using principal
components analysis discussed earlier is especially well suited to be used with qualitative
data once texts are given codes that are chosen to measure an operational definition of
culture. Quantitative methods of analysis are simply precise ways to systematically
analyze complicated data by looking for associations and patterns among variables or
informants, which may be difficult to see using strictly qualitative analysis. Used in
conjunction with qualitative analysis, quantitative analysis of qualitative data is a
powerful tool in the study of populations. Both anthropologist and demographers should
welcome the results of these types of analyses.
The Use of Quantification in Anthropological Demography
How realistic is it to expect that anthropologists will incorporate more quantitative
analyses into their research agendas? Judging by the tenor of some of the recent debates
on anthropological demography, it remains to be seen if the acceptance of quantitative
analyses of qualitative data (let alone quantitative analyses of quantitative data) will ever
rise above a small minority in anthropology. While there are some calls for a blending of
survey data with qualitative research methods as a quintessential anthropological research
91
strategy (Hammel and Friou 1997, p. 181; McNicoll 1980, p. 457), arguments for an
anthropological demography generally call for a greater use of qualitative methods in
demography but say nothing about greater use of quantitative methods in anthropology.
Quantification is characterized as a handicap (Greenhalgh 1996; Hill 1997; ScheperHughes 1997) and a “thickening” of demographic analyses and conceptualization of
variables is encouraged (Fricke 1997a; Townsend 1997). When the discussion turns to
anthropological use of quantitative methods, the tone is of resignation to the reality that
anthropologists “distrust” (Kertzer 1997b, p. 19) and are “hostile” towards quantification
(Hammel and Friou 1997, p. 178), that they have an “anti-quantification bias”
(Greenhalgh 1990, p. 100), or that “they celebrate their innumeracy” (Hammel and Friou
1997, p. 177).
Some of this resistance of quantification may be due to epistemological concerns,
but a proportion might well be due to the disciplinary culture of anthropology that
typically does not require its graduate students to have any training in quantitative
methodology (Bernard 1994a, p. 171). Clues that the anthropological resistance of
quantification may be based more on unfamiliarity rather than informed rejection is often
revealed in the joking asides of the participants of debates on anthropological
demography. While arguing for a “demography without numbers”, Scheper-Hughes
jokes that the extent of her ability to learn quantitative methods was limited to putting
page numbers on her research reports (Scheper-Hughes 1997, p. 202). When advising
demographers on the possibility of collaboration with ethnographers, Gene Hammel
characterizes this as a practical solution to the problem of integrating of qualitative and
92
quantitative methods, “if the ethnographer can be taught to count” (Hammel 1990, p.
472).
The unfamiliarity with quantitative methods has far more fundamental effects than
simply the reluctance of anthropologists to analyze their data quantitatively or the
difficulty anthropologists may have in understanding the quantitative work of
demographers. The most important effect is the reluctance of anthropologists to
operationalize their variables, especially culture. Quantification provides the ultimate
incentive for researchers to turn theoretical concepts into operational variables. Without
quantification, there is little incentive to reduce complex theoretical conceptualizations
into more limited operational definitions. However, empirical measurement of
operational definitions is at the heart of testing for the construct validity of theoretical
variables. Without empirical measurement there is no way to assure the construct
validity of the theoretical construction. Also, operational definitions allow researchers to
talk to each other in the same language and allow for testing and falsification of
theoretical concepts (Bernard 1994b, p. 32). Without this dedication to quantification,
anthropological theoretical concepts lack this common operational language. This can
lead to inconsistent uses of these concepts among researchers. Without
operationalization, no two researchers can be certain that they are talking about the same
thing when they debate theoretical concepts.
Anthropologists sometimes appear to be unaware that the possibility of
operationalization and quantification of culture exists. In a recent issue of Current
Anthropology dedicated to debating the usefulness of the concept of culture (Vol. 40,
Supplement, February 1999), Christoph Brumann argues for the continued usefulness of
93
culture and offers a strategy for operationalization that has similarities with the methods
and theories used by cognitive anthropologists (Brumann 1999). Brumann advocates the
development of matrices of traits for informants to help in detecting the patterns that
indicate cultures. In response to his article, Lila Abu-Lughod, a leading voice in the
effort to jettison the concept of culture, agrees that there are certainly patterns of thinking
and behaving which would be called culture but discounts the effort to operationalize
culture because she has doubts “that anyone today would be willing to undertake the
formidable task of drawing up matrices of shared and nonshared features” (Abu-Lughod
1999, p. S14). This argument by Abu-Lughod is not actually an argument against the
method proposed by Bruman. Rather it is a testimonial to ignorance of and disinterest in
methods of cognitive anthropology as well as multivariate statistical methods in general.
Often this ignorance of operational definitions of culture allows for the dismissal of
the possibility of the quantification of anthropological concepts out of hand. Greenhalgh
characterizes anthropological theoretical constructs as “inherently difficult to
operationalize, measure and analyze as variables”, she argues that they “simply do not
lend themselves well to being treated” as quantifiable (Greenhalgh 1996, p. 48-49), and
states that arguments that are based on culture “are not empirically falsifiable”
(Greenhalgh 1988, p. 632). My argument is that the anthropological culture of nonquantifiability has allowed anthropologists to create theoretical concepts without regard
to operationalization, which leads to lingering doubts about their connection with reality.
This lack of an operational definition of culture is likely to be the main stumbling block
to a greater sophistication in the use of culture in demography.
94
Conclusion
Although the majority of anthropologists may never fully embrace the
methodological advances being made in cognitive anthropology using multivariate
statistics, this has little baring on work that can be done in the name of anthropological
demography. In the remainder of this manuscript, I will detail my attempt to put these
theoretical and methodological principles to the task of greater understanding of the role
of culture in the global fertility decline process. I have chosen Honduras as a suitable
location to conduct a study of culture, changing gender roles and fertility decline. In the
next chapter, I discuss Honduras as a case study of the process of fertility decline. In
subsequent chapters, I discuss my operationalization of the domain of natality culture and
the data collection process I underwent to test the theory of fertility decline I discussed in
chapter 2.
CHAPTER 5
REPRODUCTIVE ATTITUDES AND BEHAVIOR IN HONDURAS
Introduction
In this chapter, I introduce Honduras as an interesting case study of the ideas I
developed in the preceding chapters. In the first section, I present the Honduran fertility
decline as a way of illustrating important theoretical and methodological issues in the
study of fertility decline. I present the history of fertility decline in the 20th Century and I
present the prevailing demographic interpretation of the fertility trends in Honduras and
Central America. In the next section, I critique traditional demographic analysis of
fertility data which is exemplified by the Honduran case. I argue that demographers
typically take an inductive approach to developing theory to explain fertility decline. I
discuss how this induction-heavy approach has consequences for measurement and
theory building.
In the final section of this chapter, I present an analysis of a data set from two
recent surveys completed in Honduras: the “Encuesta Nacional De Epidemiologia Y
Salud Familiar, 1996”(Monteith et al. 1997a) and the “Encuesta Nacional De Salud
Masculina, 1996” (Monteith et al. 1998). These surveys provide answers to reproductive
attitude and behavior questions for independent samples of Honduran men and women. I
present this analysis to illustrate a critique of the traditional demographic analyses of
fertility data. I conclude that greater attention to theory in all phases of fertility research
would help improve demographic findings in fertility research.
95
96
Honduras as a Case Study in Theories of Fertility Decline
Honduras presents a useful case in the examination of theoretical issues in fertility
research because its aggregate fertility rate has lagged behind most of its neighbors in
Central America. The entire region is theoretically interesting because it has been
characterized as a region with no single pattern of fertility decline (Guzman 1996, p.
xxiii). In this section I will summarize Honduras’ history of fertility decline in the
context of Latin American fertility decline. This analysis will serve as an example in my
critique of traditional demographic theorizing about fertility decline.
History of Fertility Decline in Honduras and Central America
Figure 5-1 shows that in Latin America during the last half of the 20th Century,
fertility declined in every country yet the pace of this decline has varied for individual
cases (Chackiel 1996, p. 4). Some countries began and ended the period with low rates
of fertility, but most countries began the period with high rates and then shifted to
medium level fertility rates. Some began the period with high rates and continued to
have high rates throughout the century. Honduras falls into this latter category, having
one of the slowest rates of fertility decline in all of Latin America.
These data contradict generally accepted predictions (Chackiel 1996, p. 6-9).
Demographers have been unable to detect an overarching pattern in these nationally
aggregated rates except for the weak link between development and fertility decline
(Bongaarts and Watkins 1996; Guzman 1996, p. xxiii). There appear, instead, to be two
patterns: 1) Costa Rica and Panama, which have experienced a decline in fertility and
mortality, and 2) Honduras, Nicaragua, Guatemala and El Salvador, which have
continued to have high rates of fertility and mortality (Latin American Demographic
Centre 1997, p. 4).
97
8.0
7.0
6.0
TFR
5.0
4.0
3.0
2.0
1.0
0.0
1950-
1955-
1960-
1965-
1970-
1975-
1980-
1985-
1955
1960
1965
1970
1975
1980
1985
1990
Year Range
Figure 5-1 Latin American TFR by country 1950-55 to 1985-90. (Source: Chackiel
1996, p. 5. Note: Bold indicates Honduras. Dashed lines indicate other
Central American Countries.)
Causes of Fertility Rate Variation in Central America
Many factors have been hypothesized to cause variation in fertility decline in
Central America. Levels of urbanization have been changing in the region with some
countries experiencing more change than others. Figure 2 shows the changes in
percentage of population living in urban areas for each country during the past fifty years.
Figure 3 shows the pace at which urban sectors have been growing in these countries.
Honduras has far and away the fastest current urban growth rate when compared with its
overall growth rate. A consistent finding across nations is that urban fertility rates are
much lower than rural fertility rates and this is especially true in Honduras (Latin
98
American Demographic Centre 1997, p. 16; Guzman 1997, p. 48). In the large urban
areas of Honduras, the total fertility rate (the average number of children being born to
each woman throughout her lifetime) is 3.14 while in the rural areas it is 6.34 (Monteith
et al. 1997a, p. 56).
70
60
Central America
50
Costa Rica
Percent
40
El Salvador
30
Guatemala
20
10
Honduras
0
2000
1995
1990
1980
1970
1960
1950
Nicaragua
Panama
Year
Figure 5-2 Degree of Urbanization expressed in percentage of total population residing in
urban areas. (Source: Latin American Demographic Centre 1997, p. 38).
Another general trend related to urbanization is that the difference between urban
and rural fertility rates is large at the onset of fertility decline, but decreases as the
transition progresses. However, in Honduras, the level of rural fertility actually increased
while the urban fertility rates declined (Chackiel 1996, p. 16-17). The difference between
urban and rural fertility in Honduras remains one of the biggest such differences in Latin
America (Chackiel 1996, p. 16). Thus, urbanization, while certainly linked to declining
99
levels of fertility, exhibits some unexpected properties as evidenced by the Honduras
case.
In addition to increasing urbanization, demographers have theorized that this and
other structural changes in social life in Central America have caused changes to
normative fertility attitudes, resulting in lower fertility (Guzman 1996, p. xii; Guzman
1997, p. 42-43). These structural changes include changes in marriage patterns, more
educational opportunities, higher living standards, more social mobility, and more
women entering the workforce (Latin American Demographic Centre 1997, p. 5). The
stress of these new social pressures causes changes in these value systems resulting in
greater social aspirations for goods and services such as education (Guzman 1996, p.
xxiv; Guzman 1997, p. 47).
Unmet Need for Family Planning
An important concept that is often invoked in demographic studies of fertility
decline is the unmet need for family planning. This concept is particularly important in
cases where the actual fertility levels of a particular country, such as Honduras, are higher
than would be predicted by demographers. This conceptual variable, “the level of unmet
need in a particular population”, is evaluated by some combination of measurements,
including desire for future child bearing, use of contraceptives during last pregnancy, and
ideal family size. These elements are then compared with actual behavior regarding
childbearing and reported behavior about the use of contraceptives. It is a common
finding that, in developing countries, many people express a desire either to terminate or
delay childbearing, yet they do nothing to satisfy this desire. This finding is generally
interpreted by demographers that there is a lack of family planning services (Potts 2000).
100
This follows from the stated goal of family planning programs: to provide services for
couples to have as many children as they want to have, when they want to have them
(Hubacher 1997, p. 534).
Central
35
America
Costa Rica
30
Percent difference
25
El Salvador
20
Guatemala
15
Honduras
10
Nicaragua
5
0
1995-00
1990-95
1980-90
1970-80
1960-70
1950-60
Panama
Year range
Figure 5-3 Tempo of urbanization expressed in difference between urban population
growth rate and the population growth rate of whole population. (Source:
(Latin American Demographic Centre 1997, p. 38)).
Several researchers have remarked on the high levels of unmet need for family
planning in Honduras and have implicated this in its continued high rates of fertility
(Latin American Demographic Centre 1997, p. 21; Monteith et al. 1997b, p. 16-17). In
these studies, researchers have noted that the ideal number of children as reported by
women is lower than the reported number of children actually born. Also, these studies
have found a percentage of women who indicate that they do not desire to have any more
101
children, either at all or in the near future, yet they are not currently using contraceptives
(Magnani et al. 1998, p. 305,316-317). These researchers have speculated that the causes
of unmet need for family planning include a low level of social interaction whereby
women do not have information about contraceptives, have not had their normative
attitudes about desired large families challenged, do not know how to use or obtain
contraceptives, or do not have social expectations to pursue educational opportunities. I
will discuss the argument for social interaction as a major force in causing fertility
decline in the next section.
Social Interaction as a Cause of Fertility Decline
An influential paper by Bongaarts and Watkins (1996) makes the case for social
interaction as a major force in causing fertility decline. Bongaarts and Watkins offer an
extensive examination of theories of fertility decline with a special emphasis on the
comparison between theories that focus on economic forces and theories that emphasize
social interaction. The debate between economic changes versus social interaction is
among the liveliest contemporary theoretical debates about fertility decline.
Bongaarts and Watkins show that economic changes are the most important overall
cause of fertility decline. The R2 between fertility and a multivariate measure of
development among 69 developing nations for the period of 1960-65 to 1985-90 is -.60
(Bongaarts and Watkins 1996, p. 642-643). However, as with the European data, this
strong negative correlation between fertility rates and economic variables fails to predict
either the onset of fertility decline or its pace for particular countries. In some cases,
variables that are associated with development, such as literacy or urbanization, are
associated with a particular country or region’s drop in fertility rates, while in others they
are not (Watkins 1986). The discovery of this pattern–underdeveloped regions often
102
experienced fertility declines shortly after the declines of adjacent, highly industrialized
regions–was the impetus for the social interaction hypotheses (Bongaarts and Watkins
1996, p. 640).
The theory behind this hypothesis is that channels of communication and
interaction are necessary for the transferal of information. Information exchange and
social influence are the “oils of the machine” of fertility decline (Bongaarts and Watkins
1996, p. 656). The pace and onset of fertility decline for a particular country is a function
of the strength and effectiveness of its channels of communication. A high level of
unmet need for family planning is a direct result of the lack of channels of
communication because it leads to a lack of information about effective contraceptives,
the persistence of cultural norms that place a negative stigma on contraceptive use, and
the exorbitant costs of contraceptives. With more channels of information and influence
exchange, people are more connected to ideas about fertility limitation through their
personal networks and the pace of fertility decline proceeds faster.
Critique of Demographic Analyses of Fertility Decline
The demographic analyses above illustrate the strategy of demographers for
understanding Honduran fertility (as well as fertility decline in general): see which
aggregate socio-economic variables correlate with overall fertility rates. When this
strategy can only explain differing fertility rates to a limited degree, the concept of unmet
need for family planning is evoked to fill in the empirical gaps. In these analyses, a
theory explaining descriptive findings is somewhat of an afterthought. Theory is
incorporated into the analysis only after the standard data analytic procedures to describe
the population are complete and found wanting. The goal of the theoretical analysis is to
explain the changes in national level statistics, not individual behavior. With this
103
approach, there is virtually no consideration of why people in Central America are less
likely to be motivated to have children than previous generations were and why this
motivation varies based on national residence.
The fundamental problem with this type of analysis is that it depends on induction
as the only source of theorizing. As I discussed in Chapter 2, the cycle of science
depends on deductive as well as inductive theorizing so that hypotheses can be falsified,
leading to adjustments in theory (Kuznar 1997). When demographers develop theories
about reproductive behavior through the analysis of aggregate national statistics, it is an
important and necessary but incomplete element in the scientific process of
understanding human reproductive behavior. Theories that explain empirical findings
should lead to testable hypotheses that drive the collection of additional data that may
falsify these hypotheses.
Demographers typically analyze data that were collected for purposes other than
falsifying hypotheses. The primary use of most demographic data is to describe large
populations. Data collected for descriptive purposes can be used to test theories and
falsify hypotheses if the variables collected and the units of analysis used are amenable to
this type of analysis. However, this is not always the case.
I suspect that the underdevelopment of demographic theory of fertility decline is at
least in part due to the common practice of explaining findings from existing data without
collecting data that can be used in deductive tests. Thus, to use a common metaphor, the
existing-data-set tail is often wagging the demographic-theory dog. Without a theory of
fertility decline to direct data collection to be used for hypothesis testing, demographic
researchers are susceptible to explaining fertility decline with spurious associations.
104
Critique of Social Interaction Theory
The Bongaarts and Watkins article is a good example of this the tendency to rely on
explanations involving spurious associations. Although Bongaarts and Watkins are
correct in highlighting the role of personal networks in influencing changes in behavior
and attitudes, they ignore the role of social networks in economic change. It is difficult to
isolate the role of networks in the transformation of information about contraceptives
versus information about economic opportunities. Their assertion that the diffusion of
information about contraceptives can affect fertility patterns independently of social and
economic circumstances would have to be proven rather than imputed from correlations
between regional development and fertility levels. It is not likely that a developed
region’s economic influence on fertility rates would simply stop at its borders, especially
if it is adjacent to an underdeveloped region with a population of people sharing “a
common language and elements of a common culture” (Bongaarts and Watkins 1996, p.
647). Social networks would likely span these regions bringing inhabitants of
underdeveloped areas connections to economic opportunities, which in turn cause
temporary migration and increased trade as well as information exchange about family
planning practices. Social networks are an essential element in the spreading of
information about economic opportunities (Granovetter 1973).
Thus, the ultimate cause of fertility change is still development and greater
industrialization. Bongaarts and Watkins have highlighted an aspect of fertility decline
that needs to be incorporated into any model of fertility transition, but social interaction
by itself cannot be used as an explanation of a change in behavior. Social interaction
itself needs to be explained.
105
Even though they admit that the proliferation of channels of social interaction is a
“concomitant of development” (Bongaarts and Watkins 1996, p. 667), they describe
social interaction as an independent force on fertility. They state that, “We would expect
more rapid fertility decline in countries where a multiplicity of channels connects
communities, and slower fertility decline where such channels are sparse” (Bongaarts and
Watkins 1996, p. 662). However, isolation from channels of communication is often an
indication of economic isolation as well, so we are left with a description rather than an
explanation.
Urbanization, increased educational opportunities, attitudinal change, and more
extensive family planning services are other examples of spurious associations that have
been used as explanations for fertility decline. Urbanization is certainly linked to fertility
decline, but not in a direct way. Often, it interacts with other factors causing it to have
differential force at different time periods (Handwerker 1992, p. 1247). Increased
educational opportunities have also been shown to depend on the level of economic
opportunities for women. When education cannot be parlayed into economic
advancement, it does not affect fertility (Handwerker 1992, p. 1247). Attitudinal
changes cannot be considered a cause of fertility decline because you cannot explain a
change in preferences for family size by a change in preferences. This is a tautology
(Handwerker 1986, p. 401).
The effect of family planning programs on fertility decline is somewhat more
complicated. Certainly the ability to obtain modern, effective means to prevent
contraception facilitates the limiting of births and is a technological necessity to achieve
below-replacement fertility levels (Handwerker 1992, p. 1247). However, there is some
106
debate about the role family planning programs play in causing fertility decline.
Throughout human history, unwanted childbearing has been prevented in a variety of
ways including cultural taboos and indirect infanticide (Harris and Ross 1987; Scrimshaw
1983). The assumption that human fertility would be unrestrained without family
planning programs is false because high fertility has always been below the biological
maximum possible (Bongaarts 1975; Hirshman 1994, p. 206). Also, family planning
programs have been shown to have ambiguous results and have even been shown to
enable users to better space births to reduce infant mortality, effectively causing an
increase in completed family size (Bledsoe 1994, p. 111; Easterlin 1985, p. 22-24;
Handwerker 1992, p. 1247).
The concept of unmet need for family planning is a related factor that may have an
ambiguous effect on fertility decline. Typically, the conceptualization and measurement
of unmet need for family planning is another area of demographic study of fertility
decline that has lacked a sufficient theoretical approach. The choices that demographers
make in the collection and analysis of data regarding unmet need for family planning
often reveal an under-emphasis on the testing and development of theories of human
fertility. Specifically, the choice to base unmet need for family planning measurements
on data collected only from women reveal a neglect of a theory of gender. I will discuss
this in more detail later in this chapter.
National Level Studies of Fertility Decline
Perhaps the main cause of the proliferation of demographic theories of fertility
decline with spurious associations is the persistence of the choice to make nations the
units of analysis without a theoretical justification. Because of its bias towards producing
studies that have policy implications (Greenhalgh 1996), demography often conducts
107
studies with nation states as the primary unit of analysis. This is understandable given
that the consumers of the results of studies, and often those funding the studies in the first
place, are the nations themselves or international funding agencies that have a vested
interest in partitioning data along national lines. Demographers often do not conduct
their own data gathering and demography graduate students are discouraged from
collecting their own data, so sources data for demographic investigation often come prepackaged in discrete, nationally defined units. Also, governments sometimes have
political needs that dictate that survey data be collected in idiosyncratic ways. This
makes the combination of survey data across countries difficult, which further influences
researchers to aggregate findings and to treat whole countries as individual cases.
This practice of considering nations as units of analysis has negative effects on
building consistent and generalizable theories of demographic processes, especially
fertility decline. The use of aggregate demographic data places important limitations on
the types of theories that can be tested. This practice is the best evidence that theory is
sometimes secondary to demographers’ analyses of data sets (Greenhalgh 1995, p. 4; Hill
1997. p. 224; Kertzer 1997, p. 2; McNicoll 1980, p. 441). Theory suffers when nations
are used as the default unit of analysis because, unless researchers can justify the
theoretical connection between national status and the studied outcomes of their data
modeling, it should not be used as a unit of analysis.
This is not to say that nations should never be used as units of analysis in fertility
studies. In nations that have explicit, highly enforced demographic and fertility policies,
like China, the use of a nation as a unit of analysis is theoretically justified. China’s onechild policy surely influences the fertility decisions of many Chinese people (Greenhalgh
108
1988). Geoffrey McNicoll’s (2001) analysis of pre- and post-transitional societies based
on pro-natalist or anti-natalist policies and interventions of governments demonstrates the
link between national identity and fertility behavior. However, a more common example
is the use of national level data in the study by Bongarts and Watkins. They state that
Costa Rica, Brazil and Chile were the first Latin American countries to enter the fertility
transition (in 1965) and Bolivia and Guatemala were the last (in 1978) (1996, p. 650).
There is no obvious explanation for how being Costa Rican produced decisions in 1965
that lowered fertility in 1965 while being Bolivian did not. Characteristics of these
countries contribute to changing aggregate fertility rates, but countries do not experience
the fertility transition. They reflect it. People experience fertility a transition and are led
to do so through their participation in culture-forming social networks that channel
information about how best to live one’s life.
Culture Theory and Fertility Decline
Rather than explaining fertility decline through a theory of social interaction, a
theory of culture change that includes social interaction is much more powerful way to
explain fertility decline. Fertility decline happens because anti-nalalist behaviors and
thoughts become culturally accepted by a group of socially interacting network members.
This happens over time as new modes of resource acquisition become available, and new
behaviors and thoughts are necessary to compete for these resources. Thus, it becomes
culturally accepted to delay and limit childbearing while pursuing economic advancement
and it becomes culturally prohibited to expect access to resources through other people
(Handwerker 1989).
Cultures (groups of socially interacting members who share similar ideas and
observe each other’s behaviors) are the units of analysis in this theoretical framework.
109
Material changes in the environment act indirectly on cultures through direct influence on
the lives of many their members. As members of a culture who are affected by material
changes in the environment adjust to these changes, they discuss their lives with other
members of their culture. After a period of dissonance between the existing culture and
the behavioral adjustments to new material realities, a culture change process begins
(Bernard and Pelto 1987). Eventually this process leads to changes in the thoughts and
behaviors of those network members who have no direct experience with the material
changes. Thus, social interaction is a key component to the process, but is not itself
causing any change.
Gender and Fertility Decline
Demography has often been criticized for its unsophisticated use of the concept of
culture (Hammel 1990). Related to this criticism is the critique of demography for its
neglect of gender in human fertility studies (Riley 1998; Riley 1997). This is important
because, as I discussed in Chapters 1 and 2, a theory that accounts for differences in
gender roles is central to theories of fertility decline. The area in which a demographic
neglect of gender is the most obvious is in its assessment of unmet need for family
planning. Over the years, many measures of unmet need for family planning have been
developed, but studies have consistently used data collected only from women to
understand this phenomenon (DeGraf 1991; Phan 1997; Ross 1997; Sai 1997; Westoff
1978; Westoff 1981a; Westoff 1988; Westoff 1981b).
It is true that women generally bear the primary responsibility for child rearing and
they have the sole responsibility for child bearing. However, the practice of
conceptualizing unmet need for family planning purely in terms of women’s attitudes and
behaviors means that no theory of gender is being used. Excluding men from the analysis
110
of reproductive desires assumes that a woman’s preferences accurately represent a
couple’s preference and that no couples have conflict over reproductive decisions. It also
means that there is no reference to gendered divisions of labor, gendered variation in
access to resources, and gendered variation in power relations–all of which can effect a
woman’s ability to realize her own childbearing intentions.
Dodoo and Seal argue for a greater inclusion of men in fertility studies. They
found that there were significant differences between men’s and women’s attitudes
within individual couples and in aggregate measures (Dodoo and Seal 1994). Dodoo and
Seal also found that a husband’s preferences may at times eclipse his wife’s preferences
for reproductive outcomes by preventing her from using contraceptives. Other studies
confirm the influence of spouses over each other in reproductive decisions (Bankole and
Singh 1998; Ezeh 1993; Razzaque 1999). Ethnographic evidence has supported the
contention that not just men but entire communities sometimes prevent women from
using contraceptives when the women would otherwise want to limit their reproduction
(Browner 1986). Zulu showed that in using traditional methods of birth control, older
female relatives play a significant part in influencing reproductive decisions while men
became involved with the introduction of modern methods (Zulu 1998).
Mason and Taj (1987) did a meta-analysis of studies in 30 developing countries
where both men’s and women’s fertility preferences were measured. Their conclusion is
that, although the prevailing literature suggests that women often want to have fewer
children than men do and the gap between their attitudes and behaviors (i.e. they have
more children than they want) is due to this difference, there is, in fact, no empirical
evidence to support this position. However, their study averages the preferred number of
111
children for men and women for entire nations. This is driven, of course, by the fact that
the individual studies in the meta-analysis had defined their populations of study as
nations. However, aggregate averages of fertility preferences may be obscuring
heterogeneity within each country (Dodoo and Seal 1994, p. 381).
An approach that takes gender into consideration would not expect consistent
behaviors and preferences for men and women regardless of context. The countries in
the meta-analysis of Mason and Taj are likely to have many different forms of gender
institutions. Even within the individual countries in the study there are likely multiple
gender systems in place at a given time. Certain circumstances may influence women to
want more children than men, such as their lack of power compared to men in highly
sexually stratified societies. High fertility in these societies may be a response by women
to increase their avenues for obtaining resources through their children rather than relying
solely on their husbands for income (Handwerker 1993, p. 42). In situations where
women produce their own resources, they may want to limit their fertility because it is a
hindrance towards pursuing educational and employment opportunities. If these two
situations were going on at the same time in the same country, an overall average would
give the impression that there is strong husband and wife agreement. Comparison of
aggregate averages of fertility preferences for men and women does not take gender into
consideration because gender implies cultural variability, which is not the same thing as
national variability.
The main research question asked by Mason and Taj is interesting: what is the role
of gender in reproductive decision-making? However, without operationalizing gender
differences, the question is impossible to answer. A measurement of gender differences
112
would consider the context of the roles, the local sexual division of labor, and the
expected attitudes and behaviors for men and women. As cultures change, these roles
and expectations change. Whether or not there are consistent differences between sexes,
regardless of gender, is an interesting question to study in its own right. However, this
would have to be done with something other than aggregate averages.
Studies in Honduras (Magnani et al. 1998; Monteith et al. 1997b) conclude that
there is a high level of unmet need for family planning, but again, these conclusions are
based on surveys conducted with only women and use aggregate data. Questions remain
about the extent of intra-national heterogeneity in the relationship between gender and
reproductive attitudes in Honduras. In the proceeding sections, I address these issues
with an analysis of two surveys conducted in Honduras in 1996, the “Encuesta Nacional
De Epidemiologia Y Salud Familiar, 1996” and the “Encuesta Nacional De Salud
Masculina, 1996.” I analyze these data sets with an emphasis on gender rather than sex
as a means of investigating the differences in attitudes between women and men
regarding reproductive intentions on a sub-national level.
Reproductive Attitudes of Honduran Men and Women
The “Encuesta Nacional De Epidemiologia Y Salud Familiar, 1996” (National
Survey of Epidemiology and Family Health or ENESF-96 (Monteith et al. 1997a))
surveyed 7,505 randomly selected women and the “Encuesta Nacional De Salud
Masculina, 1996” (National Survey of Male Health or ENSM-96 (Monteith et al. 1998))
surveyed 2,925 randomly selected men. The two surveys produced independent
samples–there were no data collected on spouses. Each survey asked the same questions
regarding the number of children ever born, attitudes about ideal number of children, and
knowledge, attitudes, and practice questions about contraceptives. Comparison of the
113
answers given by men and women at the aggregate level of the country as a whole show
similarities. This is true for both total fertility rate and ideal number of children.
Comparison between these two measures reveals a difference between ideal number of
children when compared with actual fertility rate, suggesting an unmet need for family
planning.
Table 5-1 Median Ideal Number of Children by Sex and Location
Men
Location
Women
TFR
Ideal
TFR
Ideal
Tegucigalpa/SPS
3.70
2.8
3.14
2.7
Other Urban
4.71
3.0
3.94
2.8
Rural
5.28
3.4
6.34
3.3
Total
4.77
3.2
4.94
3.0
Source: (Monteith et al. 1997a, p. 100)
Table 5-1 shows this comparison between men and women for three regions of the
country. Within regions and for the country as a whole, men and women differ only
slightly in their perceptions of the ideal number of children. The data are divided into
three regions because of the significant differences in the three environments.
Tegucigalpa and San Pedro Sula are the two main urban areas of Honduras. Tegucigalpa
is the capital and San Pedro Sula is the center for commerce. These two cities are
significantly different than any other urban area in Honduras in both size and population.
The other urban areas are much smaller and more closely tied with the rural areas of the
country providing an intermediary category between Tegucigalpa/San Pedro Sula and the
rural regions.
114
Table 5-2 Planned/Not Planned Status of Last Birth by Sex (in percentages)
Not Planned
Planned %
Desired % (Overall %)
Not Desired %
Men
82.5
10.4 (92.9)
Women
64.1
23.1 (87.2)
Odds Ratio, Planned vs. Not Planned: 2.64 (2.25, 3.10)
Odds Ratio, Desired vs. Not Desired: 1.92 (1.52, 2.44)
DK/NR %
6.8
.2
12.3
.5
Source: (Monteith et al. 1997a, p. 97)
The similarity between the answers of men and women breaks down in table 5-2.
Although table 5-1 would suggest non-significant differences between Honduran men
and women in their desires for children, further exploration reveals some important
differences. Table 5-2 shows a large difference between men and women in their
perceptions of their most recent birth. Respondents were asked if their most recent birth
(during the past five years) was planned. If the answer was “no”, the respondent was
asked if the birth was desired or not desired (Monteith et al. 1998, p. 97). Men are 2.6
times as likely as women to say that their last birth was planned. In the column labeled
“Desired”, the percentage in the parentheses indicates the combination of the percentage
who planned their last birth and the percentage who did not plan their last birth but
thought of it as desirable anyway. Men were 1.9 times as likely as women to say that the
last birth was desired. These odds ratios are significant at the 95% confidence level.
These two tables seemingly contradict each other. In table 5-1, aggregate numbers
show that, as Mason and Taj found, differences between sexes are, on the average, small.
Table 5-2 shows cultural differences between men and women regarding specific births.
These differences do not show up in aggregate data on the ideal number of children.
115
Analysis of Dios Me Mande
Because these tables only partially addressed the questions I had about the potential
for cultural and gendered differences in the desirability of children in Honduras, I decided
to conduct additional analyses of the raw data. I received SAS data sets of the raw data
used to produce the demographic reports, as well as SAS programs to analyze this data,
from the reproductive health division of the Centers for Disease Control. The SAS code
that accompanied the data created new variables from combinations of individual
questions and recoded the response categories for certain variables.
One interesting thing I noticed from looking at these programs was that not all of
the responses to the ideal number of children question were being described in the final
report. In the question that asks the ideal number of children desired by a respondent,
one of the possible responses is Dios me mande or “However many God sends.” The
SAS code that analyses the ideal number of children question effectively treated the
answer, Dios me mande as “Don’t know.” On the one hand, this answer truly is a “Don’t
know” response because the respondent is indicating an unknown quantity. However,
this answer indicates something qualitatively different than “Don’t know” with
implications for a theory of fertility decline. The final reports summarizing the findings
from these surveys reported the numerical responses alone and ignored both the Dios me
mande and “Don’t know” responses as missing values. Because I was looking for an
explanation for why this question did not conform with other questions, I thought that the
“Dios me mande” responses might provide a key to better understanding the aggregate
responses to the ideal number of children question.
Although I was familiar with the common practice in survey analyses whereby
non-numerical responses were re-coded as missing values, I asked the Centers for
116
Disease Control demographer who provided me the data, Paul Stupp, why these
responses were recoded as “Don’t know.” His response was that there were so few of
these answers that they were not important enough to analyze by themselves (personal
communication, March 20, 2001). This made sense to me in relation to the goals of these
studies: to describe the health characteristics of men and women in Honduras. Because
my goals were very different (testing hypotheses derived from theory) I recoded this
variable and re-analyzed the ideal number of children question including the Dios me
mande responses.
The inclusion of the Dios me mande responses made it necessary for me to analyze
the data with categorical statistical techniques. This is because the response Dios me
mande is a nominal level response rather than interval level. The numeric responses to
the ideal number of children question, on the other hand, are interval level. In order to
analyze the interval level data in conjunction with the nominal level data, I had to
collapse the interval level data into nominal categories.
The rule of thumb in statistical analysis is to always use a higher level of
measurement when possible because higher levels of measurement contain more
information (Bernard 2002, p. 42-47). Interval level data are higher than nominal level
data. Therefore, my choice to collapse the interval level data into nominal categories
might seem to be a violation of this rule of thumb. However, because nominal level data
cannot be converted into interval level, the nominal level data would have to be thrown
out in order to analyze the data on an interval level. This is exactly what the original
analysts of the data did. In general, it is better to use interval level data when possible,
but it is worse if doing so requires that some of the data be ignored.
117
Re-analysis of National Data
Table 5-3 compares the percentages of people who gave a numerical response to
the question about the ideal number of children to the percentage of those saying “Dios
Me Mande.” The table breaks down these percentages for each of the regions of
Honduras. Although the percentages are small in each region, people living in rural areas
are more likely to say say Dios me mande, indicating potential significance for this
response in understanding fertility behavior in Honduras.
Instead of using Dios me mande as an expression of an unknown, I treated it as an
indication of preference for a high level of fertility. That is, respondents are saying that
they are not going to do anything to prevent a birth and are leaving it up to God. In table
5-4, I have assigned cases up to 4 children as “low desired fertility.” All other cases,
either 5 or more or Dios me mande–are assigned to “high desired fertility.” (I chose the
cutoff level of five because it was higher than the TFR for the entire country.) This table
presents a compliment to Table 5-1, which showed a slight trend in the greater demand
for children among more rural respondents. Table 5-4 is similar to Table 5-1 in that there
is a substantial difference between rural and urban when indicating a preference for a
high level of fertility.
Table 5-3 Has Ideal Number vs. “Dios Me Mande”
Location
Numerical
DMM
Total
Response
Tegucigalpa/SPS
2410 (99%)
29(1%)
2439
Other Urban
2459 (98%)
57 (2%)
2516
Rural
4458 (92%)
402 (8%)
4860
118
Table 5-4 Dichotomized Ideal Number of Children. (High >= 5 or “Dios me mande”)
Location
Low
High
Total
Tegucigalpa/SPS
2259 (94%)
151(6%)
2410
Other Urban
2257 (92%)
202 (8%)
2459
Rural
3669 (82%)
1793 (18%)
4462
Tables 5-5 and 5-6 present the results of an analysis of reproductive preferences for
women and men, controlling for region of residence. Table 5-5 details the distribution of
answers to the ideal number of children question dichotomized as either a numerical
response or a Dios me mande response. The table shows the raw number of answers for
each response type, as well as the overall percentage, for men and women living in the
three regions of Honduras.
Table 5-5 Has Ideal Number vs. “Dios Me Mande” by Sex and Location
Tegucicalpa/ SPS
Other Urban
Rural
Men
Women
Men
Women
Men
Women
Has Number
834 (98%)
1576 (99%)
768 (98%)
1691 (97%)
1106 (95%)
3356 (91%)
DMM
15 (2%)
14 (1%)
12 (2%)
45 (3%)
59 (5%)
343 (9%)
Odds Ratio
(95% CI)
2.02 (.97, 4.21)
.59 (.31, 1.12)
.52 (.39, .69)
Table 5-5 also shows odds ratios comparing the likelihood of men saying Dios me
mande compared with women for each region. Table 5-5 also shows 95% confidence
intervals for the odds ratios. For Tegucigalpa/San Pedo Sula, men are 2.02 times as
likely to say Dios me mande than women. The confidence interval for this comparison
includes 1, which means that it is not significant at the 95% confidence level. However,
the interval is mostly above 1 and only dips slightly below 1 while it ranges up to 4.21.
119
Therefore, the odds ratio is probably significant at a confidence interval of slightly less
than 95%. It is reasonable to conclude that men who live in Tegucigalpa/San Pedro Sula
are more likely to say Dios me mande than the women who also live in these two cities.
For the other urban areas, men are only 59% as likely to say Dios me mande as women.
However, the confidence interval ranges well above 1 leading to the conclusion that there
is no justification for claiming significant differences between men and women’s
responses in these areas. In the rural areas, men are 52 % as likely to say Dios me mande
as rural women. The 95% confidence interval resides well below 1 giving justification to
the claim that rural women are significantly more likely to say Dios me mande than rural
men.
Table 5-6 Dichotomized Ideal Number of Children by Sex and Location
Tegucicalpa/ SPS
Other Urban
Rural
Men
Women
Men
Women
Men
Women
Low
High
771 (91%)
1488 (94%)
701 (90%)
1556 (90%)
944 (81%)
2725 (74%)
78 (9%)
102 (6%)
79 (10%)
180 (10%)
221 (19%)
974 (26%)
Odds Ratio
(95% CI)
1.47 (1.09, 2.01)
.97 (.74, 1.29)
.66 (.56, .77)
Table 5-6 represents a different dichotomization of the ideal number of children
question. In Table 5-6, the answers to this question are divided into two categories:
those who said that the ideal number of children is 4 or less and those who either said that
the ideal number of children was 5 or more or Dios me mande, as in table 5-4. Table 5-6
divides the data into the frequency of responses for men and women living in the three
areas of Honduras. Table 5-6 also gives the same type of odds ratio analyses that were
presented in Table 5-5.
120
Although saying Dios me mande is not the same as the survey respondents
indicating that they prefer a high number of children, I have included these responses
together for analytic purposes. Those deciding to have as many children as God sends
would typically shun the use of modern contraceptives because they believe that there are
religious sanctions against curtailing the reproduction that would follow naturally from
sexual intercourse. Therefore, for the purpose of this analysis, people who have attitudes
favoring a high number of children are functionally the same as people who favor having
all the children that God sends.
The results of the odds ratio analyses in Table 5-6 are also very similar to Table 55. Men living in the two major urban areas of Tegucigalpa and San Pedro Sula were 1.47
times as likely to say that they wanted a high number of children. This statistic is
significant at the 95% confidence level. The percentages of men and women who live in
other urban areas of Honduras are close to being exactly the same and show no
significant differences. On the other hand, rural men are only 65% as likely as rural
women to report that they desire a high number of children. The confidence interval for
rural women and men is also well below 1, indicating a highly significant finding.
The most striking feature of Tables 5-5 and 5-6 is the flip-flop of percentages from
men to women in the rural and major urban areas. Women are more likely to respond
with higher numbers of children and they are more likely to say Dios me mande in the
rural areas than men. However, in Tegucigalpa and San Pedro Sula, men are more likely
to fall into these two categories. The urban areas besides Tegucigalpa and San Pedro
Sula appropriately show no significant differences between the answers given by men
and women.
121
Discussion
In this chapter, I described the problems I have seen in the traditional demographic
analyses of fertility decline. The biggest problem that I have noticed is the excessively
inductive approach to investigating the causes of fertility decline. From this approach,
many other problems with measurement and theoretical implications develop. I have
demonstrated how a more theory driven approach, including a greater attention to culture,
gender, units of analysis, and levels of measurement, can produce much different
conclusions than approaches that neglect theory. I found that there is evidence of cultural
differences between men and women in Honduras regarding the desirability of
childbearing. I also found that these differences are related to rural or urban residence.
In the next chapters, I build on these findings by demonstrating my approach to
measuring the culture of natality in Central Honduras. In Chapter 6, I discuss my choice
of operationalization and my data collection strategy. In Chapter 7, I discuss the results
of my analysis.
CHAPTER 6
ADAPTATION AND ADMINISTRATION OF THE CHILDBEARING
QUESTIONNAIRE
Introduction: Natality Culture
The term pro-natal has been used to describe cultural systems that promote
childbearing (e.g. (Harris and Ross 1987, p. 11; Mason 2001, p. 308; Mason and Taj
1987, p. 619; Sangree 1997)). It is occasionally contrasted with anti-natal culture that
discourages procreation (e.g. (Harris and Ross 1987, p. 11)). The fertility transition is
often described as a process of change from a pro-natal culture to an anti-natal culture
(Handwerker 1986, p. 10). However, the assertion of one, overall cultural pattern of
natality (either pro- or anti-) assumes shared thought, feeling and behavior among the
members of a population regarding the positive and negative aspects of reproduction. To
test a theory of fertility decline that includes culture as a causal force, the existence of a
shared natality culture–or cultures–must be operationalized and demonstrated.
In this chapter, I discuss the operationalization of natality culture. I discuss my
choice to adapt the Childbearing Questionnaire (CBQ) (Miller 1995) to measure natality
culture in central Honduras. The first section of this chapter is a description of the CBQ
and the theoretical model of childbearing motivation that led to its development. I also
discuss the results of empirical studies that have used the CBQ.
In the second section, I discuss the steps I took to adapt this questionnaire in an
urban/rural field setting in central Honduras. I discuss the process of creating a culturally
122
123
appropriate translation of the original English questions into Spanish. I also discuss the
steps I took to conduct an ethnographic investigation while I was adapting the scale.
In the third section, I describe the steps I took to administer the scale to a sample of
Hondurans. I describe my iterative, multistage ethnographic approach in which
qualitative and quantitative data were continually collected and analyzed. I also describe
the selection and training of my research team, the development of two different
sampling frames (quota and multi-stage stratified random), the finalization of the survey
instrument and the administration of the survey. Finally, I describe the process of followup ethnographic interviewing which I used to clarify questions that developed during
preliminary analyses of the survey data.
The Childbearing Questionnaire
The Childbearing Questionnaire and Evolutionary Theory
Over the past two decades, Warren B. Miller has developed the Childbearing
Questionnaire to measure the psychological traits of positive and negative childbearing
motivation (Miller 1995). This questionnaire is designed to measure reproductive
motivations, which are “enduring latent dispositions to behave in a particular way under
certain circumstances” (Miller et al. 1999, p. 75). The theoretical framework that Miller
used in the development of this questionnaire is similar to the evolutionary framework
described in Chapters 2 and 3. Miller assumes that, through the process of evolution,
human beings have evolved psychological structures that developed as a consequence of
recurring environmental pressures on survival and reproduction (Miller et al. 1999, p.
58). Reproductive motivation is one of these evolved psychological structures.
Reproductive Motivation does not motivate an individual to have a specific number
of children or have children at particular times in the individual’s lifetime. Instead,
124
reproductive motivation is a general bonding mechanism that is present in all higher
animals, and is especially strong in primates (Miller et al. 1999, p. 58). This bonding
mechanism works at both general and specific levels in the Central Nervous System. It
motivates an individual to form lasting social bonds with other individuals. It also
motivates bonding for specific types of relationships, including the bonding between
infants and their parents, young children and non-caretaking members of their social
group (siblings and other family members, playmates, neighbors, etc.), adults and other
adults in sexual relationships, and adults and their children. The bonding system between
parents and their children is called the nurturant system and is assumed to be the primary
source of human reproductive motivation (Miller et al. 1999, p. 91).
The human parent-infant bonding system is different than in other animals. In nonhuman animals, the motivation to care for an infant is stimulated by direct exposure to
infants. For these animals, sexual bonding leads to reproduction, which produces the
required stimuli to trigger care-taking motivation in parents. For humans, the process is
different. Human beings are capable of being motivated to nurture from knowledge
about the process of reproduction alone. Instead of a reactive childbearing motivation,
humans have a parent-infant bonding system that is both proactive and reactive (Miller et
al. 1999, p. 55).
The Motivational Model Behind the CBQ
The theoretical perspective that Miller used to design the CBQ does not assume a
direct connection between motivational traits and behavioral outcomes. In Miller’s
theoretical framework, motivation is the first step in a five-step process leading to a
reproductive event. Miller labels this theoretical framework the T-D-I-B psychological
model. Figure 6.1 visually depicts this model. The T in the model represents the
125
psychological traits of the individual. Miller defines these traits as “neurally stored
potential for feelings, thoughts, and action that has been acquired by a biological
organism in the process of growth and development” (Miller 1994, p. 230). These traits
are numerous and sometimes have conflicting influence on behavior. Miller describes
two motivational traits that are associated with childbearing: positive childbearing
motivation (PCM) and negative childbearing motivation (NCM) (Miller 1995, p. 476).
Rather than assume a direct correspondence between genetic and motivational traits,
Miller describes motivational traits as being the end product of both genes and
environment (Miller et al. 2002, p. 4).
Motivational Traits
Desires
Intentions
Behavior
Figure 6-1 T-D-I-B Psychological Model
The D and I in the model represent desires and intentions, respectively. Desires are
an individual’s conscious wishes, wants, or concrete goals (Miller et al. 2002, p. 5).
Regarding reproduction, there are three types of desires: a general desire to have
children, a desire for a specific number of children, and a desire for the timing of
childbearing. There are also desires to avoid childbearing as well as a desire to terminate
a pregnancy if it has already occurred (Miller et al. 2002, p. 6). The development of
desires is influenced by the interaction of all of an individual’s motivational traits, not
just PCM and NCM. For example, general social-bonding or care-taking traits that are
not necessarily associated with childbearing would theoretically interact with positive
childbearing motivational traits and lead to higher childbearing desires. On the other
hand, traits such as autonomy or orderliness have the potential to lower childbearing
motivation. Intentions are similar to desires but represent a different type of conscious
126
state. While motivation is what one would like to do, intentions represent what one plans
to do. Intentions can be thought of as desires constrained by reality (Miller et al. 2002, p.
7). Desires must be converted into intentions before an individual behaves.
Regarding childbearing, one force for shaping desires into intentions is the
perception of the childbearing desires of one’s partner. Figure 6-2 depicts this interaction
by splitting the model into two halves, one for each partner. The childbearing desire of
each partner interacts with the other partner’s childbearing intentions through the
perception of these desires. Figure 6-2 is also a combined male-female model because
reproductive behavior involves the behavior of two people. The B in the model
represents this reproductive behavior. The model could include a final step, making it a
T-D-I-B-E model with the E representing a resulting reproductive event (pregnancies or
births). Reproductive behaviors fall into two categories: 1. contraceptive (where the goal
is to prevent conception) and 2. proceptive (where the goal is to achieve conception).
These behaviors can be further divided into progestational (supporting and nourishing the
fetus) or contragestational (preventing the continuation of the pregnancy) (Miller et al.
2002, p. 8) Because human beings have a long history of fertility regulation in situations
where effective and safe methods of birth control and abortion have not existed, I extend
these behaviors to pro-nurturing behaviors (care for infants) and anti-nurturing behaviors
(direct infanticide or indirect infanticide through neglect) (Harris and Ross 1987).
Miller developed the CBQ to measure the independent effect of childbearing
motivation in the T-D-I-B-E sequence. Because childbearing motivation consists of both
positive and negative motivation, the CBQ contains two subscales: the Positive
Childbearing Motivation Scale (PCM) and the Negative Childbearing Motivation Scale
127
(NCM). These scales are designed to measure the respondents’ explicit self-rating of the
desirability of a comprehensive list of positive and negative childbearing experiences. In
previous studies, the scales were found to be both valid and reliable (Miller 1995, p. 479)
Male Motivational
Traits
Male Desires
Perceived
Desires of
Female
Partner
Perceived
Desires of
Male Partner
Female
Motivational Traits
Female
Desires
Male
Intentions
Couple
Behavior
Female
Intentions
Figure 6-2 T-D-I-B Psychological Model with Partner Interaction (Source: Miller et al.
2002).
Miller also developed several sub-scales for the PCM and NCM. Miller developed
these sub-scales using a mixture of statistical (factor and cluster analyses) and qualitative
analyses to group together individual items by relevant themes (Miller 1995, p. 476). I
have listed the entire set of CBQ items, including groupings by sub-scale, in Appendix A.
The Potential Cross-Cultural Application of the CBQ
Although the CBQ was originally developed and tested on a very homogenous
population (European-Americans, 20-40 year old, middle class men and women with
either one or no children living in Silicon Valley California) the intent behind the scale
was to develop a measurement of a human universal motivation for childbearing (Miller
et al. 2002, p. 13). Miller acknowledges that the wording of the questions may have to be
adjusted to use the scale in other cultural settings. He also acknowledges that additional
128
scale items may be necessary to account for different childbearing factors in other
populations. However, Miller maintains that there is justification for assuming that the
two dimensions of positive and negative motivations towards childbearing will exist in
some form across cultures (Miller 1995, p. 483).
This assumption is only now being tested in settings that are culturally divergent
from Miller’s original sample. In this chapter, I detail my adaptation and use of this scale
to measure childbearing motivation in central Honduras. In addition to the Honduran
sample, the CBQ also has been used on a sample of Iranian men and women (Miller et al.
2002). The extreme homogeneity of the original longitudinal sample was a strategic
choice by Miller to allow for the testing and development of the model without a cultural
diversity confound. Although this choice is reasonable, so far the cultural homogeneity
of populations sampled in tests of the CBQ have been assumed rather than empirically
demonstrated. The analysis that I will report in Chapter 7 represents the first empirical
test of the construct validity of a culture of childbearing motivation for a particular
population.
Empirical Evidence Supporting the T-D-I-B-E Model
Empirical tests on the original sample validate many aspects of the T-D-I-B-E
model. The best predictor of reproductive behavior was intentions (Miller et al. 2002, p.
8). Desires were the best predictor of intentions, but had little effect on behaviors,
independent of intentions (Miller et al. 2002, pp. 7-8). The explicit self-ratings of
motivation (measured by the CBQ) produced the best prediction of childbearing desires
(Miller et al. 2002, p. 6). Miller and colleagues have also tested the effect of implicit
motivation (which can be measured by techniques asking respondents to write
129
imaginative stories which are then systematically coded for analysis) and found them to
be stronger predictors of contraceptive behavior (Miller et al. 2002, p. 5).
Other factors not related to childbearing motivation, such as age, family
composition (e.g. number of siblings), and social network composition have been found
to have a predictive effect on the various model components (Miller et al. 2002, p. 7).
For example, the employment status of the respondent’s mother was found to be
particularly predictive of childbearing motivation. This demonstrates that, although this
model and this instrument are focused on the enduring psychological antecedents of
childbearing behavior, these traits continually affect behavior through environmental
factors.
An important empirical finding that has ramifications for an evolutionary
perspective on fertility behavior is that the two scales, PCM and NCM, were found to be
orthogonal to each other (Miller 1995, p. 483). This means that positive and negative
motivation were found to vary independently of each other. This demonstrates that these
two scales are not mirror image measurements of one underlying variable labeled
childbearing motivation. They are each independently measuring two different
motivational traits. This means that there is a possibility for various combinations of
positive and negative childbearing motivations. Combinations include high positive
motivation coupled with low negative motivation, high positive motivation coupled with
high negative motivation, low positive motivation coupled with low negative motivation,
or low positive motivation coupled with high negative motivation. This offers the
possibility that low fertility may not necessarily be the result of low positive childbearing
130
motivation. Negative childbearing motivation may be sufficiently high to suppress the
effect of positive motivation on childbearing.
Genes and Childbearing Motivation
Miller et al. (1999) offer a test of the hypothesis that there is a genetic influence on
childbearing motivation. Miller et al. tested the connection between genes and positive
and negative childbearing motivation. The genes that Miller et al. chose for this study
have been linked to neurotransmitters that are associated with mammalian nurturant
systems and parental behaviors (p. 81). Also, other studies have found that these
neurotransmitters influence the release of several hormones associated with the regulation
of mammalian reproductive behavior (p. 82). Miller et al. found that there were genes
that predicted PCM, genes that predicted NCM, and genes that predicted the intensity of
both (p. 89). They also found that these genes primarily predicted the motivational
aspects of the model but not desires or intentions. Interestingly, they found very few sex
differences in these associations (p. 90).
This evidence of a genetic connection to both positive and negative childbearing
motivation is another important indicator of the non-deterministic connection between
evolution and fertility rates. Although a theoretical connection between genes selected
for in the EEA and a positive motivation to bear children is entirely consistent with
evolutionary theory, there is no necessary connection between this motivation and
specific reproductive behaviors. Life history theory asserts that evolution would have an
effect on the timing and quality of childbearing just as much as the quantity of
childbearing. Depending on environmental conditions, prevention or delay of
childbearing is a direct way of assuring higher quality for children already born or those
to be born in the future. Thus, the genetic connection to negative childbearing motivation
131
as well as positive childbearing motivation is entirely consistent with what would be
expected under evolutionary theory.
Miller’s model is also consistent with Pott’s (1998) variability selection hypothesis
discussed in Chapters 2 and 3. Genes influence both negative and positive childbearing
motivation, which interact with each other and environmental variables to influence
reproductive behaviors. Reproductive behavior variability–having more or less children
and investing more or less in each child–is assured through both genetic and
environmental variability. Thus, a genetic influence on reproductive behavior does not
imply a genetic determination of a specific level of fertility. Negative (or positive)
childbearing motivation is more likely to be expressed under certain environmental
conditions.
Some critics have stated that a problem with the suggestion that there is an
important genetic influence on human reproductive behavior is that human groups share
roughly the same set of genes. These genes, through natural selection, would consistently
promote high fertility. They argue that because human groups do not all express high
fertility behavior, cultures and environments are more important in determining fertility
rates (Harris and Ross 1987, p. 11). However, the theoretical model and empirical tests
of Miller and colleagues contradict these assertions. For practical purposes, genetic
variation among human groups may be considered irrelevant in human behavioral
studies. However, within human groups, there is important variation on the individual
level. Sexual reproduction is a means of maintaining genetic variability among members
of a species. Considering life history theory and the extreme variability in the EEA, it is
adaptively important to maintain variable genetic influence on childbearing motivation.
132
With this perspective, antinatal behaviors cannot be used to justify dismissal of the
genetic influence on human reproductive behavior.
Theories of the genetic influence on human reproductive behavior can only be
falsified through empirical testing. Miller’s motivational model provides the framework
through which tests of the independent effects of genes and environment on childbearing
can be conducted. Miller and colleagues have produced several tests of this model in
which the independent effects of genetic and environmental variables are measured.
The effect of culture on individual childbearing motivation has so far been simply
assumed in order to facilitate the testing of the model. In the original tests (Miller 1995),
the choice of a homogenous sample (in ethnicity, socio-economic status, age and parity)
was assumed to control for culture. In a later test (Miller et al. 2002), a comparison of
samples from California, Honduras and Iran were compared for possible cultural
differences in childbearing motivation. The assumption of cultural homogeneity in the
one test and cultural heterogeneity in the other seems reasonable but testing the
assumption would produce more credible results.
In the next section, I describe the process of data collection I undertook in order to
empirically demonstrate the existence or non-existence of homogeneity in natality culture
on a sample of people from central Honduras. In Chapter 7, I describe the results of the
analysis of these data.
Adaptation of the Childbearing Questionnaire in Central Honduras
Research Location
The area I chose to conduct fieldwork on natality culture was the city of Catacamas
in the departamento of Olancho, Honduras. I lived in the city for eleven months
beginning in July, 2000 while I was collecting data. I interviewed people who lived in
133
the city and in the surrounding rural areas. I chose this area for several reasons. First, I
wanted to understand the cultural differences in natality that were suggested by the
findings I discussed in Chapter 5. The findings suggest a cultural difference in gender
relations and attitudes about fertility for people living in rural areas compared with
people in urban areas. Catacamas is a good place to study this issue because it is an
urban center of a mainly rural, agricultural region. It offered convenient interviewing
access to both urban and rural people.
Another reason I chose this field site was because I had lived in the area while
conducting fieldwork on another project in July of 1998 and was familiar with local
gender and fertility conditions. In this previous research, I conducted surveys with
people about their reproductive histories and I was struck by the variety of family
compositions. I was particularly struck by the differences between those living in the
urban area and those living in the adjacent rural areas. While many urban families were
composed of two or three children, the rural families were typically composed of six,
seven, eight, or more children.
This previous research also put me in contact with many local health workers,
including medical doctors, and health organizations working in various capacities
throughout the city. Many of these health organizations were working on reproductive
health projects and were enthusiastic about facilitating my research. These organizations
typically serviced both the city population as well as the populations living in the
surrounding rural areas. These contacts made the prospects for completing a rural-urban
comparison on reproductive issues plausible.
134
From this field visit, I also realized that Catacamas was an interesting place to
explore issues relating to gender. Catacamas is located within the departamento (which
is something like a county) of Olancho. Olancho is one of the largest departamentos in
Honduras and is one of the most rural and least densely populated. From my previous
field stay, I also learned that Olancho was known throughout Honduras for its high level
of machismo. The primary local interpretation of machismo was that it indicated the
belief in the superiority of men over women, especially in household decision-making.
In some preliminary discussions with people about fertility, I heard machismo blamed for
unmet need for family planning. I heard that there were women who did not want to have
a large number of children but they did because their husbands wanted more children.
They were afraid of confronting their husbands about this because they were not
supposed to question a man’s authority. Also, I heard that many men were celosos–
jealous of their wives’ sexuality–and wanted them to be pregnant all the time so that they
would not have sex with other men. However, although I was told that machismo was
common, I also heard many people, including men, condemn it as a social evil. This
made Catacamas seem like a fruitful location to study fertility and its relation to shifting
notions of masculinity and femininity.
Instrument Translation
The first step I took in the translation process was to have the instrument translated
into Spanish by a native Spanish-speaker before I went into the field. In order to have
more confidence that the choices made by the translator captured the original intent of the
English, I had another native Spanish-speaker translate the Spanish instrument back into
English. This allowed me to see if there were any significant shifts in meaning occurring
during the translation process (Brislin 1986; Weller 1998). I discussed the items that
135
changed meaning with the translators and we came to an agreement about what word
choices would be the best.
The next step was to consult with more bilingual, native Spanish-speakers living in
Catacamas. I asked two medical doctors familiar with conducting surveys to read the
questionnaire and offer suggestions about wording changes. I explained the objectives of
my research to them, which helped improve their suggestions. Both doctors had lived in
the area for around ten years but grew up in Western Honduras. This gave them an
interesting perspective on the people in the area. They are both Honduran, but because
they grew up elsewhere they had an outsider’s perspective on the local population. They
gave me many practical suggestions on wording changes as well as tips on conducting
surveys in the area. We also had many long conversations on the behavior and attitudes
of the people in Catacamas.
Ethnography and Scale Adaptation: Instrument Pre-Test
Determining cultural validity of scale items
Once I had a working translation of the survey, I tested it on 57 people. Because I
knew that many people in the area (urban and rural) were not literate, I conducted each
interview verbally. Although I had many bilingual people contribute to the choices in
wording of the Spanish questionnaire, there was still the potential that people would
interpret the questions differently than the orignal intent of the scale’s author. The people
who helped me translate the survey were all well educated and familiar with survey
questions. Many of the people I planned on interviewing were not as educated and did
not have as much experience with surveys as the translators. Therefore, the only way to
make sure that the questions were making sense was to administer the questionnaire on a
test sample and probe for potential misunderstandings.
136
Besides finalizing the wording of the questionnaire, I also used the practice
interviews as a means of collecting ethnographic data on natality culture. Scale
adaptation and participant observation were the two methodological strategies I used
while living in Catacamas from July 2000 until June 2001. Both of these strategies
continually informed each other during the data collection process. While I was
conducting interviews, I was living in a neighborhood of Catacamas, making friends with
my neighbors, participating in their lives, interacting with their children, and watching
them interact with their children. I was continually taking field notes to record my
interactions in the community. At the same time, I was conducting interviews using the
CBQ translation. The scale adaptation and structured interviewing helped me to
understand explicit natality culture while the participant observation methods I used were
helped me to understand implicit natality culture (DeWalt et al. 1998). When I learned
something about one type of culture, it improved my understanding of the other type of
culture.
I conducted the pre-test interviews as semi-structured ethnographic interviews. My
main objective of this stage of data collection was to collect ethnographic information on
natality culture. Although I was asking survey questions, my objective was not to make
quantitative comparisons among respondents. I treated the people that I interviewed in
the pre-test as ethnographic informants. I decided to do this as a consequence of my
experience conducting structured interviews in Catacamas on other projects. I noticed
that many questionnaire respondents were more than eager to add qualifications to their
responses to structured questions. This is common for people who are not used to the
brusque style of a structured interview. I noticed that if the survey respondents did not
137
realize that the goals of the interview (to record short direct answers to questions) they
often attempted to engage me in a conversation about the question topics. When the
objectives of data collection are to produce data that can be used to make reliable,
quantitative comparisons, this extraneous conversation is a problem. However, I could
not help but notice that I learned a lot about their culture through their discussion of the
questions. I also learned which questions made sense, which questions did not, which
questions were unnecessary and what questions were not being asked. For this stage of
the project, I decided to simulate the structured questionnaire setting but encourage the
extraneous information exchange so that I could make qualitative observations about the
cultural models the informants were using to generate answers to the survey questions.
When I asked a question and the respondent gave me more than a straightforward
answer, I recorded the additional conversation in notes underneath the question on my
survey sheets. I found that I could use these notes to help me remember most of the
additional conversation and write more detailed notes after the interview. Once each
interview was completed, I sat for roughly thirty minutes and wrote out a narrative
description of the interview. The description included the additional information I
received during the interview as well as any overall impressions I had of the interview.
At the end of the day, after I had completed interviewing, I typed the notes I took during
the day into an electronic file. While I was typing, I took the opportunity to expand on
these notes with overall impressions about how the interviewing process was proceeding.
While I interviewed, I also watched for interesting non-verbal responses to the
questions. If the respondent laughed, smiled, looked surprised, looked shocked, looked
angry, or paused to think, I would ask them to explain their reaction to the question. I
138
recorded verbal and non-verbal reactions in notes on my questionnaire sheets and used
this data in my evaluation of the validity of the survey questions.
This process helped me evaluate problems with the wording of particular scale
items. I was able to recognize questions that had wording problems because these were
questions that the respondents often asked me to clarify. If they either seemed confused
by the question or they specifically asked for more clarification, I would explain the
question to them in other words. Once they recognized the concept that I was describing,
they would answer the question. When this would happen, I noted the wording variation
that made it more understandable. Sometimes the respondents would offer their own
wording as a way of confirming my question. I would note these suggestions as potential
changes to the questionnaire for future rounds of the survey.
Modification of the CBQ
This qualitative data collection process helped me to think of additional scale items
that should be included in the questionnaire. Originally, I had intended to conduct these
test interviews in two phases. The first phase was going to be a more open semistructured interview in which I asked the respondents to tell me all of the positive and
negative aspects of having children. This is a validation strategy in which the scale items
are generated from scratch (Brislin 1986; Weller 1998). If the respondents mention the
items on their own without prompting, there is more reason for confidence that the
questions are valid indicators of the overall concept that is being measured (such as
positive and negative aspects of childbearing).
However, I used this strategy only sparingly because I found it difficult to draw out
a list of items that fit the abstract categories of positive and negative childbearing
motivations. In a previous study, I was successful in using a modified freelisting (Weller
139
and Romney 1988) technique to create lists of positive and negative aspects of
childbearing (Kennedy 2002). This research was based on 26 semi-structured interviews
conducted in Oaxaca City, Mexico in 1997. Each interview lasted around an hour. In the
data collection process, I had difficulty drawing out a list of positive and negative aspects
of childbearing from a straightforward freelist prompt. The respondents were usually
inclined to say that it is natural to have children and children are expensive, but then stop.
However, I found that respondents could easily list many positive and negative aspects of
childbearing during the course of a conversation about children. This prompted me to
engage the respondents in a conversation, record it, and then create the freelists from
transcriptions of the recordings.
Because I interviewed a different population of people for the CBQ testing, I
decided to see if the traditional freelisting technique would be more successful than my
Mexican study. I experimented with a few different freelist prompts (“Why do people
have children”, “What are some benefits to having children”, etc.), but the results were
the same. Therefore, I decided to use the questionnaire items as conversation prompts
and then probe the respondents for additional aspects of childbearing motivation not
included in the original questionnaire. When I started asking people to rate the
desirability of the positive aspects of childbearing or the undesirability of the negative
aspects, they often would started describing their general feelings about childbearing. I
listened for interesting anecdotes, metaphors, rationalizations, contradictions, etc. and I
probed them for more information about things they thought were important for me to
know.
140
The respondents often reacted to the scale questions by describing their feelings
about concrete examples from their own lives. It seemed to be much easier for them to
place their feelings about childbearing in a specific, personal context rather than to
describe their feelings about hypothetical or abstract childbearing situations. When the
respondents talked about their own experiences with positive and negative aspects of
childbearing, I listened for items that were not on the original scale but were culturally
significant and deserved to be added to the questionnaire.
Before I began testing the CBQ, I added items that came up during my Mexican
interviews. The additional items that I added can be found in Appendix 1. Because these
items were not on the original CBQ scale, I paid particular attention to the respondents’
reaction when I read these questions to see if they seemed relevant to positive and
negative childbearing motivation.
I concluded that the items in the original CBQ with the addition of the items I
added from my previous study comprised a comprehensive list of positive and negative
aspects of childbearing motivation. Many of the items were instantly recognizable to the
respondents and it seemed obvious to me that they knew exactly what I was talking
about. Sometimes they nodded their heads as they answered the question or they started
to answer before I finished reading every word of the question. This was an indication
that the question pinpointed an emotional connection the respondent had with that aspect
of childbearing. A few items did not seem to make much of a connection to many of the
respondents. I discuss individual items later in this chapter and in Chapter 7.
I did not add any additional items to the CBQ besides the items I added from my
Mexican research. I did not see enough evidence that there were any significant items
141
missing from the CBQ supplemented by the additional items I added before interviewing
began. However, there is the potential that the data collection method–using structured
questions to prompt qualitative responses–could have suppressed the elicitation of
additional items.
A significant change that I made to the presentation of the CBQ was a change in
the scale’s four point rating system. The English version of the CBQ asks respondents to
rate the desirability of the positive scale items as very desirable, moderately desirable,
slightly desirable, or not desirable. The negative scale items ask respondents to rate the
undesirability of the scale items as very undesirable, moderately undesirable, slightly
undesirable and not undesirable (Miller 1995, p. 476). After consulting with local
translators, I decided to experiment with a five-point scale. The reason for this is because
we could not decide which Spanish words would be the best substitutions for slightly,
moderately, and very. We decided to use the words un poco, mas o menos, mucho, and
bastante to represent four different levels of desirability or undesirability.
Another consideration was that the original CBQ was administered as a paper
questionnaire, but I administered it orally. Many people living in Catacamas were
illiterate. The five point scale allowed me to present the scale with a visual guide: I put
my hand up with the palm facing them and I pointed to my thumb for “not desirable” (no
le gusta), my index finger for the first level of desirability or undesirability (le gusta un
poco), and so on.
However, I was soon convinced that the informants were not reliably distinguishing
between un poco or mas o menos nor were they distinguishing between mucho or
bastante. I would sometimes repeat the answer I was given and the respondents would
142
say, “Yes” and nod their heads but then say one of the other choices, as if they were the
same thing. There were a few respondents who perfectly grasped the distinction between
the response categories. Some of them even gave me the number of the response
(0,1,2,3,4) that was written on my questionnaire sheet rather than say the words
corresponding to that choice because they were more highly educated and were familiar
with standardized tests and questionnaires. However, because I had no confidence that
the majority of respondents understood the need for fine distinction in their answers, I
decided to make the scale responses as basic as possible. I decided against using the fivepoint scale and reduced the original four-point scale to a three-point scale: not desirable
or undesirable (no le gusta, no le molesta), somewhat desirable or undesirable (le gusta
mas o menos, le molesta mas o menos), and very desirable or undesirable (le gusta
bastante, le molesta bastante).
Pre-Test sample
I chose the sample of people to pre-test the scale questions using a convenience
sampling strategy. My strategy in selecting people for the pre-test was to maximize the
variety of types of people I interviewed while minimizing the cost and time necessary to
complete a sufficient number of test interviews. After consulting with my local contacts,
I decided to interview people who were in waiting rooms at various urban and rural
clinics. The clinics were run by organizations that were familiar with my project and
gave me approval to approach people in their waiting rooms. They told me that the
people sometimes have to wait several hours and many of them would be happy to have
something to do while they waited. Each day, I went to the clinics and approached
various people and asked them if they wanted to participate in my study. Most of the
people that I approached said yes, but there were a few people who refused. I tried to
143
avoid approaching people who were obviously ill. I often interviewed people who
accompanied someone else to the clinic and waited until their companion finished
consulting with a doctor.
This strategy was successful in generating a good amount of qualitative and
quantitative data in a short period of time. The clinics were good locations to interview a
wide variety of people without having to spend an excessive amount of time and energy
traveling all over the area. The population of my study was the entire Catacamas region,
including the urban inhabitants as well as those living in the outlying rural aldeas
(villages). Encountering a mixture of urban people to interview was not difficult.
However, encountering a large number of rural people from various locations would have
occupied a significant amount of time and energy. Some of the rural areas had good
clinics, but most of them either had clinics without adequate staff and equipment or no
clinics at all. Each day, a significant proportion of the patrons of the urban clinics were
from the surrounding aldeas. Because Catacamas was the center of a large rural region
that had very poor health infrastructure, the visitors to the clinics came from a variety of
locations. Because I had no way of knowing if the rural visitors to the urban clinics
differed from those who did not travel to the city clinics, I also conducted some
interviews in two rural clinics as well.
The urban visitors came from a variety of neighborhoods. I interviewed at two
different urban clinics. One was a clinic run by a local non-governmental religious
organization. This clinic was not free, but it was less expensive than the clinics run by
independent doctors. The doctors who worked in the clinics were well respected in the
area. People who were able to afford more expensive, private clinics often visited these
144
clinics instead. The combination of low cost and good reputation gave the clinic a
diverse clientele, which aided my goal of interviewing a wide variety of people.
However, because the clinic was religiously affiliated (non-Catholic, evangelical
Christian) and was not free, I had a suspicion that I might have some form of bias if I
interviewed in that clinic alone. To counteract this, I also interviewed in the city’s state
run clinic, which was free. This public clinic also had a large waiting room, which was
usually full every day and offered the same interviewing opportunity as the private clinic.
I received permission from the clinic staff to interview people in their waiting room.
To maximize the variability the types of informants I spoke with, I included a short
demographic questionnaire at the end of the scale questions. I asked questions that
measured age, wealth, religious affiliation, marital status, parity, educational background,
family composition, and employment. Each day after interviewing, I entered this data
into an electronic file. I reviewed univariate and various combinations of bivariate
statistics for these demographic categories daily. This helped me determine the potential
biases of my sample, which I attempted to correct with subsequent interviews.
For example, mid-way through interviewing I noticed that my sample was skewed
towards young women and older men. I realized that the clinics had a large amount of
young women in the waiting rooms, probably because many of them were there with
their young children or were waiting for gynecological exams. Most of the men who
came to the clinics were older men who came to the clinic with their own health problems
or accompanied family members. Not many young men came to the clinics. People told
me it was seen as unmanly for a young man to see a doctor. They would not go to a
clinic unless it was absolutely necessary. They also were not usually at the clinics
145
because they could not take time off from work. There were older women at the clinics,
but I realized that I tended to avoid interviewing them if I had another type of person
available to interview. After I analyzed these statistics and saw what my interviewing
tendencies were, I was able to direct myself towards the categories of people that were
not equally represented in my sample. I made sure that if I saw a young man or an older
woman at a clinic, I would make a special effort to ask them for an interview. Using this
strategy, I was limited some of the bias that would have unknowingly entered my nonrandom sample selection.
Because I was not trying to make quantitative comparisons among the informants
in order to make generalizations about the entire population, having a representative
sample was not important in this stage of my research. However, it was important to
limit under-representation of important categories of people because of the possibility
that they might represent a divergent viewpoint on natality culture. Making the
demographic characteristics of my convenience sample explicit also helped me in my
qualitative analysis of the qualitative data I collected. Although the sample was not
representative, knowing the overall demographic characteristics of the sample enabled
me to factor this in to my analysis of the comments made about childbearing.
The non-random nature of the pre-test data collection allowed for some important
flexibility in the direction of sample selection. As I tested the scale I noticed that rural
women gave interesting answers that diverged from the rest of the sample on certain key
variables. Towards the end of the pre-test process, once I had enough people in the
various categories that I decided were important (young men, older women) I oversampled rural women in order to explore their answers to certain questions in more detail.
146
Scale Administration
Data collection proceeded in six stages. The first stage involved building a
research team to do interviews, help determine the sample selection process, locate
interviewers, modify questionnaire instruments, transcribe interviews, and to help me
with my preliminary ethnographic and survey analysis. The next stage involved training
these interviewers to conduct the survey. The third stage was an administration of the
questionnaire to a quota sample of respondents. The fourth stage involved reinterviewing a small sub-sample of the respondents. These re-interviews were
unstructured, formal interviews and were tape recorded and transcribed. The next stage
was another administration of the questionnaire on a randomly selected sample. The
sixth and final stage was another round of unstructured re-interviews of survey
respondents.
Building a Research Team
The first step in my process of administration of the modified CBQ was to
assemble a research team. During the entire pre-test, I was the only interviewer
collecting data. I realized that the information I was able to collect was probably limited
because I was not Honduran and was not a native Spanish speaker. The ethnographer’s
status as an outsider is often to her or his advantage because nothing in the new culture is
taken for granted and informants are more willing to explain mundane aspects of their
lives with an outsider (Spradley 1979, p. 50-51). However, there are many practical
aspects of field research and scale adaptation that require insider expertise. Although my
Spanish was competent enough for me to converse with people about the questionnaire,
changes in the wording of the scale items usually required the expertise of native-Spanish
speakers who were familiar with locally used expressions. Also, there were many
147
practical aspects of interviewing that required assistance from someone who lived in the
local area. Determining such things as the locations and the best times to locate people
for interviewing, and the sensitivity of my questions required local knowledge that I did
not have. I also wanted to match interviewers and interviewees by sex to minimize the
bias that might result with a man asking a woman (or a woman asking a man) questions
about reproductive issues.
Following the suggestions of Handwerker (Handwerker 2001, pp. 253-259), I also
wanted assistants to fulfill the additional role of key informants. I was planning a multistage iterative data collection process that depended on analysis in the field. Subsequent
stages of data collection depended on analysis of the data from the previous stage. The
most efficient way to do this was to have people working on my project with me who
would be available to answer questions about what I was learning in the field. I wanted
to have people who knew my project well to discuss the findings that surprised or
confused me. I also wanted someone to help me make decisions about data collection
procedures that would build on these preliminary findings.
I hired two interviewers, a man and a woman, who were local health workers
experienced in survey administration and were interested in my project. The
administrators of the local clinic where I did my interviewing introduced me to my
assistants after I asked them if they knew of anyone who would be interested in working
on my project. I presented my project to them and had a discussion about the issues I
was investigating. It was clear from my discussions that they were both very interested in
the issues of fertility decline and changing gender roles. They both had both worked as
148
interviewers on past reproductive health surveys in the area and were very knowledgeable
about survey administration.
Research Team Training
Although these assistants were experienced interviewers, we had an extensive
period of training before beginning the survey administration. The training period was
important for them to understand my goals and expectations. It also helped me get to
know them personally and to develop a friendly teamwork relationship. This relationship
was important because I wanted to learn as much from them as I was learning from my
other informants. The best way to learn from them was to have ongoing friendly
conversations about the work they were doing.
The first step in the training process was to revise the questionnaire. I asked them
to read the questionnaire and to identify problems with question wording. I also wanted
them to look for problems that might hinder the administration of the questionnaire.
They also clarified issues that came up while I was administering the pre-test
questionnaire. I asked them to clarify questions that confused the respondents and
questions that generated answers that I was not expecting.
The respondents seemed to interpret some of the questions differently than the
original English wording. I explained to my assistants my interpretation of the questions
to help them understand what the question meant to me. For example, one of the most
problematic questions asks about the desirability of fulfilling one’s potential by having
children. The reaction I received most often was an immediate “No.” This often
confused me because the respondents were otherwise very favorable towards other
aspects of childbearing. I came to realize that they were interpreting this as a question
about their desire to fulfill their potential to have children (having as many children as
149
physically possible). I tried to change this wording to reflect my understanding of the
question but I was not able to find a suitable alternative. I discussed this question with
my assistants. After much discussion of the potential wording options we thought of a
successful wording change that reduced ambiguity.
During these discussions of wording changes, we discussed gender relations and
fertility preferences in the Catacamas area. These discussions were usually very
animated because the topic was obviously very interesting to all of us. These
conversations were also helpful in making clear the issues behind the questionnaire and
helped improve their understanding of the importance of the data collection and analysis.
We had several brainstorming sessions of questionnaire analysis and project planning
before they began to administer the questionnaire.
I decided to have them begin practicing administration of the questionnaire the
same way I conducted the pre-tests. They went to the same clinics where I interviewed
and approached people waiting in the waiting rooms. At first, I listened to them conduct
their interviews. I took notes about how they presented the study, how they read the
introduction and the questions, how they interacted with the respondents, and how they
coded the answers. After I watched them conduct a few interviews each, we discussed
problems or potential problems with their interviewing. When I was confident that they
understood what I wanted from them, They found people to interview on their own.
They both conducted ten practice interviews each.
After they finished with their practice interviews, we met again to discuss issues
they noticed while administering the interviews. Before administering the questionnaire,
there were many questions that they did not think had any wording problems. Because
150
they were both well educated, they had no problem understanding the questions on the
questionnaire themselves and thought that the respondents would not be confused.
However, after administering the questionnaire to a small number of people, they saw
problems in how some people interpreted the questions. They realized that certain words
were not commonly used in everyday speech by most of the people we interviewed.
After discussing the problems, we decided on additional wording changes that would
make the questions clearer. Once we cleared up all of the wording problems, we
finalized the format of the questionnaire.
Random and Non-random Sampling
Because the primary goal of my administration of the survey was to measure a
cultural variable (natality culture), there were unique sampling considerations. One of
the assumptions of most classic statistical procedures is that cases are independent of
each other. This implies that the residuals (or errors) of statistical models are not
correlated. If they are correlated, some type of confound exists, such as spatial or
temporal autocorrelation, and must be controlled for in order to improve the model’s
predictive power. However, cultural data by definition is non-independent because
culture is created through social interaction. If there is a culture for a domain in a
population, measurements of individuals on this domain are going to be correlated.
Therefore, understanding the correlation is the goal of ethnography and controlling for it
is undesirable (Handwerker and Wozniak 1997, p. 874).
Classical statistical tests assure independence for non-cultural data through random
sampling. Handwerker and Wozniak (1997) demonstrated that when using techniques
specifically designed to measure cultural data (such as those I discussed in Chapter 4),
random sampling produces the same results as convenience samples made up of experts
151
in the culture. Random samples are important in the study of culture only when
estimating the prevalence of cultures in a population (Handwerker 2002, p. 99).
However, until cultural boundaries are defined and understood, random samples can be
counter-productive. In an exploratory phase of cultural research, it is more important to
locate and interview individuals who can best help researchers understand the
characteristics of cultural boundaries.
I administered the modified CBQ questionnaire in two phases using both random
and non-random sampling strategies. The goal of the first phase was to identify the
culture or cultures of natality, to determine the construct validity of the concept of
natality culture, and to understand the meaning of the cultural boundaries. These goals
made it more important to find informants who represented different categories of people
and who would be good informants. I planned to use descriptive statistics techniques to
analyze the cultural data collected in this phase. In the second phase (described later in
this chapter), the analysis goals included inferential techniques in order to develop
generalizations of the entire population.
Quota Sample
In my first step towards developing a sampling strategy, I identified life event
factors that might be causing cultural diversity. My theoretical perspective (discussed in
chapter 2) dictated that sex and rural or urban location were the most important variables.
During my pre-test interviews and my discussions of these interviews with my assistants,
I identified two other variables that might be important in producing divergent cultural
views of natality: age and income. For age, I decided on an age range that matched the
traditional age range of reproductive health studies: 18 to 50 years old. I broke the age
range into two groups: young (18-34 years old) and old (35-50 years old).
152
My research assistants and I chose income as a measure of wealth because it
seemed to be the only variable that would divided people consistently across urban and
rural locations. I originally intended to use a lifestyle inventory that measured wealth by
determining how many items or services respondents had in their homes. Items included
cars, stereos, televisions, indoor plumbing, etc.). However, after lengthy discussions with
my assistants and analysis of the pre-test demographic data, I came to the conclusion that
there were rural people who did not have these items or services yet could be considered
wealthier than urban people who had them. Many rural people who owned large amounts
of cattle were very wealthy by the standards of Catacamas but did not have these items or
services because they lived in undeveloped areas. On the other hand, many poor urban
residents had these services because they were cheap and available. In the end, I took my
assistants advice and used income as the criteria to divide the sample into levels of
wealth. We used three levels of income that roughly corresponded to $0 - $69, $70 $199, and more than $200 per month. My assistants believed that these levels
successfully divided people into the three wealth classes that existed in Catacamas.
I originally intended to measure wealth as an interval level variable by asking the
respondents the exact amount they earned each month. During the pre-test interviews
and during the assistants’ practice interviews, the questionnaire had a question that asked
for this amount directly. However, my assistants and I saw that this was a problem
because many respondents seemed reluctant to give out this information. I noticed that
some respondents seemed embarrassed to say how much they made if it was very little.
People who made a relatively good income also seemed embarrassed when asked this
153
question. Respondents sometimes seem startled that they were asked about their income,
hesitated before answering, glanced around the room, whispered the answer, etc.
In Catacamas, there is an obvious contrast between people who have a lot of money
and people who are very poor. This inequitable distribution of wealth is probably the
cause of the hesitation to answer this direct question about income. During the practice
interviews, I had my assistants experiment with two questions: one direct monthly
income question and a range question. They asked the range question if the respondent
seemed reluctant to give an answer to the direct question. After the practice interviews,
my assistants agreed that this was a better way to ask the wealth question. Therefore, we
measured wealth with an ordinal-level variable.
I developed a sampling frame that divided 200 interviews into 24 categories. Each
category represented a unique combination of the four variables described above. Each
category had a target quota of either 8 or 9 interviews. At the start of interviewing, I gave
each of my interviewers a sheet of paper that had a chart detailing the individual
categories (See Appendix F). I instructed them to locate people to interview that fell into
the different categories. Each day when they returned with their completed paper
questionnaires, I entered the data into an electronic file and calculated how many
interviews in each category had been completed and how many more were left to
complete. I gave an updated chart to the interviewers (along with fresh copies of
questionnaires) each morning before they began interviewing again. The chart informed
them of how many more interviews they needed to complete in each category.
The interviewers needed to draw on their knowledge of the area as well as their
social network connections to locate people to interview in the different categories. Both
154
interviewers had to complete 100 interviews each, with the male interviewer interviewing
100 men and the female interviewer interviewing 100 women. In each 100 interviews, 50
were urban and 50 were rural residents. Finding the right number of people for each of
the age and income categories was more difficult than finding equal numbers of men and
women or rural or urban residents. Once the interview was completed, it was not
difficult to determine which age and income category the respondent belonged to because
the interviewers asked the respondents demographic questions in addition to the CBQ
scale questions. However, the interviewers did not always feel comfortable asking the
respondents for their age or income before beginning the interview, so they had to guess
what category they fit into before approaching a potential interview candidate. Most of
the time they were able to find people they needed to complete the quota for each
category. Both of the interviewers had great familiarity with the people in the city and
the surrounding areas from their work as health providers. They were able to identify
locations where they would be able to find people matching the quota criteria. They also
tapped their extensive social networks in the area to complete the quotas.
Sometimes they guessed about age or income incorrectly and ended up completing
excessive interviews in a few of the categories. Also, some of the categories were more
difficult to complete than others, resulting in some quota categories that had too few
completed interviews and others that had more than their share. Overall, 18 of the 24
categories had at least eight interviews, four had seven, and two had only six interviews.
(See Appendix F for details.)
In addition to locating people with the various combinations of sex, age, income, or
area of residence, I asked my assistants to interview people who might have interesting
155
perspectives on natality culture. I asked them to think about people they know and try to
interview people who have expressed weak or strong childbearing motivations. Also, I
encouraged them to locate respondents who might have an interesting perspective on
gender relations in Catacamas. This non-random sampling process helped us to better
define and understand the boundaries of intra- and inter-cultural diversity.
Combining Exploration with Questionnaire Administration
The phase one data collection had an important qualitative component in addition
to the collection of responses to the structured CBQ questionnaire. This qualitative
component was important because I was still exploring natality culture during phase one.
In the pre-test, I collected responses to the questionnaire, but I did not use this data to
make comparisons among respondents. I used the survey questions as a framework for
conducting ethnographic semi-structured interviews in order to generate qualitative data
on natality culture. The pre-test phase was pure exploration of natality culture and
respondents’ reaction to the scale questions.
In phase one, on the other hand, I collected data to make systematic comparisons
among respondents. During the pre-test, I came to some preliminary conclusions about
natality culture. I saw patterns in the responses to the questions and began to generate
hypotheses. For example, I hypothesized that that there was one pattern of pro-natality
culture with some important intra-cultural diversity. However, the exploratory data I
collected did not lend themselves to the testing of hypotheses. The systematic
administration of the questionnaire in phase one allowed me to go beyond pure
exploration and to test the construct validity of natality culture.
However, I still wanted to have an exploratory component in my phase one data
collection. There were two components to the phase one exploration of natality culture.
156
First, as an experiment, I decided to allow the interviewers to continue the qualitative
data collection techniques I used during the pre-test. I showed them my questionnaires
from the pre-test and described how I asked the scale question and recorded the
unprompted verbal responses in addition to the scale responses. I encouraged the
interviewers to record this explanatory discourse onto the questionnaires so that I would
be able to read it and evaluate it along with the scale responses. The experiment worked
well. The interviewers enjoyed being able to discuss questions with the respondents
rather than simply record the scale responses. This process also involved them in the data
on a much deeper level. Our conversations about the study improved and this deepened
my understanding of natality culture. Each day, the interviewers would bring me their
completed questionnaires and I would record the qualitative responses to the survey
questions in an electronic file.
The second exploratory component was a series of re-interviews I conducted with
respondents from the phase one survey. After analyzing the scale data and conducting
tests of construct validity using principal components analysis (see the discussion of this
method in Chapter 4 and the results of this analysis in Chapter 7), I chose respondents
who were representative of cultural clusters. I also chose respondents based on their
qualitative responses. It was obvious that certain respondents were more interested in the
subject and more verbose. This identified them as good candidates for follow-up
unstructured interviews. At the end of each questionnaire, the interviewers asked the
respondents if they could contact them again. I created a sub-set of respondents who said
that they would not mind being interviewed again and who seemed to be good sources of
157
more ethnographic data. I discussed these respondents with my assistants and they gave
me their opinion of these respondents’ potential as unstructured interviewees.
When we were able to locate the respondents for the follow-up interview, I
conducted a tape-recorded, unstructured interview that lasted up to ninety minutes.
Locating the rural respondents was sometimes very difficult because there was no way to
communicate with them other than driving to where they lived, seeing if they were home
and asking them to do the interview. The level of development in Catacamas was so low
at the time of my study that many urban residents did not have phone service. Phone
service in the rural areas was almost non-existent. Therefore, we identified many
possible respondents and interviewed the ones that we were able to locate. It was easier
to locate the urban residents but it was more difficult to establish a time for an interview.
On many occasions, we found the interviewee’s residence but we were unable to conduct
an interview because they were not at home (even if we had a prior appointment). We
occasionally had to return to the residence several times before we were able to complete
a follow-up interview.
Before each unstructured follow-up interview, I made a brief list of questions for
the respondent based on their questionnaire answers. However, I allowed the
conversations to drift in unexpected directions depending on the topics that the
interviewee mentioned. If they made an interesting point, I would ask them to expand on
that point so that I fully understood what they were saying and did not make false
assumptions.
Another experiment that was successful was my decision to include my assistants
in the interviews. After asking the respondents if they would be interested in being
158
interviewed again, the interviewers recorded a name and address on the questionnaire
sheet. However, addresses used in Catacamas and many other places in Honduras were
not exact. They were usually descriptions of the location of a residence rather than a
street name or house number. These descriptions usually consisted of a relation to a
landmark in the village or neighborhood (e.g. the Baptist Church, the book store, the big
mango tree). To find a residence using this type of address required a working
knowledge of the area, so I paid the interviewers to help me find the respondents and ask
them for an additional interview.
I quickly realized that having assistants with me during these interviews was an
advantage. One advantage was that the respondents usually recognized the assistants
from the interview and usually welcomed us into their homes immediately. More
importantly, having one of my assistants present at the unstructured interview helped me
to gain more knowledge from the interview than I would have gained otherwise. I was
always the primary interviewer, but there were many times that the presence of another
interviewer helped improve the interviewing session. The primary benefit of having an
assistant with me during the interview was the improved communication. Occasionally, I
did not understand something that the informant said and the assistants re-stated it. This
usually created a prompt for the informants to provide more information about what they
just said. Other times, my questions or comments made it clear that I did not understand
something that the informant said and my assistants would pointed this out. This
improved the flow of the interview.
Another benefit was that my assistants were very knowledgeable about my
interviewing objectives and actively participated in the interviews. They understood the
159
types of issues that were important to my research and they looked for opportunities to
draw out this information from the informant. Occasionally, the informant said many
interesting things quickly. Having a second person thinking of questions was helpful in
identifying remarks that were in need of clarification or points of view that needed
expansion. Over time, I developed a good interviewing rapport with my assistants and
we used our combined efforts to maximize the amount and quality of exploratory data
generated from these unstructured interviews.
Scale Administration Phase Two: Random Sample
I conducted phase two of my scale administration on a random sample of people in
Catacamas and its surrounding rural villages. I used a random sample because the main
goal of this phase was to produce explanatory rather than exploratory findings (Johnson
1998, p. 139). My analysis objectives were to use inferential statistics to produce
findings that could be generalized to the entire population. A random sample is an
important element in generalizability of findings because it minimizes the introduction of
unknown bias into the sample selection.
The distribution of the population was very different in the urban areas compared
with the rural areas. There was also significant diversity within the urban area and within
the rural area. Most of the population in the Catacamas region was concentrated in the
city. Most of the city population was concentrated in the center of the city. Some of the
rural villages were more connected to the city of Catacamas than others. The lifestyle in
these villages was more urban than other villages. In order to account for this diversity, I
conducted a multistage cluster sample.
160
Urban Sampling Frame
With the help of my assistants, I divided the neighborhoods in the city into two
groups: the center neighborhoods and the peripheral neighborhoods. The center
neighborhoods were very urban and had a higher population density than the rest of the
city. This is where most of the shop owners, doctors, lawyers and other high-income
professionals lived. This area had many city services, such as city sewage, water,
electricity, phone service, and garbage collection. The peripheral neighborhoods, on the
other hand, had a less dense population with low-income workers and a lack of city
services. Many residences did not have electricity, phone service, city sewage service, or
garbage collection and had infrequent or no city water service. Usually, the roads in
these areas were not paved. In general, the neighborhoods on the periphery had more
rural characteristics than the neighborhoods in the center of the city.
In all, there were 45 neighborhoods in Catacamas, 13 in the center and 32 on the
periphery. One of these neighborhoods extended from the center of the city to the
periphery. We decided to split this neighborhood into a central neighborhood and a
peripheral neighborhood. This pushed the number of neighborhoods to 46. I selected a
simple random sample of five central neighborhoods and five peripheral neighborhoods
using Excel’s random number generation function. From the ministry of health, I
obtained census information about the city neighborhoods. Each neighborhood had a list
of numbered blocks and each block had a list of numbered houses. For each selected
neighborhood, I added up the number of houses and selected four simple random samples
of houses. Two of these samples were for interviewing men and two were for
interviewing women. There were two samples in anticipation of cases where there was
not an eligible respondent (an adult man or woman) living in the residence. If there was
161
no eligible respondent, the interviewers selected the first residence from the secondary
sample.
The number of houses I selected for these samples differed between the central and
peripheral neighborhoods. All of the central neighborhoods had a large number of houses
(range of 98-281). I selected five houses for each of the central neighborhood samples.
For the peripheral neighborhoods, there was a big difference in the number of houses
(range 14–421). In order to account for this large difference in population size, I used a
disproportionate stratified random sample (Bernard 2002, p. 150-151). I originally
planned on having a proportionate stratified random sample, but one of the
neighborhoods I selected had seventy percent of the total number of houses in the
selected neighborhoods while three of the neighborhoods had less than seven percent of
the total houses. I wanted to sample at least two houses in each neighborhood. In order
to do this, I had to lower the number of houses selected in the large neighborhood and
select a disproportionate sample.
Once I selected the houses for the sample, finding the residences and completing
the interview was straightforward. Although numeric addresses were rarely used in
Catacamas, each house did have an assigned block and house number that corresponded
with the ministry of health census data. The average person who lived in Catacamas had
no idea which block or house number they lived in, even though the number was written
somewhere on the outside of their home. These numbers were not very prominent. They
were written with some form of permanent marker and were usually located on gates or
over the front door to the house. My interviewers had experience using this sampling
162
method from previous interviewing projects. They knew where to look on the front of
the house to identify the block and house number.
Once the interviewers found the selected the residence, they asked to speak with an
adult (18 years old to 50 years old) man or woman who lived in the home. If there were
more than one living there, they would select one at random by flipping a coin. If there
was no adult man or woman, or if the adult man or woman refused, they would add a
house from the second sample for that neighborhood. If there was an eligible adult living
in the home but they were not home at the time, the interviewers asked for a better time
to contact them and returned to complete the interview.
Rural Sample
We also divided the rural regions into groups. One of my assistants worked for the
Ministry of Health as an investigator of infectious outbreaks. His work often took him to
the surrounding regions of Catacamas. He was very familiar with the similarities and
differences of the various rural villages. We divided a map of the rural areas surrounding
Catacamas into five sectors. These sectors corresponded to the regional division the
ministry of health used in its investigations. These areas differed with each other based
on terrain–some were in mountainous areas and some were in valleys–their level of
development, and their distance from Catacamas. In order to account for the diversity
represented by each of these sectors, I selected a simple random sample of two villages
from each sector using Excel’s random number generator function.
It was impossible to select a sample of rural respondents before going to the
villages to interview. The ministry of health office in Catacamas did not have census
information for the rural villages. Each village had its own sampling issues and we
planned on making sampling decisions once we arrived to conduct interviews. My
163
assistants knew that some of the villages in the sample had their own regional
government-run health clinics and probably had census information. We decided that
this would be the best sampling frame to use if available. For the villages without census
information, we decided to create our own ad hoc sampling frame once we arrived in the
village.
Three of the five regions had health clinics with census information that provided
us with a sampling frame. The clinics had a sheet of paper for each household in the
village with information about the people who lived there. We selected households at
random by selecting sheets at random. We first eliminated the households without adults
in the 18-50 age range. Next, we shuffled the pages (which were not in any particular
order to begin with) and selected a systematic random sample of the pages (Bernard
2002, p. 146-148).
We selected a sample size for each village that was proportional to the amount of
people in the villages. We noticed that in two of the sectors, some of the selected villages
were very small (around 20 residences). We selected additional villages in those sectors
to add to the sample. For one sector, we added a third village. For the other sector, we
sampled people living in each village because they were all very small. The selected
villages for the other three sectors had enough people living in them to limit the sample to
two villages.
For the villages that did not have census information, we developed a sampling
frame on the spot with the help of a local person. We first went to the village and
surveyed the streets and counted the houses in the area. We then went to the local store
in the village and asked someone working there for information about the houses. We
164
first confirmed with them the total number of houses we counted in case we did not see
all of them or some of the houses did not actually have people living there. We then
asked for the names of the families who lived in each house. We wrote these names on
slips of paper, rolled the paper up into little balls, put them in a hat, and then selected a
sample. We determined the number of houses to select by estimating the proportional
size of that village compared with the other villages in the region.
For each sector, we interviewed ten men and ten women. The same male assistant
who conducted the phase one interviews located and interviewed all of the male
respondents in phase two. Unfortunately, between phase one and phase two, my female
assistant began working at a hospital in another city that was a forty-minute drive from
Catacamas. Because she began working there a short time before the commencement of
phase two, she was not able to take time off to continue working on my project. She
finished work each day in the early afternoon, which allowed her enough time to continue
doing the urban interviews only. The rural villages were sometimes an hour or two away
and, because some of them were accessible only via a four-wheel drive vehicle, we had to
leave early in the morning and work all day to meet our interviewing goals.
Because there was no way that my female assistant could meet this schedule, I
hired another female assistant. I hired another health worker who also worked at the
ministry of health. This assistant was someone I had worked with before on a previous
project. On this project, I trained her to conduct structured interviews and knew that she
was professional and a hard worker. After explaining the questionnaire to her I had her
conduct five practice surveys so that she would not be surprised by the nuances of the
165
questionnaire. Once we discussed and resolved all of her questions, we began the rural
interviews.
Phase Two Questionnaire Modifications
I modified the questionnaire before phase two interviewing. I added several
sections to address issues that came up in the phase-one scale analysis and the
unstructured interviews. I added a section that measured the respondents’ knowledge and
use of family planning methods. This section was used in the two Honduran health
studies that I discussed in Chapter 5 (Monteith et al. 1997; Monteith et al. 1998). I also
included a section about gender roles with questions about reproductive decision-making,
household division of labor, and acceptability of extramarital relations. These were
issues that surfaced during the unstructured interviews. I also added a section that
addressed the perceptions of female employment opportunities. I first made a list of all
the types of employment given as answers on the phase-one survey. I then discussed this
list with my assistants and with their help I modified the list to include a selection of job
types that were typically male, a selection that were typically female, and a selection that
were not necessarily male or female. The phase two interviewers presented the
respondents with this list and asked them if the job types were male, female, or could be
performed by or either men or women. The demographic section of the phase one
questionnaire remained the same with some slight modifications. (See Appendix B, C,
D, and E for the Spanish and English versions of the questionnaires for phase one and
phase two.)
The CBQ questions were not modified from phase one to phase two. However,
because of the added length of the questionnaire, I decided to eliminate some of the
questions that seemed to be unnecessary for an analysis of natality culture. I determined
166
which questions were unnecessary based on my analysis of the phase one data and from
the qualitative responses to these questions. The questions I removed did not contribute
much to the overall variance. Also, these questions seemed to be confusing to the
respondents.
CHAPTER 7
ANALYSIS
The Development of the Hypothesis of One Culture of Pronatality
After the series of ethnographic interviews in the pre-test, I hypothesized that there
was an overall pattern of childbearing motivation. The pattern I expected to find
consisted of high positive childbearing motivation scores and low negative childbearing
motivation scores. In other words, I expected to see evidence of pronatal culture. The
reason I expected this was because, as I was asking the series of positive motivation
questions, many people seemed to draw on an overall childbearing-as-positive-experience
cognitive model. They would say todo (everything) and nod their heads as I asked about
each type of positive experience. They did not appear to be considering each individual
question too deeply. Instead, they anticipated that their answer would be “very desirable”
because they assumed that they thought every aspect of childbearing was a desirable
experience.
I realized that they were not just going through the motions, however, when I asked
them questions that implied that they thought having a large family was desirable (for
example, “having as many children as God sends” or “being the center of a large and
active family”). Many times I noticed respondents giving one positive answer after
another until we asked one of these questions. At that point, they had to stop themselves
from giving another bastante answer, think for a second, and admit that they did not want
to have a large family.
167
168
I noticed the same type of answering pattern for the negative questions, only in the
opposite direction. Many respondents seemed to dismiss the idea of negative aspects of
childbearing. If they were asked about some mundane aggravation, such as putting up
with a child’s noise and disorder or taking care of them when they are sick, they would
reply, “¡Son hijos!” (They’re children!). This implied that they accepted children and the
things that children did as part of life. The daily inconveniences of children were not
something that they considered desirable or undesirable. There did seem to be a sub-set
of NCM questions that evoked “highly undesirable” answers, but these answers only
solidified my belief that there was an overall pronatal culture. The aspects of
childbearing that typically received “highly undesirable” answers were what I describe as
fears about child development. Questions that fell into the “Fears and Worries of
parenthood” NCM subscale (see Appendix A) deal with a parent’s concern that their
child is not developing into a healthy, happy member of society. Even if the respondents
of these questions were part of a highly pronatal culture, worries associated with the
successful development of children would be understandably undesirable.
Although I noticed this overall pattern in the way people were answering CBQ
questions, there were subtle differences in how each respondent answered the scale
questions. Frequently, respondents answered individual questions differently than I
would have expected given my assumptions about an overall pattern of pronatal culture.
Because of the large number of items in the CBQ, it was impossible for me, using
qualitative analysis alone, to determine if these deviations were randomly distributed or if
they represented evidence against one pattern of pronatal culture. In the next section, I
169
discuss my analysis of the phase one CBQ data and the tests for construct validity I
performed using the methods discussed in Chapter 4.
Factor 2
1
0
SEX
Men
Women
-1
-1
0
1
Factor 1
Figure 7-1 Scatter Plot of First Two Factors of PCA for all CBQ Items–Phase One
Construct Validity of Pronatality Culture
I conducted a Principle Components Analysis (PCA) of the phase one CBQ data
using SAS’s “proc factor” procedure with no rotation. I first transposed the data so that it
was in a variable-by-informant matrix rather than an informant-by-variable matrix
(Handwerker 2001, p. 190). This allows for the creation of an informant-by-informant
similarity matrix (Handwerker 2002, p. 113; Romney et al. 1986, p. 322). Figure 7-1 is a
scatter plot of the informants’ loadings on the first two factors of all of the phase one
CBQ data. This graph is a representation of the similarities and dissimilarities among the
170
respondents. Each point on the graph represents an individual who responded to the
questionnaire. Factor 1, the horizontal axis, represents the broadest range of
commonalities among the respondents (Handwerker 2001, p. 190). Factor 2, the vertical
axis, represents the second broadest range of commonalities among the respondents. The
closer the points are to each other, the closer the totality of the individuals’ CBQ
responses are, at least with regard to the first two factors. In figure 7-1, circles are men
and Xs are women.
140
120
59%
Eigenvalue
100
80
60
40
20
7%
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Factor
Figure 7-2 Scree Plot of PCA Eigenvalues for All CBQ Questions–Phase One
Figure 7-2 is a scree plot of the eigenvalues of all of the factors for the PCA
analysis of the CBQ data for phase one. This plot shows the relative importance of the
factors produced by the PCA. The evaluation of the relative importance of the first factor
171
compared with the remaining factors offers a test of the hypotheses that there is an
overall cultural pattern behind the domain under study. If the first factor is large relative
to the other factors, there is evidence of only one pattern of variance in the data, which is
evidence of a culture (Romney et al. 1986, p. 323). In Figure 7-2, there is a sharp fall in
eigenvalues from factor 1 to factor 2. The points on the plot representing factors 1 and 2
have percentages next to them. These percentages represent the amount of variance in
the data that is explained by the factor. Factor 1 explains 59% of the data and factor 2
explains 7 % of the data. The fact that factor 1 accounts for over 50% of the variance and
there are no respondents with negative values for factor 1 gives evidence that there is an
overall pattern in the data. This is evidence that there is a culture of childbearing
motivation (Handwerker 2002, p. 113; Handwerker 2001, p. 190; Romney et al. 1986, p.
323).
Table 7-1 shows the variable factor loadings produced by the PCA sorted by factor
1. Table 7-2 shows the same factor loadings sorted by factor 2. (The definitions of the
variable names that are used in Tables 7-1 and 7-2 can be found in Appendix A). The
PCM questions are displayed in bold. There are many variables that have high loadings
on factor 2 (greater than .5 or less than -.5). This indicates that, although there is
evidence of one culture of childbearing motivation, there is some important cultural
diversity represented by factor 2.
The important detail to notice about Table 7-1 is that Factor 1 almost perfectly
distinguishes between the positive and negative childbearing motivation questions. Most
of the PCM questions are loading high on factor 1 while most of the NCM questions are
loading low. Although it might seem obvious that the positive and negative childbearing
172
questions would tend to receive opposite responses from the same respondent, this is not
necessarily the case. As I discussed in Chapter 6, Miller found that the PCM and NCM
subscales were actually measuring two non-correlated traits (Miller 1995, p. 483). Thus,
there is potential that the administration of this scale in another cultural setting might
produce a different distribution for factor 1 loadings.
Table 7-1 Factor Loadings for PCA of All CBQ Items Sorted by Factor 1–Phase One
VARIABLE
EXITO
UTIL
AMOR
DEPEND
GUIAR
COMPART
APELLID
AYUDAV
COMPLET
TIEMPO
CHINEAR
FERTIL
HEREDAR
VIDA
SOCIED
FUNCION
ADMIA
VARON
JUGAR
CUSTUMB
SATIS
MUJER
AVERGU
FELIZ
FORTAL
HERMAN
NIETO
MOVER
FACTOR 1
1.026
1.011
0.995
0.992
0.984
0.983
0.964
0.958
0.954
0.948
0.946
0.935
0.924
0.923
0.895
0.886
0.861
0.827
0.823
0.823
0.808
0.771
0.756
0.722
0.672
0.645
0.607
0.494
FACTOR 2
0.125
0.035
0.124
0.092
0.031
0.012
-0.006
0.139
-0.100
0.015
0.063
-0.036
0.185
-0.198
0.098
-0.389
-0.332
-0.485
-0.256
0.259
-0.481
0.022
0.146
0.765
-0.875
-0.969
-0.626
-0.110
VARIABLE FACTOR 1
0.372
DEBERES
0.348
DARLUZ
SEGURID
-0.089
VEJEZ
-0.177
-0.193
PEPE
IMPEDIDO
-0.223
-0.284
JEFE
DOLOR
-0.345
INCAPAZ
-0.392
SUFRIR
-0.467
-0.484
TRABAJE
CRIANDO
-0.929
DINERO
-1.103
SALUD
-1.156
OTRONIN
-1.189
DESAGRA
-1.229
RUIDO
-1.317
CARRERA
-1.390
CASA
-1.397
DECUIDAR
-1.471
LIBERTAD
-1.483
CARGA
-1.489
-1.510
DIOS
ECONOMIC
-1.531
ENFERMO
-1.623
EXIGENTE
-1.642
ENERGIA
-1.735
FACTOR 2
0.837
-1.789
3.140
-0.310
2.275
1.689
-1.565
1.303
-2.648
1.261
0.298
2.455
-2.133
-1.198
-0.745
0.648
0.149
0.032
0.402
0.077
0.362
-0.918
-0.903
0.516
0.228
-0.376
-0.335
173
In this setting, figure 7-1, 7-2 and tables 7-1 and 7-2 offer evidence that there is a
pronatal culture. The culture is not perfectly pronatal, however. There are many PCM
items that overlap on factor 1 with NCM items. In a later section, I will analyze the
factor loadings in order to describe the cultural agreement and disagreement about
childbearing motivation.
Table 7-2 Factor Loadings for PCA of All CBQ Items Sorted by Factor 2–Phase One
VARIABLE
SEGURID
CRIANDO
PEPE
IMPEDIDO
DOLOR
SUFRIR
DEBERES
FELIZ
DESAGRA
ECONOMIC
CASA
LIBERTAD
TRABAJE
CUSTUMB
ENFERMO
HEREDAR
RUIDO
AVERGU
AYUDAV
EXITO
AMOR
SOCIED
DEPEND
DECUIDAR
CHINEAR
UTIL
CARRERA
GUIAR
FACTOR 1
-0.089
-0.929
-0.193
-0.223
-0.345
-0.467
0.372
0.722
-1.229
-1.531
-1.397
-1.483
-0.484
0.823
-1.623
0.924
-1.317
0.756
0.958
1.026
0.995
0.895
0.992
-1.471
0.946
1.011
-1.390
0.984
FACTOR 2
3.140
2.455
2.275
1.689
1.303
1.261
0.837
0.765
0.648
0.516
0.402
0.362
0.298
0.290
0.228
0.185
0.149
0.146
0.139
0.125
0.124
0.098
0.092
0.077
0.063
0.035
0.032
0.031
VARIABLE
MUJER
TIEMPO
COMPART
APELLID
FERTIL
COMPLET
MOVER
VIDA
JUGAR
VEJEZ
ADMIA
ENERGIA
EXIGENTE
FUNCON
SATIS
VARON
NIETO
OTRONIN
FORTAL
DIOS
CARGA
HERMAN
SALUD
JEFE
DARLUZ
DINERO
INCAPAZ
FACTOR 1
0.771
0.948
0.983
0.964
0.935
0.954
0.494
0.923
0.823
-0.177
0.861
-1.735
-1.642
0.886
0.808
0.827
0.607
-1.189
0.672
-1.510
-1.489
0.645
-1.156
-0.284
0.348
-1.103
-0.392
FACTOR 2
0.022
0.015
0.012
-0.006
-0.036
-0.100
-0.110
-0.198
-0.256
-0.310
-0.332
-0.335
-0.376
-0.389
-0.481
-0.485
-0.626
-0.745
-0.875
-0.903
-0.918
-0.969
-1.198
-1.565
-1.789
-2.133
-2.648
174
Elimination of Sex-Bias in the CBQ
As I analyzed Table 7-2 in conjunction with Figure 1, I noticed a potential problem
with the results. Figure 1 shows an extreme split between the loadings of men and
women on factor 2. There is almost no overlap between men and women on the vertical
axis. This made me wonder if there were particular questions that were sex-biased and
should be eliminated from the analysis. In Table 7-2, I noticed that there were several
questions that loaded either very high or very low on factor 2 and could be considered
sex-biased. For example, the variable PEPE (“How desirable is breast (bottle) feeding a
baby?”) stood out during the pre-test as a question that consistently received different
responses from men and women. The main problem with this question is that the male
and female versions of the question are very different. Men cannot breast feed, so bottlefeeding is used as a substitute. However, the response from men when they heard this
question was usually highly negative because they considered bottle-feeding a less
desirable option than breast-feeding. They said that it was less healthy than breastfeeding and did not consider the idea of bottle-feeding desirable. Honduras is a very poor
country and many Hondurans cannot afford to purchase baby formula in the first place. I
considered this question to be sex-biased and dropped it from subsequent analyses.
There are several other questions that are similarly sex-biased. The PCM question
DARLUZ (“How desirable is giving birth to a baby (Helping your wife give birth to a
baby)?”) and the NCM questions SALUD (“How undesirable is having a baby who
strains my (wife’s) health?”), SUFRIR (“How undesirable is experiencing (Seeing my
wife experience) the discomforts of pregnancy?”), and DOLOR (“How undesirable is
experiencing (Seeing my wife experience) the pain of childbirth?”) refer to physical
experiences that women have and men do not. Although men can empathize with their
175
wives while they experience pregnancy and childbirth, they can never feel the same
feelings. I eliminated these questions, along with PEPE, from further analysis.
Factor 2
1
0
SEX
Men
-1
-1
Women
0
1
Factor 1
Figure 7-3 Scatter Plot of First Two Factors of PCA for all non-sex biased CBQ Items–
Phase One
Figures 7-3 and 7-4 show the results of the re-analysis of the CBQ data without the
sex-biased questions. Figure 7-3 is a scatter plot of the PCA factor loadings for the
respondents to the CBQ without the sex-biased questions. Figure 7-4 is a scree plot of
the eigenvalues of the factors from the PCA on the CBQ data without the sex-biased
questions.
The results of the PCA without the sex-biased questions are not much different than
the results with the entire set of CBQ questions. Figure 7-3 shows only a slightly
different configuration of points than figure 7-1. The shape of the scree plot in figure 7-4
176
and the consistently high factor 1 loadings continue to support the hypothesis of a strong
pattern of natality culture. Even without the sex-biased questions, there remains a strong
split between the answers given by men and women on factor 2, giving evidence of
important intra-cultural diversity. The first factor explains 64% of the variance in the
answers to the CBQ questions and the second factor explains 6%. The removal of the
sex-biased questions improved the strength of the cultural pattern of natality.
140
64%
120
Eigenvalue
100
80
60
40
20
6%
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Factor
Figure 7-4 Scree Plot of PCA Eigenvalues for Non-Sex Biased CBQ Questions–Phase
One
Tables 7-3 and 7-4 show the factor loadings of the variables from the PCA after
removing the sex-biased questions. Once again, the majority of the PCM variables load
high on factor 1 while the majority of NCM questions load low. There remains a sub-set
of PCM variables that load negatively on factor 1 and a sub-set of NCM variables that
177
load positively. In the next section, I will analyze the agreement and disagreement the
respondents show regarding the desirability of the CBQ questions in order to describe the
pattern of natality culture.
Table 7-3 Factor Loadings for PCA of All Non-Sex Biased CBQ Items Sorted by Factor
1–Phase One
VARIABLE
EXITO
UTIL
AMOR
DEPEND
GUIAR
COMPART
APELLID
AYUDAV
COMPLET
TIEMPO
CHINEAR
FERTIL
HEREDAR
VIDA
SOCIED
FUNCION
ADMIA
VARON
JUGAR
CUSTUMB
SATIS
MUJER
AVERGU
FELIZ
FORTAL
FACTOR 1
0.982
0.966
0.952
0.950
0.940
0.940
0.923
0.917
0.913
0.904
0.904
0.894
0.883
0.881
0.855
0.845
0.823
0.787
0.784
0.783
0.770
0.738
0.723
0.692
0.641
FACTOR 2
0.145
0.033
0.149
0.117
0.048
0.052
-0.016
0.151
-0.092
0.080
0.116
-0.072
0.159
-0.198
0.104
-0.401
-0.317
-0.472
-0.212
0.296
-0.544
-0.025
0.175
0.888
-0.908
VARIABLE
HERMAN
NIETO
DEBERES
SEGURID
VEJEZ
IMPEDIDO
JEFE
INCAPAZ
TRABAJE
CRIANDO
DINERO
SALUD
OTRONIN
DESAGRA
RUIDO
CARRERA
CASA
DECUIDAR
LIBERTAD
CARGA
DIOS
ECONOMIC
ENFERMO
EXIGENTE
ENERGIA
FACTOR 1
0.611
0.576
0.358
-0.079
-0.171
-0.216
-0.275
-0.383
-0.463
-0.893
-1.064
-1.109
-1.139
-1.183
-1.267
-1.336
-1.347
-1.413
-1.424
-1.435
-1.454
-1.470
-1.559
-1.582
-1.670
FACTOR 2
-1.063
-0.650
1.029
3.429
-0.175
2.055
-1.524
-2.649
0.288
2.817
-2.235
-1.171
-0.671
0.787
0.279
0.225
0.591
0.194
0.501
-0.858
-0.875
0.584
0.322
-0.251
-0.236
Description of Culture via Analysis of Factors and Ethnographic Interviewing–
Phase One
In the previous section, I demonstrated the empirical evidence supporting the
hypothesis of pronatal culture among the respondents to the CBQ. However, this
pronatal culture is not absolute. The cultural pattern is more complex than a simple
178
pattern of positive childbearing aspects being rated as highly desirable and negative
childbearing aspects being rated as highly undesirable. There are several positive
childbearing questions that load with negative childbearing questions and vice versa. In
this section, I will discuss in detail the characteristics of the pattern of pronatal culture
through a combination of analysis of the factor loadings and analysis of the qualitative
data I collected while conducting ethnographic interviews.
Table 7-4 Factor Loadings for PCA of All Non-Sex Biased CBQ Items Sorted by Factor
2–Phase One
VARIABLE
SEGURID
CRIANDO
IMPEDIDO
DEBERES
FELIZ
DESAGRA
CASA
ECONOMIC
LIBERTAD
ENFERMO
CUSTUMB
TRABAJE
RUIDO
CARRERA
DECUIDAR
AVERGU
HEREDAR
AYUDAV
AMOR
EXITO
DEPEND
CHINEAR
SOCIED
TIEMPO
COMPART
FACTOR 1
-0.079
-0.893
-0.216
0.358
0.692
-1.183
-1.347
-1.470
-1.424
-1.559
0.783
-0.463
-1.267
-1.336
-1.413
0.723
0.883
0.917
0.952
0.982
0.950
0.904
0.855
0.904
0.940
FACTOR 2
3.429
2.817
2.055
1.029
0.888
0.787
0.591
0.584
0.501
0.322
0.296
0.288
0.279
0.225
0.194
0.175
0.159
0.151
0.149
0.145
0.117
0.116
0.104
0.080
0.052
VARIABLE
GUIAR
UTIL
APELLID
MUJER
FERTIL
COMPLET
VEJEZ
VIDA
JUGAR
ENERGIA
EXIGENTE
ADMIA
FUNCION
VARON
SATIS
NIETO
OTRONIN
CARGA
DIOS
FORTAL
HERMAN
SALUD
JEFE
DINERO
INCAPAZ
FACTOR 1
0.940
0.966
0.923
0.738
0.894
0.913
-0.171
0.881
0.784
-1.670
-1.582
0.823
0.845
0.787
0.770
0.576
-1.139
-1.435
-1.454
0.641
0.611
-1.109
-0.275
-1.064
-0.383
FACTOR 2
0.048
0.033
-0.016
-0.025
-0.072
-0.092
-0.175
-0.198
-0.212
-0.236
-0.251
-0.317
-0.401
-0.472
-0.544
-0.650
-0.671
-0.858
-0.875
-0.908
-1.063
-1.171
-1.524
-2.235
-2.649
Figure 7-5 shows the same graph as figure 7-3 with the addition of descriptions of
the clusters of points. I have given the factor 1 axis the general description of “Pronatal
179
Culture.” Scores on factor 1 represent how well an individual respondent agrees with this
overall pronatal culture. As opposed to an absolute pronatal culture, this pattern of
pronatality is a qualified pronatal culture. Table 7-3 shows that two NCM variables,
AVERGU (“How undesirable is having a child who embarrasses or disgraces the rest of
the family?”) and FELIZ (“How undesirable is having an unhappy and poorly adjusted
child?”) are negative childbearing motivations that were typically rated as very
undesirable. These two variables are in contrast with other negative aspects of
childbearing (e.g. ENERGIA–“How undesirable is spending time and energy involved in
childcare?”) that were not considered undesirable because they were just normal aspects
of childbearing. Table 7-3 also shows three PCM variables that loaded negatively on
factor 1. This indicates that they were not considered desirable aspects of childbearing.
These variables were JEFE (“How desirable is being the center of a large and active
family?”), TRABAJE (“How desirable is having children that work for you?”), and DIOS
(“How desirable is having all the children that God sends you?”).
My analysis of the semi-structured and unstructured interviews confirms that these
variables, which are exceptions to absolute pronatalism, are related. Each question that
does not conform to the factor pattern is in some way related to the conflict between the
quantity and quality of children. As I discussed in the previous chapter, I developed the
hypothesis that there was a pronatal cultural pattern from my analysis of the qualitative
interview data. However, I realized that many of my informants objected to the idea of
having more children than they could support. I also realized that, even though some
respondents stated that they enjoyed all aspects of childbearing and nothing about raising
180
children bothered them, they were in fact bothered by worries associated with raising
successful children.
1
Factor 2
FEMALE
PARENTAL
CULTURAL
DIVERSITY
PRONATAL
CULTURE
0
MALE
PARENTAL
CULTURAL
DIVERSITY
SEX
Men
-1
-1
Women
0
1
Factor 1
Figure 7-5 Scatter Plot of First Two Factors of PCA for all non-sex biased CBQ Items
with descriptions of cultural patterns– Phase One
When I asked them questions that implied a desire for large families, many of them
expressed their objections to this. They often gave me examples of people they knew or
had seen who were poor yet had more children than they could support. The questions
DIOS and JEFE most often invoked this type of reaction. A common reaction to DIOS
was to state that, even though it is a sin to use contraception, it was a worse sin to have
more children than you could support. For the question JEFE, the most common reaction
was to state that they liked the idea of being the center of an active family, but they did
181
not want to have a large family. They did not want to have more children than they could
support.
My interpretation of the reaction to the positive question TRABAJE and the
negative questions FELIZ and AVERGU is that people are concerned with the ability of
their children to be successful and independent adults. When hearing the question
TRABAJE, many respondents replied that it was exploitation for parents to make their
children work for them. They would often add that it was better if children worked for
their own families. They were implying that their children’s success in life was more
important than the benefits they would receive from their labor.
This concern for successful children also came through in the reactions to FELIZ
and AVERGU. Although many respondents expressed that the everyday negative aspects
of childbearing, such as putting up with noise, disorder, disagreeable children or
sacrificing their liberty or finances, did not bother them at all, worrying about their
children’s futures did bother them. Many respondents interpreted the question AVERGU
as a question about the undesirability of having a delinquent child. Delinquency was a
frequent topic of the conversations I had with people about children. Many people
worried about having children who turned out to be delinquents. They often complained
about what they perceived as a rise in delinquency in Catacamas. I often heard people
place the blame for the rise of delinquency on parents who had more children than they
could support. They suggested that children turn into delinquents when they do not
receive enough attention from their parents. FELIZ elicited the same type of responses as
AVERGU.
182
Although these reactions to the questions DIOS, JEFE, TRABAJE, FELIZ, and
AVERGU were common, it was clear from the semi-structured and unstructured
interviews that there were also respondents who differed from the majority. Some
respondents expressed their agreement that a person should have as many children as
God sends them because the Bible told them to do so. Others stated that they did want
large families. Also, some respondents reacted to the TRABAJE question by agreeing
that children should work for their parents when they are old enough because their
parents worked for them up until that point. Some respondents reacted to the AVERGU
and FELIZ questions by stating that these things did not bother them because children
were going to develop however they developed and it was not something a parent could
change. The parent should accept their children how they are. However, it seemed
obvious from the semi-structured and unstructured interviews that these points of view
were in the minority. The analysis of the factor 1 variable loadings confirmed this
assumption. The dominant cultural pattern of the respondents in phase one was that of
modified pronatalism.
Figure 7-5 also shows descriptions of the intra-cultural diversity represented by
factor 2. The figure shows two primary clusters, one with positive factor 2 scores and
one with negative factor 2 scores. The two symbols used in this graph show a clear
separation between the answer sets of men and women. There is only slight overlap
between the factor scores of men and women. Men tend to load negatively on factor 2
while women tend to load positively.
Table 7-4 shows the factor loadings of the PCA for the entire non-sex biased phaseone CBQ sorted by factor 2. I first produced these PCA loadings while I was in the field.
183
I examined the factor loadings after the phase one survey administration but before the
first round of unstructured interviews. At first, I could not explain why there were such
stark differences between men and women on the contrasting factor 2 variables. There
were certain variables that loaded either extremely high or extremely low (higher than 2
or lower than –2) on factor 2. Men tended to score highly on such variables as
INCAPAZ (“How undesirable is feeling guilty or inadequate as a parent?”) and DINERO
(“How undesirable is having a child who makes it necessary for me (my wife) to have a
job?”). Women tended to score very low on these variables. Women tended to score
high on the variables SEGURID (“How undesirable is worrying about the health and
safety of my child?”), CRIANDO (“How undesirable is worrying whether I am raising
my child the right way?”), and IMPEDIDO (“How undesirable is having a baby who is
born deformed?”). Men tended to score very low on these variables.
The only variable for which I immediately understood the split between men and
women was DINERO. This question is actually different for men and women. Men are
asked to rate the undesirability of their wives working, while women are asked to rate the
undesirability of having to work themselves. My ethnographic interviews made it clear
that many men considered it a male responsibility to be the financial provider for their
families. Therefore, for men, if their wives worked out of necessity, this would be very
undesirable because it would indicate that they were not fulfilling their responsibilities as
men.
It was more difficult for me to understand why the other variables would be central
to a split between men and women. INCAPAZ, SEGURID, CRIANDO, and IMPEDIDO
all seemed to invoke issues that would be troubling to any parent regardless of sex or
184
gender system. I could not understand why so few women indicated that it was very
undesirable to feel inadequate as a parent. Also, I could not understand why so few men
felt that it was very undesirable to worry about the health and security of their children or
worry if they were raising their children the right way. In order to understand why there
was such a stark split between the sexes on these questions, I asked my informants to
explain the split during the follow-up unstructured interviews.
I found that men and women answered these questions differently because of
different cultural expectations and responsibilities for them as parents. Men primarily
thought of their responsibilities as financial providers. Women, on the other hand, were
expected to be the primary caretakers of the children on a daily basis. Therefore, for the
men, “feeling guilty as a parent” was mainly tied to an inability to fulfill economic
responsibilities. Given the poor economic climate of Honduras at the time of my
research, many parents felt an inability to provide for their children economically. This
inability apparently translates to inadequacies and guilt for men. Women, on the other
hand, think of their motherly responsibilities mothers as including all of the childcare
activities they fulfill on a daily basis. Therefore, they were much less likely to feel
inadequate and guilty.
Women were more likely to indicate that they were bothered by worries associated
with a child’s health and security, raising their child the right way and having a deformed
or retarded child. I was told that women would often take the sole blame of having a
deformed or retarded child because the child developed in their bodies. Also, because
they bore the primary responsibility of taking care of the child on a daily basis, a child
with developmental or health problems would impact their lives more than men. This is
185
the same for worrying about the health and security. Because mothers are usually in
charge of a child’s daily life, they take more of the responsibility if there are difficulties
with the child’s health or security. Although women are very unlikely to feel bothered by
feelings of inadequacy or guilt as a parent for a child’s day-to-day activities, they
generally rated as very undesirable the feeling that they were not raising their child the
right way. I interpret this seeming contradiction as indicating that women worried about
the long term effects of their parenting (i.e. raising their children the right way) but
generally felt that they were fulfilling their daily parenting responsibilities.
Description of Culture via Analysis of Factors and Ethnographic Interviewing–
Phase Two
Figure 7-6 shows the same scatter plot of PCA factor loadings as figure 7-3, but for
the phase two respondents. The variables used in this analysis include all of the CBQ
questions administered in phase two except the same sex-biased questions I discussed
earlier. One difference between figure 7-6 and figure 7-3 is that figure 7-6 has four
symbols, rather than two. A circle represents urban men, an “X” represents urban
women, a triangle represents rural women, and a cross represents rural men.
Just as in phase one, there is visual evidence of one culture of childbearing
motivation among the phase two informants. The majority of the respondents load high
on factor 1 and cluster around a score of 0 for factor 2. Figure 7-7 confirms that factor 1
accounts for a large proportion of the overall variance in the respondents’ answers.
Factor 1 represents 58% of the variance and explains 11.6 times as much variance as the
next biggest factor. This is evidence that there is one overall pattern of natality culture.
There is also evidence of intra-cultural diversity in phase two, represented by the spread
of points scoring low on factor 1 and high on factor 2. The most interesting thing about
186
this spread of points is that nearly all of the divergent points are represented by triangles,
indicating that rural women were more likely to vary than men or urban women.
Factor 2
1
0
LOCSEX
Urban Men
Urban Women
Rural Women
-1
-1
Rural Men
0
1
Factor 1
Figure 7-6 Scatter Plot of First Two Factors of PCA for all non-sex biased CBQ Items–
Phase Two
The Threat of the Introduction of Interviewer Error
At first, when I saw that the rural women were answering differently from the rest
of the respondents, I wondered whether this was due to error introduced by hiring a new
interviewer. As I mentioned in the previous chapter, I was unable to have the same
female interviewer from phase one complete the rural interviews for phase two.
Unfortunately, it is difficult to quantitatively assess if there is error being introduced by
the interviewer or if the differences in answers is due to the characteristics of the set of
respondents. A t-test comparison between the new interviewer’s data and the data from
187
other interviewers would be pointless because rural women might systematically answer
questions differently because they are rural women and not because the interviewers are
administering the questionnaires differently.
14 0
12 0
58%
E ig en v alu e
10 0
80
60
40
20
5%
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
F ac to r
Figure 7-7 Scree Plot PCA Eigenvalues for All non-sex biased CBQ Questions–Phase
Two
Other than the differences in the CBQ data, I have no reason to suspect that this
interviewer administered the questionnaires in a less than competent manner. I trained
and worked with her on a previous, much more strenuous, project with a much more
difficult questionnaire. From my knowledge of her work, I hired her again because I felt
that she had the ability to be a successful interviewer on my project. Also, our extensive
validation of the questionnaire in previous phases of the project produced an instrument
that was relatively free of ambiguity. Thus, anyone who had a command of the language
188
and could read clearly and record answers should have been able to administer the
questionnaire successfully.
I concluded that there is a more likely explanation for rural women diverging from
the cultural pattern in phase one. Although we interviewed rural women of various ages
and economic strata in phase one, the random sampling in phase two forced us to
interview people who might not have been interviewed in phase one. In phase two, we
randomly selected villages outside of Catacamas that were very difficult and time
consuming to reach. For example, in one region, we drove for three hours along the side
of a mountain in a four-wheel drive vehicle on roads that would not support any other
type of motorized transportation. Once we reached the end of the road, we hiked further
up the side of the mountain for forty minutes to reach one of the villages that was
randomly selected. During phase one, we sought out rural people to interview, but we
did not make this type of effort to locate respondents. This leads me to conclude that a
more likely explanation for the divergent answers of rural women from the overall
cultural pattern is that they represent an intra-cultural sub-group. I am assuming that the
sample of phase one rural women was over-represented by women who lived closer to
the urban center of Catacamas.
Phase Two Factor Analysis
Table 7-5 displays the PCA factor loadings for the non-sex biased CBQ variables
for phase two sorted by factor 1. Table 7-6 displays the same loadings sorted by factor 2.
Similarly to phase one, the phase two PCA demonstrates empirical evidence of one
culture of pronatality. With only three exceptions (AVERGU, JEFE, and DIOS) the
PCM variables load positively on factor 1 and the NCM variables load negatively. The
continued existence of AVERGU, JEFE, and DIOS as exceptions to absolute pronatality
189
leads me to characterize the culture of childbearing motivation as qualified pronatality.
The phase two respondents, just as they were in phase one, are resistant to rating as
highly desirable questions that imply a desire for a large number of children. Also, they
resisted rating most negative aspects of childbearing as highly undesirable except the
aspects having to do with worrying about a child’s development and success in life.
Factor 1 for phase two is different than factor 1 for phase one, but not significantly.
Although TRABAJE loads positively on factor 1, it is the lowest positively loading PCM
variable. FELIZ and SEGURID load negatively on factor 1, but they are the NCM
variables that are closest to zero.
Table 7-5 Factor Loadings for PCA of All Non-Sex Biased CBQ Items Sorted by Factor
1–Phase Two
VARIABLE
AMOR
GUIAR
DEPEND
ADMIA
UTIL
VARON
AYUDAV
COMPART
SOCIED
FERTIL
TIEMPO
SATIS
VIDA
CHINEAR
MUJER
COMPLET
CUSTUMB
JUGAR
NIETO
FUNCION
FORTAL
DEBERES
TRABAJE
FACTOR 1
1.060
1.057
1.045
1.037
1.023
1.014
1.005
0.994
0.988
0.986
0.955
0.947
0.942
0.942
0.865
0.864
0.856
0.832
0.830
0.828
0.804
0.621
0.309
FACTOR 2
0.266
0.373
0.181
0.268
-0.135
0.142
0.357
0.251
-0.057
-0.613
-0.891
-0.667
0.001
-0.451
0.399
-0.584
-0.803
-0.508
-0.356
-2.131
-0.234
0.115
0.481
VARIABLE
AVERGU
FELIZ
SEGURID
JEFE
IMPEDIDO
VEJEZ
RUIDO
DINERO
DESAGRA
SALUD
ECONOMIC
CRIANDO
DIOS
OTRONIN
CARRERA
LIBERTAD
ENFERMO
EXIGENTE
INCAPAZ
DECUIDAR
ENERGIA
CARGA
FACTOR 1
0.250
-0.132
-0.319
-0.453
-0.517
-0.631
-0.690
-0.725
-0.812
-1.056
-1.128
-1.136
-1.233
-1.239
-1.300
-1.304
-1.335
-1.355
-1.364
-1.368
-1.470
-1.488
FACTOR 2
1.915
2.198
1.326
0.926
1.927
1.703
-0.888
-0.961
1.039
1.018
0.006
-0.295
-2.300
-0.380
-0.111
-0.540
1.081
1.184
-1.707
-0.844
-0.502
-1.198
190
The loadings on factor 2 are different for phase two than they are for phase one.
Phase two does not show the same definitive contrast between men and women which
was obvious in phase one. The phase two factor 2 does, however, show a sharp contrast
between rural women and everyone else. This change in factor loadings may seem to
indicate a lack of reliability in the characterizations of cultural patterns using this method.
However, it is important to note that, although the secondary factors for phase one and
phase two are different, these factors only represent six and five percent of the overall
variance, respectively. Subsequent factors represent less of the overall variance, but only
slightly. It is reasonable to expect that there will be some shift in the overall variance
explained by factors with repeated measurements. The most important finding using this
method–that there is an overall cultural pattern of pronatality with some intra-cultural
diversity–did not change from one sampling method to another.
The variables that load high on factor 2 (over 1.0) are all NCM variables. The
highest loading variables (FELIZ, IMPEDIDO, AVERGU) are variables that imply some
sort of concern for the long-term development of children. People who rated these
questions as highly undesirable were concerned about developing happy and welladjusted children, children who were not physically or mentally handicapped, and who
did not develop into delinquents. The variable VEJEZ (“How undesirable is being a
burden to your children in your old age?”) also loaded high on factor 2. In my
unstructured and semi-structured interviews, the respondents who rated it as very
undesirable said that they were afraid of being burdens on their children, but the only
thing they could do to prevent this was to raise their children correctly. If they did this,
their children would not think of them as burdens. I concluded from these discussions
191
that some people thought that it indicated a development issue. If children thought of
their parents as burdens and abandoned them in their old age, they were probably not
raised the right way. For those who did not rate this question as undesirable, they said
that it was the duty of children to take care of their parents in their old age. They were
not bothered with the idea of their children having to take care of them in their old age.
They accepted it as a part of life.
Table 7-6 Factor Loadings for PCA of All Non-Sex Biased CBQ Items Sorted by Factor
2–Phase Two
VARIABLE
FELIZ
IMPEDIDO
AVERGU
VEJEZ
SEGURID
EXIGENTE
ENFERMO
DESAGRA
SALUD
JEFE
TRABAJE
MUJER
GUIAR
AYUDAV
ADMIA
AMOR
COMPART
DEPEND
VARON
DEBERES
ECONOMIC
VIDA
SOCIED
FACTOR 1
-0.132
-0.517
0.250
-0.631
-0.319
-1.355
-1.335
-0.812
-1.056
-0.453
0.309
0.865
1.057
1.005
1.037
1.060
0.994
1.045
1.014
0.621
-1.128
0.942
0.988
FACTOR 2
2.198
1.927
1.915
1.703
1.326
1.184
1.081
1.039
1.018
0.926
0.481
0.399
0.373
0.357
0.268
0.266
0.251
0.181
0.142
0.115
0.006
0.001
-0.057
VARIABLE
CARRERA
UTIL
FORTAL
CRIANDO
NIETO
OTRONIN
CHINEAR
ENERGIA
JUGAR
LIBERTAD
COMPLET
FERTIL
SATIS
CUSTUMB
DECUIDAR
RUIDO
TIEMPO
DINERO
CARGA
INCAPAZ
FUNCION
DIOS
FACTOR 1
-1.300
1.023
0.804
-1.136
0.830
-1.239
0.942
-1.470
0.832
-1.304
0.864
0.986
0.947
0.856
-1.368
-0.690
0.955
-0.725
-1.488
-1.364
0.828
-1.233
FACTOR 2
-0.111
-0.135
-0.234
-0.295
-0.356
-0.380
-0.451
-0.502
-0.508
-0.540
-0.584
-0.613
-0.667
-0.803
-0.844
-0.888
-0.891
-0.961
-1.198
-1.707
-2.131
-2.300
Other questions that load highly on factor 2 include SEGURID, EXIGENTE
(“How undesirable is being responsible for a needy and demanding baby?”), ENFERMO
192
(“How undesirable is taking care of a sick child?”), DESAGRA (“How undesirable is
taking care of a baby ho is disagreeable and irritating?”), and SALUD (“How undesirable
is having a baby who strains my (wife’s) health?”). My interpretation of this block of
questions is that, in contrast with the questions discussed in the previous paragraph, they
represent day-to-day difficulties associated with childcare rather than worries of long
term child development. SALUD, which indicates a concern for a child’s health and
security, could certainly fit into both groups.
These variables that load high on factor 2 are contrasted with two PCM questions
DIOS and FUNCION (“How desirable is fulfilling my potential by having children?”)
and two NCM questions CARGA (“How undesirable is having a baby who is a burden to
my husband (wife)?”) and INCAPAZ. These load strongly negative on factor 2. At first,
the reason these questions load together on this factor was difficult for me to see. These
questions represent a mixture of types of childbearing experiences. However, when I
considered that rural women scored high on this factor and low on this set of questions as
a group, it became easier for me to interpret.
In general, women are expected to be the primary caretakers of children and have
the most responsibility for their development in Catacamas. However, from my
observations and from what I learned from my informants, this is especially true in the
rural areas. Many women in the urban area of Catacamas worked outside of their homes.
This was almost non-existent in the rural areas. Urban husbands are more likely to
participate in childcare activities than their rural counterparts. Urban women are still
considered the primary caretakers of their children, but the notion that childcare is a
193
shared activity between husbands and wives is much stronger in the city than the villages.
In the villages, this idea is almost non-existent.
With this in mind, I can interpret this block of questions as representing a contrast
to the childbearing concerns that rural women have, which are represented by the block
of questions loading high on factor 2. Rural women are concerned with two categories of
childcare: worrying about their future development and putting up with the day-to-day
struggles of raising children. On the other hand, they are not concerned about something
such as the burden a child might place on their spouse. Because rural women did most of
the childcare themselves without help from their husbands, they would understandably
not place too much importance on burdening their husbands with a child. Also, rural
women do not feel inadequate as mothers because they are dedicating their lives to
raising their children.
The inclusion of DIOS in this block of questions is interesting because of the
results of the national data I discussed in Chapter 5. The national data showed that rural
women were more likely than rural men to say that their ideal number of children was
however many God sent them. However, when I re-interviewed women who indicated
desirable to the question DIOS in phase one, I was often told that they answered this way
because of a sense of obligation. They considered using contraceptives a sin. However,
several of these women had tubal ligation surgery even though they favored having as
many children as God sent them. They told me that they would like to have as many
children as God sent them, but they could no longer continue having children because of
economic pressures. These conversations about Dios me mande gave me the impression
that rural women were feeling a sense of cognitive dissonance about the number of
194
children they had. On the one hand, they wanted to fulfill the expectation of having the
number of children that God sent. On the other hand, they felt the need to pursue the
female sterilization operation because God had already sent them too many children.
Therefore, I consider DIOS to be consistent with the block of questions that represent
aspects of childbearing that rural women would find undesirable. Having all the children
that God sends is a major burden on these women. They often fulfill this expectation out
of obligation but suffer the majority of the consequences.
FUNCION is another interesting question that made sense to me as a part of this
block only after I considered how it related to the lives of rural women. As I discussed in
Chapter 6, the wording of this question underwent the most changes from the first
translation until the actual administration of the questionnaire. My assistants and I finally
settled on a wording that stressed that the word “potential” was referring to one’s
potential as a person, not the physical potential to have children. However, I am not
convinced that many people in Catacamas could understand having children as relating to
fulfilling one’s potential as a person. For many people, especially rural people, having
children is such an integral aspect of life that the thought of not having children does not
enter their minds. This is why it was very difficult to ask people to name the positive
aspects of children. Having children is so much a part of life for many people that
contemplation of positive motivating aspects of childbearing was uncommon. I think that
many people considered having children such an integral aspect of life that, even when
we stressed that we were asking about fulfillment of their potential as a person, they
continued to interpret the question as fulfillment of their physical potential to have
195
children. Therefore, I am not surprised that this question is linked closely with DIOS as
an indication of a preference for a large number of children.
Factor Analysis of Sub-Scales: PCM
Because the overall cultural pattern of childbearing motivation is strong pronatality,
which, for the most part, contrasts the PCM questions with the NCM questions, I decided
to analyze the individual sub-scales independently in order to better understand the
cultural diversity. Figure 7-8 is a scatter plot of the first two PCA factor loadings on all
non-sex biased PCM questions. This scatter plot also shows a description of the factor
patterns, which I will discuss below. Figure 7-9 shows the scree plot of the eigenvalues
produced by the PCA of the PCM. It is clear from these figures that the cultural
agreement on positive childbearing motivation in isolation is slightly less than the overall
agreement. The points are more distributed in the scatter plot. The overall amount of
variance explained by factor 1 is less and the ratio of explained variance between factor 1
and factor 2 is lower. However, factor 1 is still 8 times as large as factor 2. The strength
of factor 1 still indicates a large amount of cultural agreement on the positive aspects of
childbearing.
Table 7-7 shows the first two PCA factor loadings sorted by factor 1 and table 7-8
shows these same loadings sorted by factor 2. Factor 1 has no variables that load
significantly high and positively (greater than .5) and only four variables that load
negatively (DIOS, JEFE, TRABAJE, and DEBERES). My interpretation of this factor is
that it represents a contrast between a preference for a large number of children and all
other positive childbearing motivations. I have discussed how DIOS, JEFE, and
TRABAJE relate to a preference for a high number of children earlier in this chapter.
DEBERES (“How desirable is fulfilling your religious feelings about family life?”) also
196
relates to number of children because the primary influence of religion on childbearing in
Catacamas was the rejection of contraception by the Catholic Church and other religious
groups. It is interesting to note that, although the overall pattern of childbearing
motivation is decidedly pronatal, those who favor the positive childbearing aspects
having to do with large family size are the cultural outliers. Factor 1, which explains
nearly half of the overall variance in the PCM questions, is primarily driven by a
rejection of these aspects.
1
Factor 2
Children
as Wealth
0
LOCSEX
Urban Men
Importance
of Family Size
-1
-1
Urban Women
Rural Women
Rural Men
0
1
Factor 1
Figure 7-8 Scatter Plot of First Two Factors of PCA for all non-sex biased PCM Items–
Phase Two
In figure 7-8, I named the factor 2 dimension of the scatter plot “Children as
Wealth” because the variables loading high and low seem to be contrasting children’s
non-economic value with their value as household providers. Table 7-8 shows that
197
TRABAJE is the variable that is loading the strongest on the positive side of factor 2.
TRABAJE is contrasted against SATIS (“How desirable is giving my husband (wife) the
satisfaction of being a father (mother)?”), JUGAR (“How desirable is playing with your
child?”), FUNCION, and DIOS. All four of these questions indicate positive aspects of
childbearing that are not financially linked. In fact, DIOS was often interpreted as
guaranteeing financial hardship.
120
100
Eigenvalue
48%
80
60
40
8%
20
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Factor
Figure 7-9 Scree Plot PCA Eigenvalues for non-sex biased PCM Questions–Phase Two
Factor Analysis of Sub-Scales: NCM
Figure 7-10 is a scatter plot of the PCA factor loadings for the non-sex biased NCM
questions. Figure 7-11 is the scree plot of the eigenvalues for the PCA of the NCM. The
PCA analysis of the NCM sub scale shows that, for negative aspects of childbearing,
198
there is much less cultural agreement. The points on the factor loading scatter plot are
much more spread than for the previous analyses. Also, the scree plot indicates that the
first factor is not significantly larger than the second factor. There is not a large drop
from the first to the second factor because it only explains 2.3 times the amount of
variance.
Table 7-7 Factor Loadings for PCA of Positive Non-Sex Biased CBQ Items Sorted by
Factor 1–Phase Two
VARIABLE
ADMIA
GUIAR
DEPEND
AMOR
UTIL
FERTIL
AYUDAV
COMPART
VARON
SOCIED
SATIS
CHINEAR
VIDA
FACTOR 1
0.515
0.507
0.504
0.497
0.456
0.444
0.434
0.432
0.411
0.410
0.350
0.333
0.323
FACTOR 2 VARIABLE
0.160
TIEMPO
0.204
CUSTUMB
0.061
MUJER
0.151
JUGAR
-0.102
COMPLET
-0.332
FUNCION
0.498
NIETO
0.359
FORTAL
-0.044
DEBERES
-0.015
TRABAJE
-0.627
JEFE
-0.103
DIOS
-0.215
FACTOR 1
0.308
0.242
0.220
0.202
0.194
0.173
0.166
0.117
-0.233
-0.752
-2.419
-3.834
FACTOR 2
-0.508
-0.379
-0.101
-0.801
-0.389
-1.044
-0.197
-0.144
0.127
4.371
0.259
-1.189
In figure 7-10, I’ve named the factor 1 axis “Quality of Children” because the
variables that load high are mainly related to long-term child development. These
variables are contrasted with variables that represent temporary parental inconveniences
and sacrifices. Table 7-9 shows the fist two PCA factor loadings sorted by factor 1 and
table 7-10 shows these same loadings sorted by factor 2. AVERGU and FELIZ are the
variables that load high on factor 1 and CARGA, ENERGIA, INCAPAZ, DECUIDAR
(“How undesirable is straining our marriage with a baby?”), and CARRERA (“How
undesirable is being kept from my (having my wife being kept from her) career or job by
a baby?”) are some of the variables that load negatively. AVERGU and FELIZ load
199
significantly higher than any other variable. Their loadings are 2.901 and 1.847,
respectively, while no other variable has a loading of .7 or higher. The variables that load
negatively, on the other hand, are only gradually more negative. None of them loads
lower than –1.0.
Table 7-8 Factor Loadings for PCA of Positive Non-Sex Biased CBQ Items Sorted by
Factor 2–Phase Two
VARIABLE
TRABAJE
AYUDAV
COMPART
JEFE
GUIAR
ADMIA
AMOR
DEBERES
DEPEND
SOCIED
VARON
MUJER
UTIL
FACTOR 1
-0.752
0.434
0.432
-2.419
0.507
0.515
0.497
-0.233
0.504
0.410
0.411
0.220
0.456
FACTOR 2 VARIABLE
4.371
CHINEAR
0.498
FORTAL
0.359
NIETO
0.259
VIDA
0.204
FERTIL
0.160
CUSTUMB
0.151
COMPLET
0.127
TIEMPO
0.061
SATIS
-0.015
JUGAR
-0.044
FUNCION
0.101
DIOS
-0.102
FACTOR 1
0.333
0.117
0.166
0.323
0.444
0.242
0.194
0.308
0.350
0.202
0.173
3.834
FACTOR 2
-0.103
-0.144
-0.197
-0.215
-0.332
-0.379
-0.389
-0.508
-0.627
-0.801
-1.044
-1.189
Visual inspection of the vertical axis (factor 2) reveals that the split between sexes
that appeared in the phase one analysis shows up again. The split between men and
women is not as stark as it was in figure 7-3 but it is clear that the majority of points on
the lower half of the plot are representing either rural men or urban men. The upper half
contains mostly points representing rural and urban women. Factor 2 is primarily driven
by the variable DINERO, which is weakly contrasted against the variables SEGURID
and IMPEDIDO. As I discussed in the analysis of phase one, DINERO was typically
rated as highly undesirable by men because they often felt it was an embarrassment for
them to have their wives work. They felt that this would represent their inability to
provide for their families sufficiently. Also, women, because of their role as the primary
200
caretakers of children, typically rated SEGURID and IMPEDIDO as highly undesirable.
AVERGU also loaded highly negative on factor 2. It is perhaps linked with DINERO
because they both (for men) represent embarrassing situations whereby parental
responsibilities are not being met.
1
Factor 2
Parental
Expectations
Quality of
Children
0
LOCSEX
Urban Men
Urban Women
Rural Women
-1
-1
Rural Men
0
1
Factor 1
Figure 7-10 Scatter Plot of First Two Factors of PCA for all non-sex biased NCM Items–
Phase Two
Examination of Intra-Cultural Variation Using OLS Regression Analysis
The preceding analyses of intra-cultural variation has relied mainly on visual
inspection of the distribution of two variables, sex and urban or rural residence. Because
there are several other variables that have a theoretical influence on cultural agreement
for the domain of childbearing motivation, I further analyzed the data with ordinary least
squares regression analysis (OLS). The dependent variables in the multiple regression
201
analyses are the factor loadings for the first factor in the overall CBQ analysis, the PCM
analysis and the NCM analysis. These factor scores represent the amount that individual
respondents conformed to the cultural agreement on the set of variables in the PCA
analysis. The regression analysis will determine which factors predict cultural
conformity.
60
28%
50
E ig e nv a lue
40
30
12%
20
10
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
F a c to r
Figure 7-11 Scree Plot PCA Eigenvalues for non-sex biased NCM Questions–Phase Two
The independent variables that are included in the regression analyses described in
this section include sex, area of residence, age, number of children, education, and
religiousness. A dummy variable FEMALE represents the independent variable sex. It’s
value is 1 if the respondent is a woman and 0 if the respondent is a man. Another dummy
variable called URBAN represents location of residence. This variable is equal to 1 if the
202
respondent lives in the city of Catacamas and 0 if the respondent lives in one of the
surrounding villages. AGE, EDUCATION, and CHILDREN are all straightforward
variables that represent the number of years living, the number of years of school, and the
number of children ever born (currently living or not), respectively. RELIG2 is a
variable that represents the answer to a question about how religious respondents
consider themselves. The answers range from very religious (equal to 1) or not religious
at all (equal to 4). I chose to present the analyses of these variables because they are
either significant predictors of the cultural measurements (the factor 1 loadings) or they
are important as control variables. There were other variables that I included in other
regression analyses that I am not presenting here. I did not find any evidence that they
were important as either predictors or control variables. I will discuss these variables
later in this section.
Table 7-9 Factor Loadings for PCA of Negative Non-Sex Biased CBQ Items Sorted by
Factor 1–Phase Two
VARIABLE
AVERGU
FELIZ
VEJEZ
SEGURID
IMPEDIDO
DINERO
RUIDO
DESAGRA
SALUD
ECONOMIC
FACTOR 1 FACTOR 2 VARIABLE
2.901
-1.070 CRIANDO
1.847
-0.198 OTRONIN
0.681
-0.519 EXIGENTE
0.638
2.893 ENFERMO
0.529
1.916 LIBERTAD
0.509
-1.933 CARRERA
0.407
0.318 DECUIDAR
0.194
0.655 INCAPAZ
-0.185
-0.082 ENERGIA
-0.289
-0.217 CARGA
FACTOR 1 FACTOR 2
-0.500
0.637
-0.534
-0.067
-0.627
-0.238
-0.645
-0.139
-0.707
-0.145
-0.735
-0.271
-0.800
-0.297
-0.802
-0.385
-0.924
-0.437
-0.960
-0.421
Table 7-11 shows the results of an OLS regression analysis using the statistics
package SYSTAT. The dependent variable (F1MFAL) is the factor loading for the first
factor from the PCA analysis. This is the factor loading for the analysis of the entire
CBQ questionnaire, except for the sex-biased questions. This is also the factor that I
203
have previously described as the measurement of the culture of (qualified) pronatality.
Higher scores on this factor represent conformity with the overall natality culture, which
considers most aspects of childbearing positive.
Table 7-10 Factor Loadings for PCA of Negative Non-Sex Biased CBQ Items Sorted by
Factor 2–Phase Two
VARIABLE
SEGURID
IMPEDIDO
DESAGRA
CRIANDO
RUIDO
OTRONIN
SALUD
ENFERMO
LIBERTAD
FELIZ
FACTOR 1
0.638
0.529
0.194
-0.500
0.407
-0.534
-0.185
-0.645
-0.707
1.847
FACTOR 2
2.893
1.916
0.655
0.637
0.318
-0.067
-0.082
-0.139
-0.145
-0.198
VARIABLE
ECONOMIC
EXIGENTE
CARRERA
DECUIDAR
INCAPAZ
CARGA
ENERGIA
VEJEZ
AVERGU
DINERO
FACTOR 1
-0.289
-0.627
-0.735
-0.800
-0.802
-0.960
-0.924
0.681
2.901
0.509
FACTOR 2
-0.217
-0.238
-0.271
-0.297
-0.385
-0.421
-0.437
-0.519
-1.070
-1.933
The adjusted R-Squared for the model is .514. This represents the proportion of
intra-cultural variation that the model explains, controlling for the number of independent
variables (Handwerker 2001, p. 226-227). The most important and significant (p < .001)
independent variables are FEMALE and URBAN. CHILDREN is also another
significant variable (p < .05) but does not have as large a coefficient as FEMALE or
URBAN. The model also controls for AGE, RELIG2, and EDUCATION. These are not
significant predictors in this model.
Table 7-12 shows the results of another OLS regression model. The dependent
variable (F1MFP) is the factor loadings for the first factor of a PCA analysis on the PCM
variables (non-sex biased) alone. Previously, I described this variable as the
measurement for the “Importance of Family Size” culture. Higher scores on this factor
represent conformity with the culture of family size manageability. Low scores represent
cultural diversity that favors larger families.
204
Table 7-11 Sources of Intra-cultural Variation for Pronatality
Dep Var: F1MFAL N: 199 Multiple R: 0.727 Squared multiple R: 0.528
Adjusted squared multiple R: 0.514 Standard error of estimate: 0.157
Std
Effect
Coefficient Error
CONSTANT 0.872
0.062
FEMALE
-0.256
0.023
URBAN
0.190
0.025
AGE
-0.001
0.002
CHILDREN
-0.011
0.005
RELIG2
-0.003
0.013
EDUCATION -0.006
0.003
Analysis of Variance
Sum-ofSource
Squares
Regression
5.326
Residual
4.755
df
6.000
192.000
Std
Coef
0.000
-0.569
0.422
-0.041
-0.154
-0.013
-0.105
Tolerance
.
0.956
0.787
0.522
0.470
0.937
0.720
MeanSquare
0.888
0.025
t
14.092
-11.225
7.557
-0.592
-2.130
-0.248
-1.794
F-ratio
35.841
P(2 Tail)
0.000
0.000
0.000
0.555
0.034
0.805
0.074
P
0.000
Table 7-12 Sources of Intracultural Variation for Manageable Family Size Culture
Dep Var: F1MFP N: 198 Multiple R: 0.718 Squared multiple R: 0.515
Adjusted squared multiple R: 0.500 Standard error of estimate: 0.211
Effect
CONSTANT
FEMALE
URBAN
AGE
CHILDREN
RELIG2
EDUCATION
Coefficient
0.778
-0.195
0.165
-0.004
-0.035
0.044
0.001
Analysis of Variance
Sum-ofSource
Squares
Regression
9.049
Residual
8.509
Std
Error
0.083
0.031
0.034
0.002
0.007
0.017
0.004
df
6.000
191.000
Std
Coef
0.000
-0.328
0.278
-0.137
-0.370
0.132
0.012
Tolerance
.
0.956
0.782
0.521
0.470
0.937
0.716
MeanSquare
1.508
0.045
t
9.377
-6.368
4.877
-1.957
-5.033
2.545
0.195
F-ratio
33.852
P(2 Tail)
0.000
0.000
0.000
0.052
0.000
0.012
0.846
P
0.000
The adjusted R-Squared for the model is .500, which indicates that the model
explains 50 percent of the intra-cultural diversity, controlling for the number of
205
independent variables. Once again, the variables FEMALE and URBAN are the most
significant (p < .001) and strongest predictive variables in the model. CHILDREN and
RELIG2 are also highly significant (p < .001), but with less predictive power. Age is
significant at the .10 level but is very weakly predictive. The only variable that is not
significant is EDUCATION, which is in the model for control purposes only.
Table 7-13 shows the results of an OLS regression model on the first factor’s
loadings from the PCA analysis of the non-sex biased NCM questions. I describe this
factor as “Fears about quality of Children.” Higher scores on this factor relate to
conformity with the cultural pattern of considering long-term developmental childcare
concerns more important than short-term childcare burdens.
Table 7-13 Sources of Intra-cultural Variation for Fears About Quality of Children
Culture
Dep Var: F1MFN N: 199 Multiple R: 0.638 Squared multiple R: 0.407
Adjusted squared multiple R: 0.389 Standard error of estimate: 0.210
Effect
CONSTANT
FEMALE
URBAN
AGE
CHILDREN
RELIG2
EDUCATION
Coefficient
0.625
-0.308
-0.155
0.002
0.006
-0.012
0.003
Analysis of Variance
Sum-ofSource
Squares
Regression
5.800
Residual
8.446
Std Error
0.082
0.03
0.034
0.002
0.007
0.017
0.004
df
6.000
192.000
Std Coef
0
-0.575
-0.29
0.066
0.065
-0.04
0.048
Tolerance
.
0.956
0.787
0.522
0.47
0.937
0.72
MeanSquare
0.967
0.044
t
7.584
-10.122
-4.628
0.862
0.8
-0.695
0.731
F-ratio
21.973
P(2 Tail)
0
0
0
0.39
0.425
0.488
0.465
P
0.000
The adjusted R-Squared for the model is .389, which indicates that the model only
accounts for around 39 percent of the cultural diversity, controlling for the number of
206
independent variables. The variables FEMALE and URBAN are once again the most
highly significant and highly predictive of the model’s independent variables. None of
the other controlling variables are significant in the model.
I also ran modified versions of these regression models with two other variables.
These two variables represented variables that I considered to be theoretically important
in the determination of cultural diversity. The first variable was a measurement of the
family’s income, based on three levels: low, medium, high. These three levels match the
levels that I used to separate the sample in phase one. I did not find this variable to be
strongly predictive or significant in any of the above models. The inclusion of this
variable did not change the results that I discussed above. I decided to not include it as a
control variable because it is correlated with the urban/rural variable.
The other variable that I measured and included in preliminary regression models
was a variable that measured perception of employment opportunities for women. As I
discussed in Chapter 6, each respondent was asked about a series of job types. They were
asked if the job type was something that should be done by women, men or could be done
by both. I dichotomized each of these job variables into another variable that had a value
of one if the respondent said that the job could be done by a woman or could be done by
both. The variable had a value of zero if they said that only men could do the job.
I ran a Guttman scale analysis on the set of dichotomized variables using the
statistics package Anthropac (Borgatti 1992). The coefficient of Reproducibility was
.879. This gives evidence that the list of employment types represented a set of items that
measure an underlying unidimensional variable (Bernard 2002, p. 302-304). I call this
variable, “Perception of female employment opportunity.” However, when I aggregated
207
the individual job type variables into an overall female employment opportunity
measurement and included it into regression models as an independent variable, I found
that the variable had no predictive ability. Therefore, I removed it from subsequent
regression models.
Discussion of Analysis
The main conclusion that I draw from the analysis presented above is that the two
factors that are the most important in determining agreement and diversity for natality
culture in Catacamas are sex and location of residence. For each cultural domain, sex
and location were highly significant predictors and were responsible for a large
percentage of the intra-cultural variance. There were other variables that were significant
predictors, but none had anywhere near the effect on the overall model that sex and
location had.
The interpretation of the effect that sex and location have on natality culture is
complicated. For both men and women, there appear to be some contradictory
tendencies. Although participating in the overall pronatal culture, they are also
participating in intra-cultural patterns that suggest conflicting tendencies.
For women, especially rural women, there appears to be some level of cognitive
dissonance regarding the issue of family size. Through analysis of the entire CBQ data it
is clear that rural women are more likely to diverge from the overall pattern of
pronatality. This does not imply that they have more antinatal tendencies than other
categories of people. As I discussed earlier, the pattern of pronatality culture is qualified
because there is agreement that family sizes need to be limited to a manageable number.
The analysis of the PCM questions alone shows that rural women were more likely to
diverge from this cultural pattern of manageable family size. I interpret this finding as
208
indicating that rural women are even more pronatal than the overall pattern of natality–at
least with regard to family size.
My ethnographic experience confirms that rural women in Catacamas have much to
gain from being the parent of a large family. This is mainly because of the lack of social
roles for rural women outside of motherhood. There is very little female employment for
women in the villages outside of Catacamas. Even domestic work, which employs a
large number of poor urban women, is not much of an option in rural areas. Urban areas
have more opportunities for poor women to be domestic workers because urban women
are more likely to work outside of the home. The households with two working parents
create a demand for domestic workers that does not exist in the rural areas. This lack of
economic opportunity makes rural women more dependent on their children for
economic support, especially because their husbands do not always share their income
with their families. Therefore, rural women experience a benefit from large families that
is not present to other categories of people in Catacamas.
On the other hand, women are more likely to diverge from the pattern of negative
childbearing motivation that considers worries about long-term development of children
to be more important than the temporary inconveniences of raising them. This is
understandable considering women’s place in Catacamas as the parent primarily
responsible for childcare. Because they are more likely to experience the inconveniences
and sacrifices associated with childcare, it makes sense that they would be more likely to
diverge from this pattern. Urban women are more likely than rural women to diverge
from this pattern, which also makes sense considering the availability of non-childcare
related opportunities in the urban area. Rural men are the most likely to conform to the
209
overall pattern, which is also understandable given that they are the least likely to be
burdened with childcare. Although urban men are less likely to partake in household or
childcare activities than urban women, they are more likely than their rural counterparts.
Therefore, although rural women are more likely to be favorable to larger families,
women in general are more likely to be bothered by the day-to-day inconveniences
associated with childcare.
There are also some interesting sources of cognitive dissonance at work for men,
especially urban men. From the analysis of the PCM data, men are more likely to
conform to the pattern of considering worries associated with the long-term development
of their children to be the primary sources of negative childbearing motivation.
However, the second factor of the NCM analysis also makes it clear that men consider
their wives having to work to be a negative concern as well. As I discussed earlier, many
men considered that their duties as fathers and husbands included being the sole provider
for the household. As a result of my participant observation and ethnographic
interviewing, I came to realize that these two concerns are sometimes in conflict.
My male assistant, who was also one of my key informants, confirmed this
conclusion. He told me about his personal conflict with these two issues. When he was
young and newly married, he absolutely refused to allow his wife to work, primarily
because of the ridicule he thought he would receive from other men. He said that he
resisted this because he was a machista when he was younger. His machista behavior
included heavy drinking, being abusive to his wife, and having sexual relations with other
women. However, he eventually changed his behavior and supported his wife to
continue with school. She now works nights as an accountant and he willingly takes care
210
of their youngest child while she is at work. He also proudly claims that he has stopped
behaving like a machista in other ways. He credits his change of heart to his desire for
his children to have better lives than he had. His main motivating force is that he wants
them to get University degrees. He realized that his behavior, including his treatment of
his wife, was having a negative effect on his children’s development. He also strongly
credits his desire for his wife’s income to help pay for his children’s education expenses
as a turning point in his attitude towards her and his family. He strongly believes that this
is a common thing for other men in Catacamas as well.
This also is consistent with the finding that rural men are more likely than urban
men to conform to the pattern of placing long term developmental concerns above the
temporary inconveniences associated with childcare. Urban men participate in childcare
to a much higher degree than rural men and are more likely to diverge from the pattern
that de-emphasizes day-to-day childcare aggravations.
It is possible that these contradictions, which hint at the possibility of cognitive
dissonance for both men and women, are key elements in the culture change process that
occurs during a fertility decline. As Honduras becomes more urbanized and as there are
more employment opportunities besides agricultural production, women are having
access to more roles besides motherhood and are relying less on their children for
economic support. The demands of childbearing and childrearing are in conflict with
attaining these new roles and opportunities. Men and women are seeing educational and
employment opportunities for their children that did not exist before. They are also
realizing that large families have a diminished opportunity to pursue these opportunities.
Also, men are realizing that they have to resolve the dissonance between their perception
211
of their place in the household and the goals they have for their children. Resolution
often comes from their rejection of previously held notions that men need to be the sole
providers for their families and that they have the right to avoid participating in childcare
because of their provider status.
Men like my assistant realize that, in order for their children to have successful
lives, they will have to give them this opportunity through education. Higher education
will allow their children to become more competitive for the few well paying jobs that
exist in Honduras. They realize that, in order to give their children these opportunities,
they cannot afford to resist allowing their wives to work. Because of greater educational
and employment opportunities for women, sharing the economic burden with their wives
becomes even more attractive. This changes the intra-household dynamics between men
and women. It also leads to the re-definition of male and female parental roles.
As men and women continually work to resolve these conflicts, both in their
behaviors and their minds, they participate in the discourse on parenting and childcare
with members of their social network. It is through this process that new cultural models
of motherhood and fatherhood are created. This redefinition of masculine and feminine
roles is the epicenter of the fertility decline process.
CHAPTER 8
CONCLUSION AND DIRECTIONS FOR FUTURE RESEARCH
Findings
In this work, I presented an investigation of fertility decline in Honduras. I argued
for a theoretical model of fertility decline that ranges across social science disciplines.
Also, I promoted a particular style of research, called anthropological demography, in
which fieldwork is combined with secondary data analysis. In the style of research,
typically un-operationalized concepts, such as culture, are fully operationalized and
measured. As an example, I offer here what I believe to be the first extended
operationalization and measurement of natality culture.
My findings indicate that cultural differences regarding childbearing motivation are
primarily driven by two variables: sex (male and female) and location (urban or rural).
This is true at the national level as well as at the local level in Honduras. These cultural
differences are manifest in the quality and quantity of the children that people say they
want.
My goal in this project, and one of the primary goals of anthropological
demography in general, is to provide a context for demographic changes that are ongoing.
The analysis of secondary data sources is a valuable and necessary part of the study of
population change. However, as my analysis in Chapter 5 demonstrated, this type of
analysis cannot answer all our questions. In fact, new questions were created through the
analysis of secondary data, and I have argued that anthropological demography, firmly
based on field research, is a tool for answering many such questions.
212
213
The fieldwork component of anthropological demography allows us to understand
how people are evaluating their options, on the ground, while they decide how many
children to have. It allows us to investigate how ideas–such as what it is to be a
man/father or woman/mother–have an effect on people’s fertility decisions and how those
ideas change in response to material circumstances. In other words, fieldwork-based
demography gives a window into how ideas are constructed and re-constructed by groups
of people as the world changes around them.
Although the findings of this study are not an exhaustive test of any one theory of
fertility decline, they support a combination of resource access theory and embodied
capital theory (outlined in Chapter 2). As Kaplan’s embodied capital theory would
predict, I found that people are constantly evaluating tradeoffs between quantity and
quality of children. There was a clear understanding of the tradeoffs between having a
large family and having a manageable amount of children. People decide what is a
manageable amount of children based primarily on their evaluation of how much
education they can afford. People recognize that their children are living in a changing
economy that now requires adults to have certain skills to be competitive in the
marketplace.
As Handwerker’s resource access theory would also predict, parents are concerned
with education for their children because they recognize that there are opportunities to
access resources for people with certain skills. Also, as Handwerker has argued, it is
clear that changing channels of resource access are causing a culture change process. I
found cultural diversity in people’s conception of what constituted suitable family size
and what constituted good quality children. Ideas about what constitutes a large or small
214
family and what constitutes sufficient investment in children are shared–that is, they are
cultural. However, this sharing is not perfect because the cultural definitions of what
constitutes good quality children and what constitutes a good quantity of children are
constantly being modified as people live their lives, engage in conversations, and interact
with people around them. This interaction creates the meaning that people need in order
to face the challenge of raising children in an environment with changing channels to
resource access.
Another set of cultural definitions that are shaped by resource access is the
understanding of what it is to be a man/father/husband and women/mother/wife. The
people whom I studied are negotiating between accepted definitions of masculinity and
femininity and new opportunities to access resources for themselves and their children.
These new opportunities are causing people in individual households to reorganize their
personal definitions of masculinity and femininity into more workable models. As life
history theory and resource access theory predict, changing opportunities do not affect
men and women equally. They each have their own reproductive objectives, which
sometimes conflict. I have demonstrated that sex is a major force in determining cultural
diversity of childbearing motivation. However, the expression of sex differences in
natality culture is a function of environmental forces.
Another important finding in my work is that the realms of positive and negative
childbearing motivation have different cultural trajectories. I found that there was strong
cultural agreement on positive childbearing motivation coupled with important cultural
diversity associated with negative childbearing motivation. This demonstrates that
fertility decline does not necessarily imply that children are less desirable to people with
215
lower fertility. People are not necessarily having fewer children because they do not
want them. Instead, their negative motivations, which include worries that they are
giving their children the opportunity to have good lives, are growing stronger.
I conclude that, at this point, the people of Catacamas are going through a
negotiating process with these positive and negative motivations vying for primacy. My
field experience and my data analysis suggest that the people there are resolving some
contradictions about parenthood/motherhood/fatherhood. I expect to see fertility
continue to decline in the coming years as many people look for ways to give their
children better lives. Also, I expect to find that cultural changes will proceed as people
re-define what constitutes a successful family.
Future Research
My original intent in this research was to produce a building block of knowledge
about the process of culture change and fertility decline. Both culture change and fertility
decline are processes that occur over time and space. Each occurrence of a fertility
decline will have certain differences from other instances of fertility decline. This makes
repetition of data collection and replicability of findings of utmost importance to produce
confidence in conclusions about the general process of fertility decline. This forces the
work of determining general laws about this process to go beyond the scope of any one
project, any one researcher, or even any one discipline. I have presented my theory,
methods, and findings in a way that I hope will facilitate future research on fertility
decline.
There are certain immediate projects that can build on the findings presented here.
First, a complete repetition of the data collection for testing the entire T-D-I-B
psychological model of childbearing motivation is required to determine the validity of
216
the model across cultures. The data collected in this work measured the “T” element of
the model, but determining the relationship between motivational traits and desires,
intentions, and behaviors in a variety of cultural settings would allow us to generalize or
modify the conclusions of this study. The measurements of cultural agreement that I
developed for this project can be used to empirically determine the effect of culture on
the T-D-I-B sequence.
Further measurement of T-D-I-B model will also help to illuminate the process of
culture change and fertility decline. Theories that link infrastructural and structural
changes to changes in reproductive behavior typically assume changes in cognitive
evaluations of reproduction. The exact nature of these changes has been mostly assumed.
The direct measurement of the T-D-I-B model in various contexts is a way to test
theoretical links among infrastructure/structure, reproductive behavior, and cognition.
Perhaps the most important direction for future research is the repetition of this
study. Firstly, this study should be repeated in other areas of Honduras. This will allow
for greater understanding of the connection between local findings and national level
data. Repetition in the Honduran cities of Tegucigalpa and San Pedro Sula are the most
important locations for a repetition of this study. This is because of the prominence of
these two cities in the lives of most Hondurans. Also, because of the high rate of
urbanization in Honduras, it is likely that the lives of the average Honduran of the future
will become more like the lives of those now living in these two major cities.
Another fertile area for a repetition of this study is in a more remote and rural area
than those surrounding Catacamas. Although areas of this study were relatively isolated
from Catacamas, the degree of isolation from an urban lifestyle was miniscule compared
217
with other areas of Honduras. Repetition of data collection in the major urban areas and
isolated rural areas will help to refine the conclusions of this study regarding the effect of
urbanization on natality culture.
Also, this study should be repeated in Honduras on targeted ethnic groups. In the
interest of simplicity, I designed this study to ignore the issue of ethnicity. The
respondents were all assumed to be Mestizo. However, there are likely to be aspects of
natality culture that vary according to ethnicity. Therefore, it is important to repeat this
study on a population of ethnic minorities in Hondudras. Honduras has many indigenous
populations living in various areas throughout the country. Also, the northern coast of
Honduras has a substantial population of Garifuna people and other descendents of
African slave populations. Comparisons among these various ethnic groups may or may
not result in substantial differences in natality culture. It is important that the hypothesis
that ethnicity influences natality culture in Honduras undergo a rigorous test.
Repetition of this data collection and analysis outside of Honduras is also
important to build on the findings of this study. Because fertility decline is a global
phenomenon, it is important to have comparative data from as many places around the
world as possible. In particular, it would be beneficial to have this study repeated in
areas that have populations with different fertility rates than Honduras. Repetition of this
study among western European populations, which have some of the lowest fertility rates
in recorded history, would help to better define the cultural changes that parallel fertility
decline. Also, it would be important to repeat this study with populations that have even
higher fertility rates than Honduras.
218
The greatest benefits of a repetition of this research may come from a re-study of
the same population at several points in time in the future. This is the only way to
definitively describe a process of change over time. With this type of data collection,
conclusions about culture change will have more validity. A major benefit of this
repetition is that changes in the cultural data can be compared with changes in other
variables, such as the TFR of Honduras. Also, cultural trends can be compared with
trends in other macro-level economic indicators. The most important types of economic
variables would be those that indicated trends in employment opportunities for women.
A re-study of the same population at several points in time in the future would also allow
for the tracking of a natural experiment. If Honduras undergoes a precipitous fertility
decline in the near future, this study can serve as a baseline data source for the cultural
conditions before the TFR drop.
In any case, repetition of this study will help to fill in the blanks of fertility decline
theory across the social sciences. In the end, the success of this work will be judged by
its utility to other researchers who study fertility decline. In order to develop a more
complete understanding of fertility decline, it is important that fertility research continue
in as many disciplines as possible. It is my hope that my interest in conducting
interdisciplinary research on fertility decline will facilitate the further collaboration of
social science researchers. In particular, I hope that this work provides a model for the
conduct of anthropological demography research. Anthropologists have much to offer an
interdisciplinary science of fertility and I hope that this work plays some small role in the
improvement of research in anthropological demography.
APPENDIX A
THE CHILDBEARING QUESTIONNAIRE (CBQ)
The following list was originally printed in Miller (1999, pp. 100-102). It
contains the items from Positive Childbearing Motivation (PCM) and Negative
Childbearing Motivation (NCM) questionnaires. The items consist of those making up
the five PCM subscales, the four NCM subscales, as well as un-scaled items. The list
also contains the items I added during this study. The wording of the items is for women
respondents, with the male respondents in parentheses.
The variable names used in the analysis of the CBQ data are shown in front of
each item in bold. The items marked with a * were used during phase one only.
PCM: Joys of Pregnancy, Birth and Infancy
Feeling a baby move and kick inside me (your wife’s body)
MOVER
Giving my husband (wife) the satisfaction of being a father (mother)
SATIS
DARLUZ Giving birth to a baby (Helping your wife give birth to a baby)
Breast (bottle) feeding a baby
PEPE
CHINEAR Holding and cuddling a baby
Dedicating myself and much of my time raising children and being a
TIEMPO
mother (father)
PCM: Traditional Parenthood
CUSTUMB Having a child who will carry on my family traditions
Being the center of a large, active family
JEFE
Strengthening our marriage through a child
FORTAL
DEBERES Fulfilling my religious feelings about family life
Providing my parents with a grandchild
NIETO
FUNCION Fulfilling my potential by having children
219
220
PCM: Satisfactions of child rearing
Having my child be a success in life
EXITO
Playing with my child
JUGAR
Having my child contribute to society
SOCIED
Guiding and teaching my child
GUIAR
COMPART Sharing child raising with my husband (wife)
Experiencing the special love and closeness that a child provides
AMOR
PCM: Feeling needed and connected
Feeling needed and useful through my baby
UTIL
Having my child provide me with companionship and support later in
life
AYUDAV
Having a helpless baby to love and protect
DEPEND
COMPLET Feeling more complete as a woman (man) through my baby
Living a fuller, more enriched life through my baby
VIDA
PCM: Instrumental values of children
FERTIL Knowing that I am fertile
ADMIA Having my family and friends admire me with my baby
VARON Having a son
MUJER Having a daughter
PCM: Unscaled
HERMAN
Giving our first child a brother or sister
PCM: Additional items on modified CBQ
TRABAJE Having children that work for the family
APELLID *Having children to continue your family name (the family name of
your spouse)
HEREDAR *Having children to inherit your things
Having all the children that God sends
DIOS
NCM: Discomforts of pregnancy and childbirth
Experiencing (Seeing my wife experience) the discomforts of
SUFRIR pregnancy
DOLOR Experiencing (Seeing my wife experience) the pain of childbirth
NCM: Fears and worried of parenthood
Having an unhappy and poorly adjusted child
FELIZ
SEGURIDAD Worrying about the health and safety of my child
Having a baby who is born deformed
IMPEDIDO
Worrying whether I am raising my child the right way
CRIANDO
Having a child who embarrasses or disgraces the rest of the
family
AVERGU
Feeling guilty or inadequate as a mother (father)
INCAPAZ
221
NCM: Negatives of Childcare
Being kept from my (having my wife being kept from her) career
CARRERA
or job by a baby
EXIGENTE Being responsible for a needy and demanding baby
Spending time and energy involved in childcare
ENERGIA
Having to put up with the mess and noise that children make
RUIDO
ECONOMIC Burdening our family finances with a child
Taking care of a baby who is disagreeable and irritating
DESAGRA
Taking care of a sick child
ENFERMO
LIBERTAD Having a baby who takes away my freedom to do other things
NCM: Parental stress
DESCUIDAR Straining our marriage with a baby
Having a baby who strains my (wife’s) health
SALUD
Having a baby who is a burden to my husband (wife)
CARGA
Having a child who makes it necessary for me (my wife) to have a
job
DINERO
NCM: Unscaled
OTRONIN Having a baby who takes away from how much I can give my other
child
NCM: Additional Items on Modified CBQ
CASA *Having a child that makes it necessary to have a larger house?
VEJEZ Being a burden on my children in my old age?
APPENDIX B
SPANISH VERSION OF QUESTIONNAIRE FOR PHASE ONE
Lugar ______________________________________________
Fecha: ___________________
Hora: ____________________
Entrevistador: _____________________
# Hoy: ______
Consentimiento
Buenos días (tardes), mi nombre es ______________________, soy investigador
trabajando en un estudio de la Universidad de Florida y PREDISAN. Me gustaría hablar
con Ud. sobre cosas positivas y negativas sobre tener hijos. Intentamos aprender porque
algunas familias tienen muchos hijos y otras tienen pocos hijos aquí en Catacamas.
Yo lo seleccioné a usted entre personas que viven cerca o alrededor de Catacamas. Esta
entrevista durará aproximadamente 30 minutos.
Su participación es completamente voluntaria. A usted le haré una serie de preguntas que
puede interrumpir en cualquier momento sin consecuencia alguna.
Sus comentarios quedarán anónimos. Por lo que no hay riesgos de que participe, aunque
tampoco hay beneficios directos por su colaboración.
Si posteriormente tiene alguna duda Ud. puede contactarnos en PREDISAN.
PREDISAN tiene todos los datos de este estudio y la información sobre sus derechos
como participante.
222
223
[Solamente en casos de grabación de la entrevista:]
(Con su permiso, me gustaría grabar nuestra conversación. Si Ud. no desea que se grabe
la entrevista, todavía me gustaría charlar y anotar sus respuestas. Las cintas son para uso
exclusivo de la investigación y servirán para darnos una idea precisa de que nos ha dicho
la gente. No pondré su nombre adjunto; con la entrevista grabada se utilizará un código
para que nadie pueda reconocerlo. Guardaré las cintas en armario cerrado con llave en la
Universidad de Florida hasta que se complete la transcripción, y después las borraré.)
¿Puedo seguir y hacerle las preguntas?”
Si
No
224
( PARA MUJERES )
CUESTIONARIO SOBRE MOTIVACIÓN DE MATERNIDAD
Por favor valore cuanto le gusta lo siguiente.
0. No le gusta
1. Le gusta mas o menos
2. Le gusta bastante
_____a. ¿Cuánto le gusta...Sentir un niño moverse o patear dentro
de su cuerpo?
_____b. ¿Cuánto le gusta...Darle a su esposo la satisfacción ser
padre?
_____c. ¿Cuánto le gusta...Dar a luz un niño?
_____d. ¿Cuánto le gusta...Dar pecho a un niño?
_____e. ¿Cuánto le gusta...Chinear y abrazar un niño?
_____f. ¿Cuánto le gusta...Dedicarse mucho tiempo a criar hijos y
ser madre?
_____g. ¿Cuánto le gusta...Tener un hijo que continúe sus
costumbres familiares? (Por ejempleo...celebrar
navidad, cumpleaños, dia de madre, dia de finados)
_____h. ¿Cuánto le gusta...Ser el centro de una familia grande y
activa?
_____i. ¿Cuánto le gusta...Fortalecer su matrimonio a través de
un hijo?
_____j. ¿Cuánto le gusta...Cumplir sus deberes religiosos sobre
la vida familiar?
_____k. ¿Cuánto le gusta...Darle un nieto a sus padres? (que sus
padres puedan conocer sus nietos )
_____l. ¿Cuánto le gusta...Desarrollar su capacidad como persona
teniendo hijos?
_____m. ¿Cuánto le gusta...Tener un hijo que tenga éxito en la
vida?
_____n. ¿Cuánto le gusta...Jugar con su hijo?
225
_____o. ¿Cuánto le gusta...Contribuir a la sociedad a través de
tener un hijo?
_____p. ¿Cuánto le gusta...Guiar y enseñar a su hijo?
_____q. ¿Cuánto le gusta...Compartir la crianza de sus hijos con
su esposo?
_____r. ¿Cuánto le gusta...Experimentar el amor especial y la
cercanía que un hijo provee?
_____s. ¿Cuánto le gusta...Sentirse necesitada y útil a través de
su hijo?
_____t. ¿Cuánto le gusta...Disfrutar la compañía y ayuda de sus
hijos en su vejez?
_____u. ¿Cuánto le gusta...Tener un niño que dependa de usted
para su cuidado y protección?
_____v. ¿Cuánto le gusta...Sentirse mas completa como mujer a
través de sus hijos?
_____w. ¿Cuánto le gusta...Vivir una vida mas satisfactoria a
través de su hijo?
_____x. ¿Cuánto le gusta...Saber que usted es fértil?
_____y. ¿Cuánto le gusta...Tener la admiración de su familia y
amistades por medio de su hijo?
_____z. ¿Cuánto le gusta...Tener un hijo varón?
_____aa. ¿Cuánto le gusta...Tener una hija mujer?
_____bb. **¿Cuánto le gusta...Darle a su primer hijo un hermano o
una hermana?
_____cc. ¿Cuánto le gusta...Tener hijos que trabajen para usted?
_____dd. ¿Cuánto le gusta...Tener hijos que siguen el apellido de
su esposo?
_____ee. ¿Cuánto le gusta...Tener hijos para heredar sus cosas?
_____ff. ¿Cuánto le gusta...Tener todos los hijos que Dios le
mande?
226
( PARA MUJERES )
Por favor valore cuanto le molesta lo siguiente.
0. No le molesta
1. Le molesta mas o menos
2. Le molesta bastante
_____a. ¿Cuánto le molesta...Sufrir las molestias del embarazo?
_____b. ¿Cuánto le molesta...Sufrir con los dolores del parto?
_____c. ¿Cuánto le molesta...Tener un hijo que no es feliz y no
muy contento en la vida?
_____d. ¿Cuánto le molesta...Estar preocupada por la salud y
seguridad de sus hijos?
_____e. ¿Cuánto le molesta...Tener un hijo que nazca impedido o
retrasado?
_____f. ¿Cuánto le molesta...Estar preocupada de si está criando
a su hijo de la manera correcta?
_____g. ¿Cuánto le molesta...Tener un hijo que avergüence o
deshonre a la familia?
_____h. ¿Cuánto le molesta...Sentirse incapaz o inadecuada como
Madre?
_____i. ¿Cuánto le molesta...Tener que interrumpir su carrera o
su trabajo a causa de un hijo?
_____j. ¿Cuánto le molesta...Ser responsable por un niño
dependiente y exigente que necesita de su cuidado?
_____k. ¿Cuánto le molesta...Gastar tiempo y energía en criar
niños?
_____l. ¿Cuánto le molesta...Tener que aguantar el desorden y el
ruido que hacen los niños?
_____m. ¿Cuánto le molesta...Soportar la carga económica de la
familia con niños?
_____n. ¿Cuánto le molesta...Cuidar niños que son desagradables e
irritantes? (molestón)
_____o. ¿Cuánto le molesta...Tener que cuidar un niño enfermo?
227
_____p. ¿Cuánto le molesta...Tener niños que quita su libertad
para hacer otras cosas?
_____q. ¿Cuánto le molesta...Descuidar su matrimonio por cuidar
niños?
_____r. ¿Cuánto le molesta...Tener niños que afecta negativamente
su salud?
_____s. ¿Cuánto le molesta...Tener niños que sean carga para su
esposo?
_____t. ¿Cuánto le molesta...Tener niños que le obliguen a
trabajar afuera la casa o ganar dinero con trabajo
adentro la casa?
_____u. ¿Cuánto le molesta...Tener un niño que quita de lo que
usted puede dar a sus otros niños?
_____v. ¿Cuánto le molesta...Tener niños que le obliguen a tener
una casa más grande?
_____w. ¿Cuánto le molesta...Ser una carga para sus hijos en su
vejez?
228
( PARA HOMBRES )
CUESTIONARIO SOBRE MOTIVACIÓN DE PATERNIDAD
Por favor valore cuanto le gusta lo siguiente.
0. No le gusta
1. Le gusta mas o menos
2. Le gusta bastante
_____a.
¿Cuánto le gusta...Sentir un niño moverse o patear
dentro de su esposa?
_____b.
¿Cuánto le gusta...Darle a su esposa la satisfacción
de ser madre?
_____c.
¿Cuánto le gusta...Ayudar a su esposa dar a luz un
niño?
_____d.
¿Cuánto le gusta...Criar con un pepe a un niño?
_____e.
¿Cuánto le gusta...Chinear y abrasar un niño?
_____f.
¿Cuánto le gusta...Dedicarse mucho tiempo a criar
hijos y ser padre?
_____g.
¿Cuánto le gusta...Tener un hijo que continúe sus
custumbres familiares? (Por ejempleo...celebrar
navidad, cumpleaños, dia de madre, dia de finados)
_____h.
¿Cuánto le gusta...Ser el centro de una familia grande
y activa?
_____i.
¿Cuánto le gusta...Fortalecer su matrimonio a través
de sus hijos?
_____j.
¿Cuánto le gusta...Cumplir sus deberes religiosos
sobre la vida familiar?
_____k.
¿Cuánto le gusta...Darle un nieto a sus padres(que sus
padres puedan conocer sus nietos)
_____l.
¿Cuánto le gusta... Desarrollar su capacidad como
persona teniendo hijos?
_____m.
¿Cuánto le gusta...Tener un hijo que tenga éxito en la
vida?
_____n.
¿Cuánto le gusta...Jugar con sus hijos?
229
_____o.
¿Cuánto le gusta...Contribuir a la sociedad a través
de tener un hijo?
_____p.
¿Cuánto le gusta...Guiar y enseñar a sus hijos?
_____q.
¿Cuánto le gusta...Compartir la crianza de sus hijos
con su esposa?
_____r.
¿Cuánto le gusta...Experimentar el amor especial y la
cercanía que un hijo provee?
_____s.
¿Cuánto le gusta...Sentirse necesitado y útil a través
de sus hijos?
_____t.
¿Cuánto le gusta...Disfrutar la compañía y ayuda de
sus hijos en su vejez?
_____u.
¿Cuánto le gusta...Tener un niño que dependa de usted
para su cuidado y protección?
_____v.
¿Cuánto le gusta...Sentirse mas completo como hombre a
través de sus hijos?
_____w.
¿Cuánto le gusta...Vivir una vida mas satisfactoria a
través de sus hijos?
_____x.
¿Cuánto le gusta...Saber que usted es fértil?
_____y.
¿Cuánto le gusta...Tener la admiración de su familia y
amistades por medio de sus hijos?
_____z.
¿Cuánto le gusta...Tener un hijo varón?
_____aa.
¿Cuánto le gusta...Tener una hija mujer?
_____bb.
¿Cuánto le gusta...Darle a su primer hijo un hermano o
una hermana?
_____cc.
¿Cuánto le gusta...Tener un hijo que trabaje para
usted?
_____dd.
¿Cuánto le gusta...Tener un hijo que sigue su
apellido?
_____ee.
¿Cuánto le gusta...Tener un hijo para heredar sus
cosas?
_____ff.
¿Cuánto le gusta...Tener todos los hijos que Dios le
mande?
230
( PARA HOMBRES )
Por favor valore cuanto le molesta lo siguiente.
0. No le molesta
1. Le molesta mas o menos
2. Le molesta bastante
_____a. ¿Cuánto le molesta...Ver a su esposa sufrir las molestias
del embarazo?
_____b. ¿Cuánto le molesta...Ver a su esposa sufrir con los
dolores del parto?
_____c. ¿Cuánto le molesta...Tener un hijo que no es feliz y no
muy contento en la vida?
_____d. ¿Cuánto le molesta...Estar preocupado por la salud y
seguridad de su hijo?
_____e. ¿Cuánto le molesta...Tener un hijo que nazca impedido o
retrasado?
_____f. ¿Cuánto le molesta...Estar preocupado de si está criando
a su hijo de la manera correcta?
_____g. ¿Cuánto le molesta...Tener un hijo que avergüence o
deshonre a la familia?
_____h. ¿Cuánto le molesta...Sentirse incapaz o inadecuado como
Padre?
_____i. ¿Cuánto le molesta...Tener que interrumpir la carrera o
trabajo de su esposa a causa de un hijo?
_____j. ¿Cuánto le molesta...Ser responsable por un niño
dependiente y exigente que necesita de su cuidado?
_____k. ¿Cuánto le molesta...Gastar tiempo y energía en criar
niños?
_____l. ¿Cuánto le molesta...Tener que aguantar el desorden y el
ruido que hacen los niños?
_____m. ¿Cuánto le molesta...Soportar la carga económica de la
familia con niños?
_____n. ¿Cuánto le molesta...Cuidar un niño que es desagradable e
irritante? (molestón)
231
_____o. ¿Cuánto le molesta...Tener que cuidar un niño enfermo?
_____p. ¿Cuánto le molesta...Tener niños que quita su libertad
para hacer otras cosas?
_____q. ¿Cuánto le molesta...Descuidar su matrimonio por cuidar
niños?
_____r. ¿Cuánto le molesta...Tener niños que afecte negativamente
la salud de su esposa?
_____s. ¿Cuánto le molesta...Tener niños que sean carga para su
esposa?
_____t. ¿Cuánto le molesta...Tener niños que obligue a su esposa
a trabajar afuera de la casa o ganar dinero con
trabajo adentro la casa
_____u. ¿Cuánto le molesta...Tener un niño que quita de lo que
usted puede dar a sus otros niños?
_____v. ¿Cuánto le molesta...Tener niños que le obligue a tener
una casa más grande?
_____w. ¿Cuánto le molesta...Ser una carga para sus hijos en su
vejez?
232
Demográfico
ENCUESTA DE HOMBRE / MUJER
1. ¿Dónde vive? ________________________
Es un: a. Barrio de Catacamas
b. Aldea de Catacamas
c. Otro ________________________
2. ¿Qué tipo de agua tiene en su vivienda?
1. Llave dentro de la vivienda
2. Llave fuera de la vivienda pero dentro de la propiedad
3. Llave fuera de la propiedad a menos de 100 metros
4. Llave fuera de la propiedad a 100 metros o más
5. Fuente natural: río, quebrada, naciente, vertiente, lago
6. Pozo malacate (sin bomba)
7. Poso con bomba (eléctrica o manual)
8. La compran / carro cisterna
9. Fuente de agua protegida
10. Otro ______________________________
3. ¿Qué clase de servicio sanitario tiene en su vivienda?
1.
2.
3.
4.
5.
Inodoro (lavable)
Letrina hidráulica / tasa campesina
Letrina / fosa simple
No tiene / al aire libre
Otro ___________________________
4. En su casa hay
a. Luz elélectrica?
b. Radio?
c. Televisión?
d. Refrigeradora?
e. Teléfono?
f. Vehículo propio? (con motor)
5. ¿Con que cocinan ustedes?
1. Electricidad
2. Gas butano
3. Carbon
4. Gas kerosene
5. Leña
6. No cocina
7. Otro ________________________________
6. ¿Cuántos años complidos tiene Ud.?
______ Edad
7. ¿Cual fue el grado o año más alto que Ud. aprobó en la escuela, colegio o universidad?
0. Ninguno
1. Primaria
2. Secundaria
3. Universidad
4. Alfabetización
0
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
233
8. ¿Cual es su estado civil actual?
1. Casada
2. Unida / unión libre
3. Divorciada
4. Separada
5. Viuda
6. Soltera
9. ¿Cuál es su religión?
_____a.
_____b.
_____c.
_____d.
Católica
Evangélica
No tiene
Otra ________________
10. ¿Ha tenido algún hijo que nació vivo?
1. Si
2. No
11. ¿Cuántos de sus hijos (varones o mujeres) viven con usted en su casa? ____________
12. ¿Cuántos de sus hijos (varones o mujeres) viven en otra parte? ____________
13. ¿Ha tenido algún niño(a) que nació vivo y murió después, incluyendo algún hijo(a) que nació
vivo(a) y murió solo minutos y horas después?
1. Si
2. No
14. ¿Cuántos de sus hijos que nacieron vivos se le murieron después? ____________
15. ¿Cuantos hermanos (varones y mujeres) tiene, incluyendo los que murieron? ____________
16. ¿Cuantos hermanos menores de Ud. ? ____________
17. ¿Realiza o hace usted algún trabajo o actividad por el cual recibe pago en dinero ó en otra
forma?
1. Si (anote actividad:
_______________________________________)
2. No
17b. ¿Cuántos, mas o menos, gana en dinero en un mes?
1. Menos que 3,000 Lempiras
2. 3,000–9,999 Lempiras
3. 10,000 Lempiras o mas
No Sé
No responde
234
18. ¿Realiza o hace su esposo(a) algún trabajo o actividad por el cual recibe pago en dinero ó en
otra forma?
1. Si (anote actividad:
__________________________________________)
2. No
18b. ¿Cuántos, mas o menos, gana en dinero en un mes?
1. Menos que 3,000 Lempiras
2. 3,000–9,999 Lempiras
3. 10,000 Lempiras o mas
No Sé
No responde
18c. (SI LOS DOS TRABAJEN) ¿Cuántos, mas o menos, ganan ustedes en dinero en un mes?
1. Menos que 3,000 Lempiras
2. 3,000–9,999 Lempiras
3. 10,000 Lempiras o mas
No Sé
No responde
19. ¿Si pudiera elegir exactamente el número de hijos que tendría en todo su vida, cuántos serían?
__ __ hijos
Los que Dios me mande
Los que tengo
No sé
20. ¿Cuántos hijos cree usted que deberían de tener las parejas que estan empezando a
formar una familia ahora?
__ __ hijos
Los que Dios les mande
No sé
21. ¿Si tenemos más preguntas en el futuro, podemos contactarle para hablar más?
1. Si
2. No
22. ¿Cómo se llama, usted?
____________________________________________________________________
23. ¿Donde podemos contactarle?
DIRECCIÓN CON PUNTO DE REFERENCIA:
____________________________________________________________________
APPENDIX C
ENGLISH VERSION OF QUESTIONNAIRE FOR PHASE ONE
Informed Consent
Good morning (afternoon), my name is ______________________, I’m an investigator
working in a study for the University of Florida and PREDISAN. I would like to speak
with you about positive and negative aspects of having children.. We intend to learn why
some families have many children and why other have few children here in Catacamas..
You were selected at random from people who live here in Catacamas and around
Catacamas. The interview will last approximately 30 minutes.
Your participation is completely voluntary. I will ask you a series of questions that you
can interrupt en whatever moment with no negative consequences.
Your comments will be anonymous. We will use a code number for each interview and
your name will never appear associated with any of the data. There are no positive
benefits and no negative risks associated with your collaboration.
If, after the interview, you have any doubt or question, you can contact us at PREDISAN.
PREDISAN has all the information about this study and they have information about how
you can find out about your rights as a research participant.
[Only in cases of tape recorded interviews:]
(With your permission, I would like to tape record our conversation. If you would no
like to have the interview taped, I would still like to speak with you and write your
answers. The tapes are for exclusive use of the investigation and will give us an idea
precisely of what is said during interviews. We will not put your name with the tape and
we will use a code number for each tape so that no one will recognize it. We will guard
235
236
the tapes in a locked box in the University of Florida until a transcription is completed.
Afterwards, the tapes will be erased.)
¿Can I begin and ask you questions?”
Yes
No (record why________________________________________)
237
CHILDBEARING MOTIVATION QUESTIONNAIRE
Please rate how desirable are the following
0.
1.
2.
Not desirable
Somewhat desirable
Very desirable
_____a.
How desirable is...Feeling a baby move and kick
inside your body (your wife’s body)?
_____b.
How desirable is...Giving your spouse the
satisfaction of fatherhood (motherhood)?
_____c.
How desirable is...Giving birth to a baby
(Helping your wife give birth to a baby)?
_____d.
How desirable is...Breast feeding a baby (bottle
feeding a baby)?
_____e.
How desirable is...Holding and cuddling a baby?
_____f.
How desirable is...Devoting yourself and much of
your time to raising children and being a mother
(father)?
_____g.
How desirable is...Having a child who will carry
on your family traditions?
_____h.
How desirable is...Being the center of a large,
active family?
_____i.
How desirable is...Strengthening your marriage
through a child?
_____j.
How desirable is...Fulfilling your religious
feelings about family life?
_____k.
How desirable is...Providing your parents with a
grandchild?
_____l.
How desirable is...Fulfilling your potential by
having children?
_____m.
How desirable is...Having a child be a success in
life?
_____n.
_____o.
How desirable is...Playing with your child?
How desirable is...Having your child contribute
to society?
238
_____p.
How desirable is...Guiding and teaching your
child?
_____q.
How desirable is...Sharing childraising with your
spouse?
_____r.
How desirable is...Experiencing the special love
and closeness that a child provides?
_____s.
How desirable is...Feeling needed and useful
through your child?
_____t.
How desirable is...Having my child provide me
with companionship and support later in life?
_____u.
How desirable is...Having a helpless baby to love
and protect?
_____v.
How desirable is...Feeling more complete as a
woman (man) through my baby?
_____w.
How desirable is...Living a fuller, more enriched
life through my child?
_____x.
_____y.
How desirable is...Knowing that you are fertile?
How desirable is...having my family and friends
admire me with my baby?
_____z.
How desirable is...Having a son?
_____aa.
How desirable is...Having a daughter?
_____bb.
How desirable is...Giving your first child a
brother or sister?
_____cc.
How desirable is...Having children that work for
the family?
_____dd.
How desirable is...Having children to continue
your family name (the family name of your spouse)?
_____ee.
How desirable is...Having children to inherit
your things?
_____ff.
How desirable is...Having all the children that
God sends?
239
Please rate how undesirable are the following.
0.
1.
2.
Not undesirable
Somewhat undesirable
Very undesirable
_____a. How undesirable is...Experiencing (seeing my wife
experience) the discomforts of pregnancy?
_____b. How undesirable is...Experiencing (seeing my wife
experience) the pain of child birth?
_____c. How undesirable is...Having an unhappy and poorly
adjusted child?
_____d. How undesirable is...Worrying about the health and safety
of your child?
_____e. How undesirable is...Having a baby who is born deformed?
_____f. How undesirable is... Worrying whether I am raising my
child the right way?
_____g. How undesirable is... Having a child who embarrasses or
disgraces the rest of the family?
_____h. How undesirable is... Feeling guilty or inadequate as a
parent?
_____i. How undesirable is... Being kept from my (having my wife
being kept from her) career or job by a baby?
_____j. How undesirable is... Being responsible for a needy and
demanding baby?
_____k. How undesirable is... Spending time and energy involved
in childcare?
_____l. How undesirable is... Having to put up with the mess and
noise that children make?
_____m. How undesirable is... Burdening our family finances with
a child?
_____n. How undesirable is... Taking care of a baby who is
disagreeable and irritating?
_____o. How undesirable is... Taking care of a sick child?
240
_____p. How undesirable is...Having a baby who takes away my
freedom to do other things?
_____q. How undesirable is...Straining your marriage with a baby?
_____r. How undesirable is...Having a baby who strains my
(wife’s) health?
_____s. How undesirable is... Having a child who is a burden to
my spouse?
_____t. How undesirable is... Having a child who makes it
necessary for me (my wife) to have a job?
_____u. How undesirable is... Having a baby who takes away from
how much I can give to my other child?
_____v. How undesirable is...Having a child that makes it
necessary to have a larger house?
_____w. How undesirable is...Being a burden on my children in my
old age?
241
Demographics
Questionnaire of MEN / WOMEN
1. Where do you live? ________________________
It is a : a. Neighborhood of Catacamas
b. Village of Catacamas
c. Other ________________________
2. What type of water do you have in your living area?
1. Tap inside
2. Tap outside but close
3. Tap off of the property under 100 meters away
4. Tap off of the property over 100 meters away
5. Natural source
6. Well without pump
7. Well with pump
8. Purchased
9. Protected source of water
10. Other______________________________
3. ¿What type of sanitary services do you have?
1.
2.
3.
4.
5.
Indoor
Latrine with water
Latrine without water
Open field
Other ___________________________
4. In your house there is...
a. Electricity?
b. Radio?
c. Television?
d. Refrigerator?
e. Telephone?
f. Vehicle
5. ¿With what do you cook?
1. Electricity
2. Gas butane
3. Charcoal
4. Kerosene
5. Wood
6. Don’t cook
7. Other ________________________________
242
6. ¿How old are you?
___ ___
7. ¿What was the highest year or grade you finished in school?
0. None
1. Primary
2. Secondary
3. University
4. Reading and Writing
0
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
8. ¿What is your marital status?
1.
2.
3.
4.
5.
6.
Married
Life partners
Divorced
Separated
Widow
Single
9. What is your religion?
1.
2.
3.
4.
Catholic
Evangelical
None
Other ________________
10. ¿Have you had a child who was born alive?
1. Yes
2. No
11. ¿How many of your children live with you? ____________
12. ¿How many live somewhere else? ____________
13. ¿Have you had a child born alive which died later?
1. Yes
2. No
14. ¿How many? ____________
15. ¿How many siblings do you have? ____________
16. How many younger ? ____________
17. Do you have some type of work or activity for which you receive pay in money or some other
form?
243
1. Yes _____________________
2. No
17b. How much do you make in one month?
1.
2.
Less than 3,000 Lempiras
3,000–9,999 Lempiras
Don’t know
Refused
18. Does your spouse have some type of work or activity for which she/he receives pay in money
or some other form?
1. Yes (specify activity: _____________________)
2. No
18b. How much, more or less, does she/he make in one month?
1.
2.
Less than 3,000 Lempiras
3,000–9,999 Lempiras
Don’t know
Refused
19. ¿If you could decide exactly the number of children to have in your life, how many would it
be?
__ __ children
As many as God sends
The number I have
Don’t know
20. ¿If a couple was beginning their family now, how many children should they have?
__ __ children
As many as God sends
Don’t know
21. ¿If we have more questions later, can we contact you to speak more?
1. Yes
2. No
22. What is your name?
___________________________________________________________________
23. Where can we contact you?
ADDRESS WITH POINT OF REFERENCE:
_____________________________________________________________________
APPENDIX D
SPANISH VERSION OF QUESTIONNAIRE FOR PHASE TWO
Lugar
_______________________________________________________________________
Barrio: _______________
Manzana: ____
Fecha: ___________________
Casa:_____
Hora: ____________________
Entrevistador: _____________________
# Hoy: ______
Encuesta de
a. hombre
b. mujer
Consentimiento
Buenos días (tardes), mi nombre es ______________________, soy investigador
trabajando en un estudio de la Universidad de Florida y PREDISAN. Me gustaría hablar
con Ud. sobre cosas positivas y negativas sobre tener hijos. Intentamos aprender porque
algunas familias tienen muchos hijos y otras tienen pocos hijos aquí en Catacamas.
Yo lo seleccioné a usted entre personas que viven cerca o alrededor de Catacamas. Esta
entrevista durará aproximadamente 30 minutos.
Su participación es completamente voluntaria. A usted le haré una serie de preguntas que
puede interrumpir en cualquier momento sin consecuencia alguna.
Sus comentarios quedarán anónimos. Por lo que no hay riesgos de que participe, aunque
tampoco hay beneficios directos por su colaboración.
244
245
Si posteriormente tiene alguna duda Ud. puede contactarnos en PREDISAN.
PREDISAN tiene todos los datos de este estudio y la información sobre sus derechos
como participante.
¿Puedo seguir y hacerle las preguntas?”
Si
No
246
( PARA MUJERES )
CUESTIONARIO SOBRE MOTIVACIÓN DE MATERNIDAD
Por favor valore cuanto le gusta lo siguiente.
0.
1.
2.
No le gusta
Le gusta mas o menos
Le gusta bastante
_____a.
¿Cuánto le gusta...Dedicarse mucho tiempo a criar hijos
y ser madre?
_____b.
¿Cuánto le gusta...Tener hijos que trabajen para usted?
_____c.
¿Cuánto le gusta...Disfrutar la compañía y ayuda de sus
hijos en su vejez?
_____d.
¿Cuánto le gusta...Vivir una vida mas satisfactoria a
través de su hijo?
_____e.
¿Cuánto le gusta...Tener un niño que dependa de usted
para su cuidado y protección?
_____f.
¿Cuánto le gusta...Compartir la crianza de sus hijos
con su esposo?
_____g.
¿Cuánto le gusta...Guiar y enseñar a su hijo?
_____h.
¿Cuánto le gusta...Experimentar el amor especial y la
cercanía que un hijo provee?
_____i.
¿Cuánto le gusta...Ser el centro de una familia grande
y activa?
_____j.
¿Cuánto le gusta...Tener todos los hijos que Dios le
mande?
_____k.
¿Cuánto le gusta...Sentirse mas completa como mujer a
través de sus hijos?
_____l.
¿Cuánto le gusta...Jugar con su hijo?
_____m.
¿Cuánto le gusta...Dar a luz un niño?
_____n.
¿Cuánto le gusta...Saber que usted es fértil?
_____o.
¿Cuánto le gusta...Tener un hijo que continúe sus
costumbres familiares? (Por ejemplo...celebrar navidad,
cumpleaños, día de madre, día de finados)
247
_____p.
¿Cuánto le gusta...Sentir un niño moverse o patear
dentro de su cuerpo?
_____q.
¿Cuánto le gusta...Darle a su esposo la satisfacción
ser padre?
_____r.
¿Cuánto le gusta...Dar pecho a un niño?
_____s.
¿Cuánto le gusta...Darle un nieto a sus padres? (que
sus padres puedan conocer sus nietos )
_____t.
¿Cuánto le gusta...Fortalecer su matrimonio a través de
un hijo?
_____u.
¿Cuánto le gusta...Tener un hijo varón?
_____v.
¿Cuánto le gusta...Tener la admiración de su familia y
amistades por medio de su hijo?
_____w.
¿Cuánto le gusta...Chinear y abrazar un niño?
_____x.
¿Cuánto le gusta...Contribuir a la sociedad a través de
tener un hijo?
_____y.
¿Cuánto le gusta...Tener una hija mujer?
_____z.
¿Cuánto le gusta...Cumplir sus deberes religiosos sobre
la vida familiar?
_____aa.
¿Cuánto le gusta...Desarrollar su capacidad como
persona teniendo hijos?
_____bb.
¿Cuánto le gusta...Sentirse necesitada y útil a través
de su hijo?
248
( PARA MUJERES )
Por favor valore cuanto le molesta lo siguiente.
0.
1.
2.
No le molesta
Le molesta mas o menos
Le molesta bastante
_____a.
¿Cuánto le molesta...Tener que aguantar el desorden y
el ruido que hacen los niños?
_____b.
¿Cuánto le molesta...Sentirse incapaz o inadecuada como
Madre?
_____c.
¿Cuánto le molesta...Soportar la carga económica de la
familia con niños?
_____d.
¿Cuánto le molesta...Estar preocupada de si está
criando a su hijo de la manera correcta?
_____e.
¿Cuánto le molesta...Tener niños que quita su libertad
para hacer otras cosas?
_____f.
¿Cuánto le molesta...Tener niños que sean carga para su
esposo?
_____g.
¿Cuánto le molesta...Ser responsable por un niño
dependiente y exigente que necesita de su cuidado?
_____h.
¿Cuánto le molesta...Tener un hijo que no es feliz y no
muy contento en la vida?
_____i.
¿Cuánto le molesta...Tener un hijo que avergüence o
deshonre a la familia?
_____j.
¿Cuánto le molesta...Cuidar niños que son desagradables
e irritantes?
_____k.
¿Cuánto le molesta...Sufrir con los dolores del parto?
_____l.
¿Cuánto le molesta...Estar preocupada por la salud y
seguridad de sus hijos?
_____m.
¿Cuánto le molesta...Ser una carga para sus hijos en su
vejez?
_____n.
¿Cuánto le molesta...Tener un hijo que nazca impedido o
retrasado?
249
_____o.
¿Cuánto le molesta...Tener un niño que quita de lo que
usted puede dar a sus otros niños?
_____p.
¿Cuánto le molesta...Tener niños que afecta
negativamente su salud?
_____q.
¿Cuánto le molesta...Tener que interrumpir su carrera o
su trabajo a causa de un hijo?
_____r.
¿Cuánto le molesta...Descuidar su matrimonio por cuidar
niños?
_____s.
¿Cuánto le molesta...Tener niños que le obliguen a
trabajar afuera la casa o ganar dinero con trabajo
adentro la casa?
_____t.
¿Cuánto le molesta...Sufrir las molestias del embarazo?
_____u.
¿Cuánto le molesta...Gastar tiempo y energía en criar
niños?
_____v.
¿Cuánto le molesta...Tener que cuidar un niño enfermo?
250
( PARA HOMBRES )
CUESTIONARIO SOBRE MOTIVACIÓN DE PATERNIDAD
Por favor valore cuanto le gusta lo siguiente.
0.
1.
2.
No le gusta
Le gusta mas o menos
Le gusta bastante
_____a.
¿Cuánto le gusta...Dedicarse mucho tiempo a criar hijos
y ser padre?
_____b.
¿Cuánto le gusta...Tener hijos que trabajen para usted?
_____c.
¿Cuánto le gusta...Disfrutar la compañía y ayuda de sus
hijos en su vejez?
_____d.
¿Cuánto le gusta...Vivir una vida mas satisfactoria a
través de su hijo?
_____e.
¿Cuánto le gusta...Tener un niño que dependa de usted
para su cuidado y protección?
_____f.
¿Cuánto le gusta...Compartir la crianza de sus hijos
con su esposa?
_____g.
¿Cuánto le gusta...Guiar y enseñar a su hijo?
_____h.
¿Cuánto le gusta...Experimentar el amor especial y la
cercanía que un hijo provee?
_____i.
¿Cuánto le gusta...Ser el centro de una familia grande
y activa?
_____j.
¿Cuánto le gusta...Tener todos los hijos que Dios le
mande?
_____k.
¿Cuánto le gusta...Sentirse mas completa como hombre a
través de sus hijos?
_____l.
¿Cuánto le gusta...Jugar con su hijo?
_____m.
¿Cuánto le gusta...Ayudar a su esposa a dar a luz un
niño?
_____n.
¿Cuánto le gusta...Saber que usted es fértil?
_____o.
¿Cuánto le gusta...Tener un hijo que continúe sus
costumbres familiares? (Por ejemplo...celebrar navidad,
cumpleaños, día de madre, día de finados)
251
_____p.
¿Cuánto le gusta...Sentir un niño moverse o patear
dentro de su esposa?
_____q.
¿Cuánto le gusta...Darle a su esposa la satisfacción
ser madre?
_____r.
¿Cuánto le gusta...Criar con un pepe a un niño?
_____s.
¿Cuánto le gusta...Darle un nieto a sus padres? (que
sus padres puedan conocer sus nietos )
_____t.
¿Cuánto le gusta...Fortalecer su matrimonio a través de
un hijo?
_____u.
¿Cuánto le gusta...Tener un hijo varón?
_____v.
¿Cuánto le gusta...Tener la admiración de su familia y
amistades por medio de su hijo?
_____w.
¿Cuánto le gusta...Chinear y abrazar un niño?
_____x.
¿Cuánto le gusta...Contribuir a la sociedad a través de
tener un hijo?
_____y.
¿Cuánto le gusta...Tener una hija mujer?
_____z.
¿Cuánto le gusta...Cumplir sus deberes religiosos sobre
la vida familiar?
_____aa.
¿Cuánto le gusta...Desarrollar su capacidad como
persona teniendo hijos?
_____bb. ¿Cuánto le gusta...Sentirse necesitada y útil a través
de su hijo?
252
( PARA HOMBRES )
Por favor valore cuanto le molesta lo siguiente.
0.
1.
2.
No le molesta
Le molesta mas o menos
Le molesta bastante
_____a. ¿Cuánto le molesta...Tener que aguantar el desorden y el
ruido que hacen los niños?
_____b. ¿Cuánto le molesta...Sentirse incapaz o inadecuada como
Padre?
_____c. ¿Cuánto le molesta...Soportar la carga económica de la
familia con niños?
_____d. ¿Cuánto le molesta...Estar preocupada de si está criando a
su hijo de la manera correcta?
_____e. ¿Cuánto le molesta...Tener niños que quita su libertad
para hacer otras cosas?
_____f. ¿Cuánto le molesta...Tener niños que sean carga para su
esposa?
_____g. ¿Cuánto le molesta...Ser responsable por un niño
dependiente y exigente que necesita de su cuidado?
_____h. ¿Cuánto le molesta...Tener un hijo que no es feliz y no
muy contento en la vida?
_____i. ¿Cuánto le molesta...Tener un hijo que avergüence o
deshonre a la familia?
_____j. ¿Cuánto le molesta...Cuidar niños que son desagradables e
irritantes?
_____k. ¿Cuánto le molesta...Ver a su esposa sufrir con los
dolores del parto?
_____l. ¿Cuánto le molesta...Estar preocupada por la salud y
seguridad de sus hijos?
_____m. ¿Cuánto le molesta...Ser una carga para sus hijos en su
vejez?
_____n. ¿Cuánto le molesta...Tener un hijo que nazca impedido o
retrasado?
253
_____o. ¿Cuánto le molesta...Tener un niño que quita de lo que
usted puede dar a sus otros niños?
_____p. ¿Cuánto le molesta...Tener niños que afecta negativamente
la salud de su esposa?
_____q. ¿Cuánto le molesta...Tener que interrumpir la carrera o
trabajo de su esposa a causa de un hijo?
_____r. ¿Cuánto le molesta...Descuidar su matrimonio por cuidar
niños?
_____s. ¿Cuánto le molesta...Tener niños que obligue a su esposa a
trabajar afuera de la casa o ganar dinero con trabajo
adentro la casa?
_____t. ¿Cuánto le molesta...Ver a su esposa sufrir las molestias
del embarazo?
_____u. ¿Cuánto le molesta...Gastar tiempo y energía en criar
niños?
_____v. ¿Cuánto le molesta...Tener que cuidar un niño enfermo?
254
Ahora, quisiera solicitar su opinión sobre algunos temas que nos
interesan.
1. ¿Si pudiera elegir exactamente el número de hijos que tendría
en todo su vida, cuántos serían?
__ __ hijos
Los que Dios me mande
Los que tengo
No sé
2. ¿Cuántos hijos cree usted que deberían de tener las parejas
que están empezando a formar una familia ahora?
__ __ hijos
Los que Dios les mande
No sé
3. Con cuantos hijos cree usted que una familia es:
Grande: __________en adelante
Pequeño: menos que __________
4. Ahora, le voy a preguntar sobre obligaciones y actividades de
hombres y mujeres en una pareja o en un hogar. ¿Quién debería
decidir el número de niños que una pareja debería tener? ¿El
hombre, La mujer, los dos, o cualquier otra persona?
NO LEER.
1.
2.
3.
4.
5.
6.
7.
8.
9.
Esposo/ compañero/ hombre
Esposa/ compañera/ mujer
Los dos juntos
Cualquier de los dos
El Médico
Lo que pase / lo que mande Dios
Otro ____________________
Especifique
NO SABE
5. ¿Quién debería decidir si una pareja utiliza métodos de
planificación familiar?
NO LEER
a)
b)
c)
d)
e)
f)
g)
Esposo/ compañero/ hombre
Esposa/ compañera/ mujer
Los dos juntos
Cualquier de los dos
El Médico
Otro ____________________
Especifique
NO SABE
255
6. ¿Quién debería soportar la carga económica de la familia?
NO LEER
a)
b)
c)
d)
e)
f)
Esposo/
Esposa/
Los dos
Æ
compañero/ hombre
compañera/ mujer
juntos--SI LOS DOS, LEA: a) ¿Los dos iguales? O...
b) ¿La mayoría el hombre? O..
c) ¿La mayoría la mujer?
Cualquier de los dos
Otro ____________________
Especifique
NO SABE
7. ¿Quién debería cuidar la casa de la familia?
NO LEER
a)
b)
c)
d)
e)
f)
Esposo/
Esposa/
Los dos
Æ
compañero/ hombre
compañera/ mujer
juntos--SI LOS DOS, LEA: a) ¿Los dos iguales? O...
b) ¿La mayoría el hombre? O..
c) ¿La mayoría la mujer?
Cualquier de los dos
Otro ____________________
Especifique
NO SABE
8. ¿Quién debería cuidar los niños de la familia?
a)
b)
c)
d)
e)
f)
Esposo/
Esposa/
Los dos
Æ
compañero/ hombre
compañera/ mujer
juntos--SI LOS DOS, LEA: a) ¿Los dos iguales? O...
b) ¿La mayoría el hombre? O..
c) ¿La mayoría la mujer?
Cualquier de los dos
Otro ____________________
Especifique
NO SABE
9. ¿Ud. Cree que es aceptable para hombres unidos tener
relaciones extramaritales?
a) Nunca
b) A veces
c) Siempre
d) NO SABE
10. ¿Ud. Cree que es aceptable para mujeres unidos tener
relaciones extramaritales?
256
a) Nunca
b) A veces
c) Siempre
d) NO SABE
11. Ahora, voy a mencionar una serie de tipos de actividad.
¿Quien cree usted que debe desarrollar el siguiente trabajo:
Hombres, Mujeres, o ambos?
Mujeres
Hombres
Ambos
a
b
c
d
e
f
g
h
I
j
k
l
m
n
o
p
q
r
s
t
u
v
w
x
y
z
aa
bb
cc
dd
ee
Jornalero
Maestro
Peluquería
Lava Ropa
Perito Mercantil
Ganadero
Vendedor ambulante
Motorista
Agricultura
Pulpería
Hortaliza
Albañil
Obrero
Secretaria
Venta de Golosinas
Agente de Policía
Periodista
Sastrería
Venta de helados
Pastor Evangélico
Empacadora de carne
Taxista
Vigilancia
Médico
Mecánico
Lic. en Derecho
Comerciante
Fabrica
Trabajo Domestico
Dentista
Enfermera
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
257
Ahora, le voy hacer algunas preguntas acerca de la planificación
familiar, es decir de las cosas que usan las parejas para evitar
que la mujer quede embarazada.
a. La píldora o pastillas
anticonceptivas
b. El DIU ó dispositivo
c. El condón
d. La inyección
e. Los métodos vaginales
jaleas, espumas,
cremas, óvulos,
tabletas.
f. La esterilización
femenina u operación
g. La vasectomía ó
esterilización
masculina
h. Ritmo o calendario
i. Billings (moco
cervical)
j. Retiro (escupir fuera)
k. Norplant
10. Ha oído
hablar de:
Si
No
11. Alguna vez ha
usado?
Si
No
a)
b)
c)
d)
1
1
1
1
2
2
2
2
a)
b)
c)
d)
1
1
1
1
2
2
2
2
e)
1
2
e)
1
2
f)
1
2
f)
1
2
g)
h)
1
1
2
2
g)
h)
1
1
2
2
i)
j)
k)
1
1
1
2
2
2
i)
j)
k)
1
1
1
2
2
2
Demográfico
1. ¿Dónde nació?
Es un:
________________________
a. Barrio de Catacamas
b. Aldea de Catacamas
c. Otro ________________________
2. ¿Qué tipo de agua tiene en su vivienda?
1. Llave dentro de la vivienda
2. Llave fuera de la vivienda pero dentro de la propiedad
3. Llave fuera de la propiedad a menos de 100 metros
4. Llave fuera de la propiedad a 100 metros o más
5. Fuente natural: río, quebrada, naciente, vertiente, lago
6. Pozo malacate (sin bomba)
7. Poso con bomba (eléctrica o manual)
8. La compran / carro cisterna
9. Fuente de agua protegida
10. Otro ______________________________
258
3. ¿Qué clase de servicio sanitario tiene en su vivienda?
1.
2.
3.
4.
5.
Inodoro (lavable)
Letrina hidráulica / tasa campesina
Letrina / fosa simple
No tiene / al aire libre
Otro ___________________________
4. En su casa hay
a. ¿Luz eléctrica? (o planta)
b. ¿Radio?
c. ¿Televisión?
motor)
d. ¿Refrigeradora?
e. ¿Teléfono?
f. ¿Vehículo propio? (con
5. ¿Con que cocinan ustedes?
1. Electricidad
5. Leña
2. Gas butano
6. No cocina
3. Carbón
7. Otro
________________________________
4. Gas kerosene
6. ¿Cuántos años cumplidos tiene Ud.?
______
años de edad
7a. ¿Cuál fue el grado o año más alto que Ud. aprobó en la
escuela, colegio o universidad?
0.
1.
2.
3.
4.
Ninguno
Primaria
Secundaria
Universidad
Alfabetización
0
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
7b. ¿Cuál fue el grado o año más alto que su ESPOSO(A) aprobó en
la escuela, colegio o universidad?
0.
1.
2.
3.
4.
Ninguno
Primaria
Secundaria
Universidad
Alfabetización
0
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
7c. ¿Cuál fue el grado o año más alto que su MADRE aprobó en la
escuela, colegio o universidad?
0.
1.
2.
3.
4.
Ninguno
Primaria
Secundaria
Universidad
Alfabetización
0
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
259
8. ¿Cuál es su estado civil actual?
1. Casada
4. Separada
2. Unida / unión libre
5. Viuda
3. Divorciada
6. Soltera
9a. ¿Cuál es su religión?
1.
2.
3.
4.
Católica
Evangélica
No tiene
Otra ________________
9b. ¿Usted como se considera?
1.
2.
3.
4.
(LEER LAS OPCIONES)
Muy religiosa
Religiosa
Poco religiosa
Nada religiosa
10. ¿Ha tenido algún hijo que nació vivo?
1. Sí
2. No
11. ¿Cuántos de sus hijos (varones o mujeres) viven con usted en
su casa? ____________
12. ¿Cuántos de sus hijos (varones o mujeres) viven en otra
parte?_____
13. ¿Ha tenido algún niño(a) que nació vivo y murió después,
incluyendo algún hijo(a) que nació vivo(a) y murió solo
minutos y horas después?
1. Sí
2. No
14. ¿Cuántos de sus hijos que nacieron vivos se le murieron
después?____
15. ¿Realiza o hace usted algún trabajo o actividad por el cual
recibe pago en dinero ó en otra forma?
1. Si
2. No
(anote actividad:_________________________________)
15b. ¿Cuántos, mas o menos, gana en dinero en un mes?
1. Menos que 3,000 Lempiras
260
2.
3.
4.
5.
3,000–9,999 Lempiras
10,000 Lempiras o mas
No Sé
No responde
16. ¿Realiza o hace su esposo(a) algún trabajo o actividad por el
cual recibe pago en dinero ó en otra forma?
1. Si
2. No
(anote actividad:_____________________________________)
16b. ¿Cuántos, mas o menos, gana en dinero en un mes?
1.
2.
3.
4.
5.
Menos que 3,000 Lempiras
3,000–9,999 Lempiras
10,000 Lempiras o más
No Sé
No responde
17. (SI LOS DOS TRABAJEN) ¿Cuántos, mas o menos, ganan ustedes en
dinero en un mes?
1.
2.
3.
4.
5.
Menos que 3,000 Lempiras
3,000–9,999 Lempiras
10,000 Lempiras o más
No Sé
No responde
18. ¿Si tenemos más preguntas en el futuro, podemos contactarle
para hablar más?
1. Si
2. No
19. ¿Cómo se llama, usted?
________________________________________________________________
20. ¿Donde podemos contactarle?
DIRECCIÓN CON PUNTO DE REFERENCIA:
______________________________________________________________
APPENDIX E
ENGLISH VERSION OF QUESTIONNAIRE FOR PHASE TWO
Informed Consent
Good morning (afternoon), my name is ______________________, I’m an investigator
working in a study for the University of Florida and PREDISAN. I would like to speak
with you about positive and negative aspects of having children.. We intend to learn why
some families have many children and why other have few children here in Catacamas..
You were selected at random from people who live here in Catacamas and around
Catacamas. The interview will last approximately 30 minutes.
Your participation is completely voluntary. I will ask you a series of questions that you
can interrupt en whatever moment with no negative consequences.
Your comments will be anonymous. We will use a code number for each interview and
your name will never appear associated with any of the data. There are no positive
benefits and no negative risks associated with your collaboration.
If, after the interview, you have any doubt or question, you can contact us at PREDISAN.
PREDISAN has all the information about this study and they have information about how
you can find out about your rights as a research participant.
[Only in cases of tape recorded interviews:]
(With your permission, I would like to tape record our conversation. If you would no
like to have the interview taped, I would still like to speak with you and write your
answers. The tapes are for exclusive use of the investigation and will give us an idea
precisely of what is said during interviews. We will not put your name with the tape and
we will use a code number for each tape so that no one will recognize it. We will guard
261
262
the tapes in a locked box in the University of Florida until a transcription is completed.
Afterwards, the tapes will be erased.)
¿Can I begin and ask you questions?”
Yes
No (record why________________________________________)
263
CHILDBEARING MOTIVATION QUESTIONNAIRE
Please rate how desirable are the following
0. Not desirable
1. Somewhat desirable
2. Very desirable
_____a.
How desirable is...Devoting yourself and much of your
time to raising children and being a mother (father)?
_____b.
How desirable is...Having children that work for the
family?
_____c.
How desirable is...Having my child provide me with
companionship and support later in life.
_____d.
How desirable is...Living a fuller, more enriched
life through your child?
_____e.
How desirable is...Having a helpless baby to love and
protect?
_____f.
How desirable is...Sharing childraising with your
spouse?
_____g.
How desirable is...Guiding and teaching your child?
_____h.
How desirable is...Experiencing the special love and
closeness that a child gives?
_____i.
How desirable is...Being the center of a large,
active family?
_____j. How desirable is...Having all the children
that God sends?
_____k.
How desirable is...Feeling more complete as a woman
(man) through my baby?
_____l.
How desirable is...Playing with your child?
_____m.
How desirable is...Giving birth to a baby (Helping
your wife give birth to a baby)?
264
_____n.
How desirable is...Knowing that you are fertile?
_____o.
How desirable is...Having a child who will carry on
your family traditions?
_____p.
How desirable is...Feeling a baby move and kick
inside your body (your wife’s body)?
_____q.
How desirable is...Giving your spouse the
satisfaction of fatherhood (motherhood)?
_____r.
How desirable is...Breast feeding a baby (bottle
feeding a baby)?
_____s.
How desirable is...Providing your parents with a
grandchild?
_____t.
How desirable is...Strengthening your marriage
through a child?
_____u.
How desirable is...Having a son?
_____v.
How desirable is...having my family and friends
admire me with my baby?
_____w.
How desirable is...Holding and cuddling a baby?
_____x.
How desirable is...Having my child contribute to
society?
_____y.
How desirable is...Having a daughter?
_____z.
How desirable is...Fulfilling your religious feelings
about family life?
_____aa. How desirable is...Fulfilling your potential by
having children?
_____bb. How desirable is...Feeling needed and useful through
your baby?
265
Please rate how undesirable are the following.
0.
1.
2.
Not undesirable
Somewhat undesirable
Very undesirable
_____a.
How undesirable is...Having to put up with the mess and
noise that children make?
_____b.
How undesirable is...Feeling guilty or inadequate as a
parent?
_____c.
How undesirable is...Burdening our family finances with
a child?
_____d.
How undesirable is...Worrying whether I am raising my
child the right way?
_____e.
How undesirable is...Having a baby who takes away my
freedom to do other things?
_____f.
How undesirable is...Having a child who is a burden to
my spouse?
_____g.
How undesirable is...Being responsible for a needy and
demanding baby?
_____h.
How undesirable is...Having an unhappy and poorly
adjusted child?
_____i.
How undesirable is...Having a child who embarrasses or
disgraces the rest of the family?
_____j.
How undesirable is...Taking care of a baby who is
disagreeable and irritating?
_____k.
How undesirable is...Experiencing (seeing my wife
experience) the pains of child birth?
_____l.
How undesirable is...Worrying about the health and
safety of your child?
_____m.
How undesirable is...Being a burden on my children in
my old age?
_____n.
How undesirable is...Having a baby who is born
deformed?
266
_____o.
How undesirable is...Having a baby who takes away from
how much I can give to my other child?
_____p.
How undesirable is...Having a baby who strains my
(wife’s) health?
_____q.
How undesirable is...Being kept from my (having my wife
being kept from her) career or job by a baby?
_____r.
How undesirable is...Straining your marriage with a
baby?
_____s.
How undesirable is...Having a child who makes it
necessary for me (my wife) to have a job?
_____t.
How undesirable is...Experiencing (seeing my wife
experience) the discomforts of pregnancy?
_____u.
How undesirable is...Spending time and energy involved
in childcare?
_____v.
How undesirable is...Taking care of a sick child?
267
Now, we would like to know your opinion of a few subjects we are
interested in.
1. If you could decided exactly the number of children that you
would have in your whole life, how many would that be?
__ __ children
The number that God sends
The number that I have
Don’t know
2. If a couple was beginning their family now, how many children should
they have?
__ __ children
The number that God sends
Don’t know
3.
With how many children do you think that a family is:
Large: __________or more
Small: less than __________
4.
Now, I’m going to ask about obligations and activities of
men and women in a couple or in a home. Who should decide
the number of children that a couple should have? The man,
the woman, both, or some other person? DON’T READ
1.
2.
3.
4.
5.
6.
7.
8.
9.
Husband/ man
Wife/ woman
Both together
Someone else
A doctor
Whatever happens / however many God sends
OTHER: ____________________
Specify
DON’T KNOW
5. ¿Who should decide if a couple uses family planning methods?
DON’T READ
a)
b)
c)
d)
e)
f)
g)
h)
Husband/ man
Wife/ woman
Both together
One or the other
A doctor
OTHER ____________________
SPECIFY
DON’T KNOW
268
6. ¿Who should support the economic burden of the family?
DON’T READ
a) Husband/ man
b) Wife/ woman
c) Both together---Æ IF BOTH, READ:
a) Both equally?
Or...
b) The majority the man, or..
c) The majority the woman?
d) One of the two
e) Other ____________________SPECIFY
f) DON’T KNOW
7. Who should take care of the family’s house?
DON’T READ
a) Husband/ man
b) Wife/ woman
c) Both together---Æ IF BOTH, READ:
a) Both equally?
Or...
b) The majority the man, or..
c) The majority the woman?
d) One of the two
e) Other ____________________SPECIFY
f) DON’T KNOW
8. Who should care for the children?
DON’T READ
a) Husband/ man
b) Wife/ woman
c) Both together---Æ IF BOTH, READ:
a) Both equally?
Or...
b) The majority the man, or..
c) The majority the woman?
d) One of the two
e) Other ____________________SPECIFY
f) DON’T KNOW
9. Do you believe that it is acceptable for married men to have
extra-marital sexual relationships?
a) Never
b) Sometimes
c) Always
d) Don’t Know
10. Do you believe that it is acceptable for married women to
have extra-marital sexual relationships?
269
a) Never
b) Sometimes
c) Always
d) Don’t Know
11. Now, I’m going to mention a series of types of work. Who do
you believe should do these types of work? Men, Women or
both?
a
b
c
d
e
f
g
h
I
j
k
l
m
n
o
p
q
r
s
t
u
v
w
x
y
z
aa
bb
cc
dd
ee
Day laborer
Teacher
Barber
Clothes washer
Accountant
Cattle Rancher
Walking Sales
Driver
Farmer
Shop keeper
Gardener
Bricklayer
Worker
Secretary
Candy sales
Police
Journalist
Taylor
Ice cream sales
Pastor
Meat packer
Taxi driver
Security guard
Doctor
Mechanic
Lawyer
Merchant
Factory worker
Domestic worker
Dentist
Nurse
Women
Men
Both
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
______
270
Now, I’m going to ask you a few questions about family planning,
meaning the things that couples use to avoid a pregnancy.
a) Birth control pills
b) IUD
c) condom
d) injection
e) vaginal methods
(jellies, foams,
creams, ovules,
tablets).
f) Female sterilization
or operation
g) Vasectomy or male
sterilization
h) Rhythm or calendar
i) Billings (cervical
mucus)
j)Withdraw
k)Norplant
10. Have you
heard of:
Yes No
11. Have you
ever used:
Yes No
a)
b)
c)
d)
e)
1
1
1
1
1
2
2
2
2
2
a)
b)
c)
d)
e)
1
1
1
1
1
2
2
2
2
2
f)
1
2
f)
1
2
g)
1
2
g)
1
2
h)
i)
1
1
2
2
h)
i)
1
1
2
2
j)
k)
1
1
2
2
j)
k)
1
1
2
2
Demographics
1. Where were you born?
It’s a :
________________________
a. Neighborhood of Catacamas
b. Village of Catacamas
c. Other ________________________
2. What type of water do you have in your house?
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Tap inside
Tap outside but close
Tap off of the property under 100 meters away
Tap off of the property over 100 meters away
Natural source
Well without pump
Well with pump
Purchased
Protected source of water
Other______________________________
3.¿What type of sanitary services do you have?
271
1.
2.
3.
4.
5.
Indoor
Latrine with water
Latrine without water
Open field
Other ___________________________
4. In your house there is...
a. Electricity?
b. Radio?
c. Television?
d. Refrigerator?
e. Telephone?
f. Vehicle
5. ¿With what do you cook?
1. Electricity
2. Gas butane
3. Charcoal
4. Kerosene
5. Wood
6. Don’t cook
7. Other ________________________________
6. ¿How old are you?
___ ___
7a. ¿What was the highest year or grade you finished in school?
0. None
1. Primary
2. Secondary
3. University
4. Reading and Writing
0
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
7b. ¿What was the highest year or grade your SPOUSE finished in school?
0. None
1. Primary
2. Secondary
3. University
4. Reading and Writing
0
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
7c. ¿What was the highest year or grade your MOTHER finished in school?
0. None
1. Primary
2. Secondary
3. University
4. Reading and Writing
0
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
272
8. ¿What is your marital status?
1.
2.
3.
4.
5.
6.
Married
Life partners
Divorced
Separated
Widow
Single
9a. What is your religion?
1.
2.
3.
4.
Catholic
Evangelical
None
Other ________________
9b. ¿Do you consider yourself:
1.
2.
3.
4.
(READ THE OPTIONS)
Very religious
Religious
Somewhat religious
Not religious
10. ¿Have you had a child who was born alive?
3. Yes
4. No
11. ¿How many of your children live with you? ____________
12. ¿How many live somewhere else? ____________
13. ¿Have you had a child born alive which died later?
1.
2.
Yes
No
14. ¿How many? ____________
15. Do you have some type of work or activity for which you receive pay in money or some other
form?
1.
2.
Yes _____________________
No
15b. How much do you make in one month?
1. Less than 3,000 Lempiras
273
2. 3,000–9,999 Lempiras
Don’t know
Refused
16. Does your spouse have some type of work or activity for which she/he receives pay in money
or some other form?
1. Yes (specify activity: _____________________)
2. No
16b. How much, more or less, does she/he make in one month?
1. Less than 3,000 Lempiras
2. 3,000–9,999 Lempiras
Don’t know
Refused
17. (IF THEY BOTH WORK) How much, more or less, do you both make in one month
1. Less than 3,000 Lempiras
2. 3,000–9,999 Lempiras
Don’t know
Refused
18. ¿If we have more questions later, can we contact you to speak more?
1.
2.
Yes
No
19. What is your name?
_________________________________________________________________
20. Where can we contact you?
ADDRESS WITH POINT OF REFERENCE:
_________________________________________________________________
APPENDIX F
PHASE ONE SAMPLE SELECTION
The following chart details the target number of completed questionnaires for the 200
phase one interviews. The 200 interviews are divided equally among four life experience
variables: sex, location, economic level (high, middle, low) and age (< 35 years old, 35
years old and older).
Men
100
Urban
Rural
50
50
Economic Level Low Middle High Low Middle High
16-17 16-17 16-17 16-17 16-17 16-17
Age
18-34
7-9
7-9
7-9
7-9
7-9
7-9
35-50
7-9
7-9
7-9
7-9
7-9
7-9
Women
100
Urban
Rural
50
50
Economic Level Low Middle High Low Middle High
16-17 16-17 16-17 16-17 16-17 16-17
Age
18-34
7-9
7-9
7-9
7-9
7-9
7-9
35-50
7-9
7-9
7-9
7-9
7-9
7-9
274
275
The chart below details the actual distribution of completed interviews for the phase one
interviews.
Men
100
Urban
Rural
50
50
Economic Level Low Middle High Low Middle High
16-17 16-17 16-17 16-17 16-17 16-17
Age
18-34
6
9
8
8
9
7
35-50
7
12
8
9
9
8
Women
100
Urban
Rural
50
50
Economic Level Low Middle High Low Middle High
16-17 16-17 16-17 16-17 16-17 16-17
Age
18-34
9
8
7
9
11
6
35-50
9
8
9
8
9
7
LIST OF REFERENCES
Abu-Lughod, L. (1999). Commentary on an article by Christoph Brumann. Current
Anthropology, 40(Supplement), S13-S15.
Alcock, J. (1999). The myth of genetic determinism–again. Behavioral and Brain
Sciences, 22, 885-886.
Aunger, R. (1999). Against idealism/contra consensus. Current Anthropology, 40(Special
Supplement), S93-S101.
Bachrach, C. A. (1999). Threads in the tapestry of life: Understanding Fertility in
Context. Advances in Population: Psychosocial Perspectives, 3, 1-29.
Bankole, A., & Singh, S. (1998). Couples' fertility and contraceptive decision-making in
developing countries: hearing the man's voice. International Family Planning
Perspectives, 24(1), 15-23.
Barash, D. P. (2002, May 24). Evolution, males and violence. The Chronicle of Higher
Education, pp. B7-B10.
Barkow, J. H. (1989). Darwin, Sex and Status: Biological Approaches to Mind and
Culture. Toronto: University of Toronto Press.
Barkow, J. H., Cosmides, L., & Tooby, J. (Eds.). (1992). The Adapted Mind. Oxford:
Oxford University Press.
Bartholomew, D. J., Steele, F., Moustaki, I., & Galbraith, J. I. (2002). The Analysis and
Interpretation of Multivariate Data for Social Scientists. Boca Raton, FL: Chapman
& Hall/CRC.
Basu, A. M., & Aaby, P. (Eds.). (1998). The Methods and Uses of Anthropological
Demography. Oxford: Clarendon Press.
Basu, A. M. A., Peter. (1998). Introduction: Approaches to anthropological demography.
In A. M. Basu & P. Aaby (Eds.), The Methods and Uses of Anthropological
Demography (pp. 1-21). Oxford: Clarendon Press.
Bernard, H. R. (1994a). Methods belong to all of us. In R. Borofsky (Ed.), Assessing
Cultural Anthropology (pp. 168-179). New York: McGraw-Hill Inc.
276
277
Bernard, H. R. (1994b). Research Methods in Anthropology: Qualitative and
Quantitative Approaches (2nd ed.). Walnut Creek: Alta Mira Press.
Bernard, H. R. (1996). Qualitative data, quantitative analysis. Cultural Anthropology
Methods, 8(1), 9-11.
Bernard, H. R. (2002). Research Methods in Anthropology: Qualitative and Quantitative
Approaches (3rd ed.). Walnut Creek: Alta Mira Press.
Bernard, H. R., & Killworth, P. D. (1979). Why are there no social physics? Journal of
the Steward Anthropological Society, 11(1), 33-58.
Bernard, H. R., & Pelto, P. J. (1987). Technology and anthropological theory. In H. R.
Bernard & P. J. Pelto (Eds.), Technology and Social Change (Second ed., pp. 359376). Prospect Heights, Illinois: Waveland Pres.
Bernard, H. R., & Ryan, G. W. (1998). Text analysis: qualitative and quantitative
methods. In H. R. Bernard (Ed.), Handbook of Methods in Cultural Anthropology
(pp. 595-646). Walnut Creek: AltaMira Press.
Bird, R. (1999). Cooperation and conflict: The behavioral ecology of the sexual division
of labor. Evolutionary Anthropology, 8(2), 65-75.
Bledsoe, C. (1994). 'Children are like young bamboo trees': Potentiality and reproduction
in sub-Saharan Africa. In K. Lindahl-Kiessling & H. Landberg (Eds.), Population,
Economic Development, and the Environment (pp. 105-138). Oxford: Oxford
University Press.
Bock, J. (1999). Evolutionary approaches to population: implications for research and
policy. Population and Environment, 21(2), 193-222.
Bongaarts, J. (1975). Why high birth rates are so low. Population and Development
Review, 1, 289-296.
Bongaarts, J. (1978). A framework for analyzing the proximate determinants of fertility.
Population and Development Review, 4(1), 105-132.
Bongaarts, J., & Watkins, S. C. (1996). Social interactions and contemporary fertility
transitions. Population and Development Review, 22(4), 639-682.
Borgatti, S. P. (1992). AnthroPac (Version 4.0). Columbia, South Carolina: Analytic
Technologies.
Borofsky, R. (1994). Assessing Cultural Anthropology. New York: McGraw-Hill.
Boster, J. S. (1985). Requiem for the omniscient informant: there's life in the old girl yet.
In J. Dougherty (Ed.), Directions in Cognitive Anthropology (pp. 177-197). Urbana:
University of Illinois Press.
278
Boyd, R., & Richerson, P. J. (1985). Culture and the Evolutionary Process. Chicago: The
University of Chicago Press.
Bradley, C. (1997). Why fertility is going down in Maragoli. In T. S. Weisner & C.
Bradley & P. L. Kilbride (Eds.), African Families and the Crisis of Social Change
(pp. 227-252). Westport, Connecticut: Bergin & Garvey.
Brislin, R. W. (1986). The wording and translation of research instruments. In W. J.
Lonner & J. W. Berry (Eds.), Field Methods in Cross-Cultural Research (pp. 137164). Thousand Oaks, CA: Sage Publications.
Brown, D. E. (1991). Human Universals. New York: McGraw-Hill, Inc.
Browner, C. H. (1986). The politics of reproduction in a Mexican village. Signs: Journal
of Women in Culture and Society, 11(4), 710-724.
Brumann, C. (1999). Writing for culture: Why a successful concept should not be
discarded. Current Anthropology, 40(Special Supplement), S1-S27.
Bulatao, R. A. (2001). Introduction. Population and Development Review, 27, 1-14.
Caldwell, J. C. (1982). Theory of Fertility Decline. New York: Academic Press.
Canton, H. (1986). Sound and shoddy sociobiology. Behavioral and Brain Sciences, 9,
188-189.
Carmines, E. G., & Zeller, R. A. (1979). Reliability and Validity Assessment (Vol. 17).
Newbury Park: Sage Publications.
Chackiel, J., & Schkolnik, S. (1996). Latin America: Overview of the fertility transition,
1950-1990. In J. M. Guzman & S. Singh & G. Rodrigues & E. A. Pantelides (Eds.),
The Fertility Transition in Latin America (pp. 3-26). Oxford: Clarendon Press.
Charlton, B. G. (1999). No short cuts to science. Behavioral and Brain Sciences, 22, 889.
Chase, P. G. (1999). Symbolism as reference and symbolism as culture. In R. Dunbar &
C. Knight & C. Power (Eds.), The Evolution of Culture (pp. 34-49). New
Brunswick, New Jersey: Rutgers University Press.
Cleland, J., & Wilson, C. (1987). Demand theories of the fertility transition: An
iconoclastic view. Population Studies, 41(1), 5-30.
Coale, A. J., & Watkins, S. C. (Eds.). (1986). The Decline of Fertility in Europe.
Princeton, New Jersey: Princeton University Press.
Cohen, M. N. (1977). The Food Crisis in Prehistory: Overpopulation and the Origins of
Agriculture. New Haven: Yale University Press.
279
Colapinto, J. (2000). As Nature Made Him: The Boy Who Was Raised as a Girl. New
York: Harper Collins.
Coney, N. S., & Mackey, W. C. (1998). Cultural evolution and gender roles:
Advantage...patriarchy. Mankind Quarterly, 39(1), 45-69.
Cronk, L., Chagnon, N., & Irons, W. (Eds.). (2000). Adaptation and Human Behavior:
An Anthropological Perspective. New York: Aldine de Gruyter.
Daly, M., & Wilson, M. (1986). A theoretical challenge to a caricature of Darwinism.
Behavioral and Brain Sciences, 9, 189-190.
D'Andrade, R. (1995a). The Development of Cognitive Anthropology. Cambridge:
Cambridge University Press.
D'Andrade, R. (1995b). Moral models in anthropology. Current Anthropology, 36(3),
399-408.
D'Andrade, R. (1999). Culture is not everything. In E. L. Cerroni-Long (Ed.),
Anthropological Theory in North America (pp. 85-103). Westport, Connecticut:
Bergin & Garvey.
Davis, K. (1963). The theory of change and response in modern demographic history.
Population Index, 29, 345-366.
Dawkins, R. (1976). The Selfish Gene. Oxford: Oxford University Press.
de Wall, F. (2001). The Ape and the Sushi Master. New York: Basic Books.
DeGraf, D. S., & de Silva, V. (1991). Unmet need for contraception in Sri Lanka.
International Family Planning Perspectives, 17(4), 123-130.
DeWalt, K. M., DeWalt, B. R., & Wayland, C. (1998). Participant Observation. In H. R.
Bernard (Ed.), Handbook of Methods in Cultural Anthropology (pp. 259-299).
Walnut Creek: AltaMira Press.
DiMaggio, P. (1997). Culture and cognition. Annual Review of Sociology, 23, 263-287.
Dodoo, F. N.-A., & Seal, A. (1994). Explaining spousal differences in reproductive
preferences: A gender inequality approach. Population and Environment, 15(5),
379-394.
Dressler, W. W., & Bindon, J. R. (2000). The health consequences of cultural
consonance: Cultural dimensions of lifestyle, social support, and arterial blood
pressure in an African American community. American Anthropologist, 102(2),
244-260.
280
Dunbar, R., Knight, C., & Power, C. (Eds.). (1999). The Evolution of Culture: An
Interdisciplinary View. New Brunswick, New Jersey: Rutgers University Press.
Dunteman, G. H. (1989). Principal Components Analysis (Vol. 69). Newbury Park, CA:
Sage Publications.
Durham, W. H. (1991). Coevolution: Genes, Culture, and Human Diversity. Stanford:
Stanford University Press.
Dyson, T., & Moore, M. (1983). On kinship structure, female autonomy, and
demographic behavior in India. Population and Development Review, 9(1), 35-60.
Easterlin, R. A., & Crimmins, E. M. (1985). The Fertility Transition. Chicago: University
of Chicago Press.
Ellison, P. T. (2001a). Introduction: Ecology, reproduction, and human evolution. In P.
T. Ellison (Ed.), Reproductive Ecology and Human Evolution (pp. 1-11). New
York: Aldine de Gruyter.
Ellison, P. T. (Ed.). (2001b). Reproductive Ecology and Human Evolution. New York:
Aldine de Gruyter.
Ezeh, A. C. (1993). The influence of spouses over each other's contraceptive attitudes in
Ghana. Studies in Family Planning, 24(3), 163-174.
Fisher, H. (1992). The Anatomy of Love: A Natural History of Mating, Marriage, and
Why We Stray. New York: Fawcett Columbine.
Foster, C. (2000). The limits to low fertility: A biosocial approach. Population and
Development Review, 26(2), 209-234.
Fricke, T. E. (1990). Darwinian transitions? A comment. Population and Development
Review, 16(1), 107-119.
Fricke, T. E. (1997a). Culture theory and demographic process: toward a thicker
demography. In D. I. Kertzer & T. E. Fricke (Eds.), Anthropological Demography:
Toward a New Synthesis (pp. 248-277). Chicago: University of Chicago Press.
Fricke, T. E. (1997b). The uses of culture in demographic research: a continuing place for
community studies. Population and Development Review, 23(4), 825-832.
Gillian, R., T, G., & Grazyna, J. (1993). The fertility of agricultural and non-agricultural
traditional societies. Population Studies, 47, 269-281.
Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology,
78(3), 1360-1380.
281
Greenhalgh, S. (1988). Fertility as mobility: Sinic transitions. Population and
Development Review, 14(4), 629-674.
Greenhalgh, S. (1990). Toward a political economy of fertility: anthropological
contributions. Population and Development Review, 16(1), 85-106.
Greenhalgh, S. (1995a). Anthropology theorizes reproduction: Integrating practice,
political economic, and feminist perspectives. In S. Greenhalgh (Ed.), Situating
Fertility: Anthropology and demographic inquiry. (pp. 3-28). Cambridge:
Cambridge University Press.
Greenhalgh, S. (Ed.). (1995b). Situating Fertility: Anthropology and Demographic
Inquiry. Cambridge: Cambridge University Press.
Greenhalgh, S. (1996). The Social construction of population science: An intellectual,
institutional, and political history of Twentieth-Century demography. Comparative
Studies in Society and History, 38(1), 26-66.
Greenhalgh, S. (1997). Methods and meanings: reflections on disciplinary difference.
Population and Development Review, 23(4), 819-824.
Guzman, J. M. (1996). Introduction: social change and fertility decline in Latin America.
In J. M. Guzman & S. Singh & G. Rodrigues & E. A. Pantelides (Eds.), The
Fertility Transition in Latin America (pp. xxii-xxxi). Oxford: Clarendon Press.
Guzman, J. M. (1997). A rapid look at the fertility transition by zone of residence in
Central America and Panama. In A. R. Pebley & L. Rosero-Bixby (Eds.),
Demographic Diversity and Change in the Central American Isthmus (pp. 41-58).
Santa Monica: RAND.
Haaga, J. G. (2001). Comment: the pace of fertility decline and the utility of
evolutionary approaches. Population and Development Review, 27(Supplement),
53-59.
Hammel, E. A. (1990). A theory of culture for demography. Population and Development
Review, 16(3), 455-485.
Hammel, E. A. (1995). Economics 1, culture 0: Fertility change and differences in the
northwest Balkans, 1700-1900. In S. Greenhalgh (Ed.), Situating Fertility:
Anthropology and demographic inquiry (pp. 225-258). Cambridge: Cambridge
University Press.
Hammel, E. A., & Friou, D. S. (1997). Anthropology and demography: marriage, liaison,
or encounter? In D. I. Kertzer & T. E. Fricke (Eds.), Anthropological Demography:
Toward a New Synthesis (pp. 175-200). Chicago: University of Chicago Press.
282
Handwerker, W. P. (1986a). Culture and reproduction: exploring micro / macro linkages.
In W. P. Handwerker (Ed.), Culture and Reproduction: An Anthropological
Critique of Demographic Transition Theory (pp. 1-28). Boulder, Colorado:
Westview Press.
Handwerker, W. P. (1986b). The modern demographic transition: An analysis of
subsistence choices and reproductive consequences. American Anthropologist, 88,
400-417.
Handwerker, W. P. (1989a). The origin and evolution of culture. American
Anthropologist, 91, 313-326.
Handwerker, W. P. (1989b). Women's Power and Social Revolution. Newbury Park, CA:
Sage.
Handwerker, W. P. (1991). Women's power and fertility transition: the cases of Africa
and the Caribbean. Population and Environment, 13, 55-78.
Handwerker, W. P. (1992). West Indian gender relations, family planning programs and
fertility decline. Social Science and Medicine, 35(10), 1245-1257.
Handwerker, W. P. (1993). Empowerment and fertility transition on Antigua, WI:
education, employment, and the moral economy of childbearing. Human
Organization, 52(1), 41-52.
Handwerker, W. P. (1998, November). The genesis of cultural difference: an hypothesis
about stress, learning, cultural evolution, social change, and rotten men &
promiscuous women. Paper presented at the Annual Meeting of the American
Anthropological Association, Philadelphia.
Handwerker, W. P. (1998). Why violence? A test of hypotheses representing three
discourses on the roots of domestic violence. Human Organization, 57(2), 200-208.
Handwerker, W. P. (2001a). Child abuse and the balance of power in parental
relationships: An evolved domain-independent mental mechanism that accounts for
behavioral variation. American Journal of Human Biology, 13(5), 679-689.
Handwerker, W. P. (2001b). Quick Ethnography: A Guide to Rapid Multi-Method
Research. Walnut Creek: AltaMira Press.
Handwerker, W. P. (2002). The construct validity of cultures: Cultural diversity, culture
theory, and a method for ethnography. American Anthropologist, 104(1), 106-122.
Handwerker, W. P. (Ed.). (1986c). Culture and Reproduction: An Anthropological
Critique of Demographic Transition Theory. Boulder, CA: Westview Press.
283
Handwerker, W. P., & Borgatti, S. P. (1998). Reasoning with numbers. In H. R. Bernard
(Ed.), Handbook of Methods in Cultural Anthropology (pp. 549-593). Walnut
Creek: Alta Mira Press.
Handwerker, W. P., & Wozniak, D. F. (1997). Sampling guidelines for cultural data.
Current Anthropology, 38, 869 - 875.
Harris, M. (1989). Our Kind: Who We Are, Where We Came From, Where We Are
Going. New York: Harper & Row.
Harris, M. (2001a). Cultural Materialism: The Struggle for a Science of Culture
(Updated ed.). Walnut Creek: AltaMira Press.
Harris, M. (2001b). The Rise of Anthropological Theory: A History of Theories of
Culture (Updated ed.). Walnut Creek, California: AltaMira Press.
Harris, M., & Ross, E. B. (1987). Death, Sex, and Fertility: Population Regulation in
Preindustrial and Developing Societies. New York: Columbia University Press.
Hill, A. G. (1997). "Truth lies in the eye of the beholder": the nature of evidence in
demography and anthropology. In D. I. Kertzer & T. Fricke (Eds.), Anthropological
Demography: Toward a New Synthesis (pp. 223-247). Chicago: University of
Chicago Press.
Hill, K., & Hurtado, A. M. (1996). Ache Life History. New York: Aldine de Gruyter.
Hill, K., & Kaplan, H. (1999). Life history traits in humans: theory and empirical studies.
Annual Review of Anthropology, 28, 397-430.
Hirshman, C. (1994). Why fertility changes. Annual Reviews of Sociology, 20, 203-233.
Holcomb III, H. R. (1998). Testing evolutionary hypotheses. In C. Crawford & D. L.
Krebs (Eds.), Handbook of Evolutionary Psychology (pp. 303-334). Mahwah, New
Jersey: Lawrence Erlbaum Associates, Publishers.
Hubacher, D., Suazo, M., & Terrell, S. (1997). Examining the increasing popularity of
traditional contraceptive methods in honduras. In A. R. Pebley & L. Rosero-Bixby
(Eds.), Demographic Diversity and Change in the Central American Isthmus (pp.
533-556). Santa Monica: RAND.
Irons, W., & Cronk, L. (2000). Two decades of a new paradigm. In L. Cronk & N.
Chagnon & W. Irons (Eds.), Adaptation and Human Behavior: An Anthropological
Perspective (pp. 3-26). New York: Aldine de Gruyter.
Iverson, S. (1992). Evolutionary Demographic Transition Theory: Comparative Causes
of Prehistoric, Historic and Modern Demographic Transitions. Unpublished
doctoral dissertation, University of Florida, Gainesville, Florida.
284
Johnson, J. C. (1998). Research design and research strategies. In H. R. Bernard (Ed.),
Handbook of Research Methods in Cultural Anthropology (pp. 131-171). Walnut
Creek: Alta Mira Press.
Kaplan, H. (1994). Evolutionary and wealth flows theories of fertility: Empirical tests
and new models. Population and Development Review, 20(4), 753-791.
Kaplan, H. (1996). A theory of fertility and parental investment in traditional and modern
human societies. Yearbook of Physical Anthropology, 39, 1-45.
Kaplan, H., & Lancaster, J. B. (1999). The evolutionary economics and psychology of the
demographic transition to low fertility. In L. Cronk & W. Irons & N. Chagnon
(Eds.), Adaptation and Human Behavior: An Anthropological Perspective (pp.
283-322). New York: Aldine de Gruyter.
Kaplan, H., Lancaster, J. B., Johnson, S. E., & Bock, J. A. (1995). Does observed fertility
maximize fitness among New Mexican men? A test of an optimality model and a
new theory of parental investment in the embodied capital of offspring. Human
Nature, 6(4), 325-360.
Kennedy, D. P. (2002). Cognitive models of fertility decline in Oaxaca City, Mexico.
Population and Environment, Under Review.
Kennelly, I., Merz, S. N., & Lorber, J. (2001). What is gender? American Sociological
Review, 66, 598-605.
Kertzer, D. I. (1995). Political-economic and cultural explanations of fertility decline. In
S. Greenhalgh (Ed.), Situating Fertility: Anthropology and demographic inquiry
(pp. 29-52). Cambridge: Cambridge University Press.
Kertzer, D. I. (1997). Qualitative and quantitative approaches to historical demography.
Population and Development Review, 23(4), 839-846.
Kertzer, D. I., & Fricke, T. (Eds.). (1997). Anthropological Demography: Toward a New
Synthesis. Chicago: University of Chicago Press.
Kertzer, D. I., & Fricke, T. E. (1997). Toward an anthropological demography. In D. I.
Kertzer & T. E. Fricke (Eds.), Anthropological Demography: Towards a New
Synthesis (pp. 1-35). Chicago: University of Chicago Press.
Knight, C., Dunbar, R., & Power, C. (1999). An evolutionary approach to culture. In R.
Dunbar & C. Knight & C. Power (Eds.), The Evolution of Culture (pp. 1-11). New
Brunswick, New Jersey: Rutgers University Press.
Knodel, J. (1997). A case for nonanthropological qualitative methods for demographers.
Population and Development Review, 23(4), 847-853.
Krippendorff, K. (1980). Content Analysis. Beverly Hills: Sage Publications, Inc.
285
Kuznar, L. (1997). Reclaiming a Scientific Anthropology. Walnut Creek: AltaMira Press.
Latin American Demographic Centre (1997). The demographic situation of Central
America. In A. R. Pebley & L. Rosero-Bixby (Eds.), Demographic Diversity and
Change in the Central American Isthmus (pp. 1-40). Santa Monica: RAND.
Leacock, E. B. (1981). The Myth of Male Dominance: Collected Articles of Women
Cross-Culturally. New York: Monthly Review Press.
Lewis, O. (1950). An anthropological approach to family studies. American Journal of
Sociology, 55(5), 468-475.
Low, B. S. (2000). Why Sex Matters: A Darwinian Look at Human Behavior. Princeton,
New Jersey: Princeton University Press.
MacDonald, K. (1999). An evolutionary perspective on human fertility. Population and
Environment, 21(2), 223-246.
Magnani, R. J., McCann, H. G., Hotchkiss, D. R., & Florence, C. S. (1998). The effects
of monetized food aid on reproductive behavior in rural Honduras. Population
Research and Policy Review, 17(4), 305-328.
Margolis, M. L. (2000). True to Her Nature: Changing Advice to American Women.
Prospect Heights, Illinois: Waveland Press, Inc.
Mason, K. O. (1997). Explaining fertility transitions. Demography, 34(4), 443-454.
Mason, K. O. (2001). Gender and family systems in the fertility transition. Population
and Development Review, 27(Supplement), 160-176.
Mason, K. O., & Taj, A. M. (1987). Differences between women's and men's
reproductive goals in developing countries. Population and Development Review,
13(4), 611-638.
McNicoll, G. (1980). Institutional determinants of fertility change. Population and
Development Review, 6(3), 441-462.
McNicoll, G. (1994). Institutional analysis of fertility. In K. Lindahl-Kiessling & H.
Landberg (Eds.), Population, Economic Development, and the Environment (pp.
199-230). Oxford: Oxford University Press.
McNicoll, G. (2001). Government and fertility in transitional and post-transitional
societies. Population and Development Review, 27(Supplement), 129-159.
Miller, E. M., & Costello, C. Y. (2001). The limits of biological determinism. American
Sociological Review, 66, 592-598.
286
Miller, G. (1956). The magical number seven, plus or minus two: some limits on our
capacity for processing information. Psychological Review, 63(2), 81-97.
Miller, W. B. (1994). Childbearing motivations, desires, and intentions: A theoretical
framework. Genetic, Social, and General Psychology Monographs, 120(2), 223258.
Miller, W. B. (1995). Childbearing motivation and its measurement. Journal of Biosocial
Science, 27, 473-487.
Miller, W. B., Pasta, D. J., MacMurray, J., Chiu, C., Wu, S., & Comings, D. E. (1999).
Genetic influences on childbearing motivation: A theoretical framework and some
empirical evidence. Advances in Population: Psychosocial Perspectives, 3, 53-103.
Miller, W. B., Pezeshki, M. Z., & Kennedy, D. P. (2002). A motivational framework for
childbearing in cross-cultural perspective. American Psychologist, In press.
Montagu, A. (1997). Man's Most Dangerous Myth: The Fallacy of Race (6th ed.).
Walnut Creek: AltaMira Press.
Monteith, R., Stupp, P., Corrales, G., & Caceres, J. M. (1997a). Honduras, Encuesta
Nacional De Epidemiologia Y Salud Familiar - 1996: Informe Final. Tegucigalpa,
Honduras and Atlanta, GA: Ministerio de Salud, Association Hondurena de
Planificacion de la Familia, United States Agency for International Development,
Centers for Disease Control and Prevention.
Monteith, R., Stupp, P., Corrales, G., & Caceres, J. M. (1997b). Honduras, Encuesta
Nacional De Epidemiologia Y Salud Familiar - 1996: Informe Resumido.
Tegucigalpa, Honduras and Atlanta, GA: Ministerio de Salud, Association
Hondurena de Planificacion de la Familia, United States Agency for International
Development, Centers for Disease Control and Prevention.
Monteith, R., Stupp, P., Dickerman, S., Corrales, G., & Caceres, J. M. (1998). Honduras,
Encuesta Nacional De Salud Masculina - 1996: Informe Final. Tegucigalpa,
Honduras and Atlanta, GA: Ministerio de Salud, Association Hondurena de
Planificacion de la Familia, United States Agency for International Development,
Centers for Disease Control and Prevention.
Moore, T. W. (1998). Fertility in China 1982-1990: Gender equality as a complement to
wealth flows theory. Population Research and Policy Review, 17(2), 197-222.
Morgan, L. H. (1985). Ancient Society. Tucson: The University of Arizona Press.
Murphy, M. F., & Margolis, M. L. (1995). An introduction to cultural materialism. In M.
F. Murphy & M. L. Margolis (Eds.), Science, Materialism, and the Study of Culture
(pp. 1-4). Gainesville, FL: University of Florida Press.
287
Neter, J., Kutur, M. H., Nachtsheim, C. J., & Wasserman, W. (1996). Applied Linear
Statistical Models (4th ed.). Boston: McGraw-Hill.
Notestein, F. (1945). Population: The Long View. In T. Schultz (Ed.), Food For the
World (pp. 39-41). Chicago: Chacago University Press.
Obermeyer, C. M. (1997). Qualitative methods: a key to a better understanding of
demographic behavior? Population and Development Review, 23(4), 813-818.
Pérusse, D. (1993). Cultural and reproductive success in industrial societies: Testing the
relationship at the proximate and ultimate levels. Behavioral and Brain Sciences,
16, 267-322.
Phan, S. B. (1997). Unmet need for contraception in Vietnam: who needs what and
where? Social Biology, 44(1/2), 111-123.
Pollak, R. A., & Watkins, S. C. (1993). Cultural and economic approaches to fertility:
proper marriage or mésalliance? Population and Development Review, 19(3), 467496.
Potts, M. (1997). Sex and the birth rate: Human biology, demographic change, and
access to fertility-regulation methods. Population and Development Review, 23(1),
1-39.
Potts, M. (2000, January). The unmet need for family planning. Scientific American, 282,
88-93.
Potts, R. (1998). Variability selection in hominid evolution. Evolutionary Anthropology,
7, 81-96.
Rao, V. (1997). Can economics mediate the relationship between anthropology and
demography? Population and Development Review, 23(4), 833-838.
Razzaque, A. (1999). Preference for children and subsequent fertility in Matlab: does
wife-husband agreement matter? Journal of Biosocial Science, 31, 17-28.
Ridley, M. (1993). The Red Queen: Sex and the Evolution of Human Nature. New York:
Penguin Books.
Riley, N. E. (1997). Similarities and differences: anthropological and demographic
perspectives on gender. In D. I. Kertzer & T. E. Fricke (Eds.), Anthropological
Demography: Toward a New Synthesis (pp. 115-138). Chicago: University of
Chicago Press.
Riley, N. E. (1998). Research on gender in demography: Limitations and constraints.
Population Research and Policy Review, 17(6), 521-538.
288
Risman, B. J. (2001). Calling the bluff of value-free science. American Sociological
Review, 66, 605-611.
Romney, A. K. (1999). Culture consensus as a statistical model. Current Anthropology,
40(Special Supplement), S103-S115.
Romney, A. K., Weller, S. C., & Batchelder, W. H. (1986). Culture as consensus: A
theory of culture and informant accuracy. American Anthropologist, 88, 313-338.
Rose, S. (1998). Lifelines: Biology Beyond Determinism. Oxford: Oxford University
Press.
Rose, S. (1999). Précis of Lifelines: Biology, freedom, determinism. Behavioral and
Brain Sciences, 22, 871-921.
Ross, J., & Heaton, L. (1997). Intended contraceptive use among women without an
unmet need. International Family Planning Perspectives, 23(4), 148-154.
Sahlins, M. (1976). The Use and Abuse of Biology. Ann Arbor: University of Michigan
Press.
Sai, F. T. (1997). The International conference on population and development
programme of action: pious hope or a workable guide? Health Transition Review,
7(Supplement 4), 1-5.
Sangree, W. H. (1997). Pronatalism and the elderly in Tiriki, Kenya. In T. S. Weisner &
C. Bradley & P. L. Kilbride (Eds.), African Families and the Crisis of Social
Change (pp. 184-207). Westport, Connecticut: Bergin & Garvey.
Scheper-Hughes, N. (1992). Death Without Weeping: The Violence of Everyday Life in
Brazil. Berkeley: University of California Press.
Scheper-Hughes, N. (1997). Demography without numbers. In D. I. Kertzer & T. E.
Fricke (Eds.), Anthropological Demography: Toward a New Synthesis (pp. 201222). Chicago: University of Chicago Press.
Schoen, R., Kim, Y. J., Nathanson, C. A., Fields, J., & Astone, N. M. (1997). Why do
Americans want children? Population and Development Review, 23, 333-358.
Scrimshaw, S. (1983). Infanticide as deliberate fertility control. In R. Bulatao & R. Lee
(Eds.), Determinants of Fertility in Developing Countries: Fertility Regulation and
Institutional Influences (Vol. 2, pp. 245-266). New York: Academic Press.
Spradley, J. P. (1979). The Ethnographic Interview. New York: Holt, Reinheart and
Winston.
Strauss, C., & Quinn, N. (1997). A Cognitive Theory of Cultural Meaning (Vol. 9).
Cambridge: Cambridge University Press.
289
Thompson, W. S. (1929). Population. American Journal of Sociology, 34(6), 959-975.
Tomasello, M. (1999). The human adaptation for culture. Annual Review of
Anthropology, 28, 509-529.
Tooby, J., & Cosmides, L. (1992). The psychological foundations of culture. In J. H.
Barkow & L. Cosmides & J. Tooby (Eds.), The Adapted Mind (pp. 19-136).
Oxford: Oxford University Press.
Townsend, N. (1997). Reproduction in anthropology and demography. In D. I. Kertzer &
T. Fricke (Eds.), Anthropological Demography: Toward a New Synthesis. Chicago:
University of Chicago Press.
Turke, P. W. (1989). Evolution and the demand for children. Population and
Development Review, 15, 61-90.
Turke, P. W. (1991). Theory and evidence on wealth flows and old-age security: A
Reply to Fricke. Population and Development Review, 17, 687-702.
Udry, J. R. (1996). Biosocial models of low-fertility societies. Population and
Development Review, 22(Supplement), 325-336.
Udry, J. R. (1999). How to alienate your natural allies and attract your enemies.
Behavioral and Brain Sciences, 22(5), 911-911.
Udry, J. R. (2000). Biological limits of gender construction. American Sociological
Review, 65(June), 443-457.
Udry, J. R. (2001). Feminist critics uncover determinism, positivism, and antiquated
theory. American Sociological Review, 66, 611-618.
Vining, D. R. J. (1986). Social versus reproductive success: The central theoretical
problem of human sociobiology. Behavioral and Brain Sciences, 9, 187-216.
Watkins, S. C. (1986). Conclusions. In A. J. Coale & S. C. Watkins (Eds.), The Decline
of Fertility in Europe (pp. 420-449). Princeton, New Jersey: Princeton University
Press.
Weidman, N. M. (2002, June 28). Who are you calling violent? Want to make something
of it? [Letter to the editor]. The Chronicle of Higher Education, pp. B4.
Weller, S. C. (1998). Structured interviewing and questionnaire construction. In H. R.
Bernard (Ed.), Handbook of Methods in Cultural Anthropology (pp. 363-407).
Walnut Creek, CA: AltaMira.
Weller, S. C., & Romney, A. K. (1988). Systematic Data Collection (Vol. 10). Newbury
Park: Sage Publications.
290
Westoff, C. F. (1978). The unmet need for birth control in five Asian countries. Family
Planning Perspectives, 10(3), 173-181.
Westoff, C. F. (1981). Unwanted fertility in six developing countries. International
Family Planning Perspectives, 7(2), 43-52.
Westoff, C. F. (1988). The potential demand for family planning: A New Measure of
Unmet Need and Estimates For Five Latin American Countries. International
Family Planning Perspectives, 14(2), 45-53.
Westoff, C. F., & Pebley, A. R. (1981). Alternative measures of unmet need for family
planning in developing countries. International Family Planning Perspectives, 7(4),
126-136.
Williams, G. C. (1966). Adaptation and Natural Selection: A Critique of Some Current
Evolutionary Thought. Princeton, N.J.: Princeton University Press.
Wilson, E. O. (1998). Consilience. New York: Alfred A. Knopf.
Yaukey, D. (1985). Demography: The Study of Human Population. Prospect Heights,
Illinois: Waveland Press, Inc.
Zulu, E. (1998). The Role of Men and Women in Decisionmaking About Reproductive
Issues in Malawi (Vol. 2). New York: Population Council.
BIOGRAPHICAL SKETCH
David Kennedy was born in the suburbs of Philadelphia, Pennsylvania, in 1969.
He received a Bachelor of Arts degree from the University of Notre Dame with a double
major in anthropology and computer applications. After receiving his BA in 1992, he
worked in South Bend, Indiana, and Austin, Texas, as a computer programmer and
systems analyst. In 1996, he began graduate training in anthropology at the University of
Florida. While pursuing his graduate degrees, he worked as a field director and
programmer for the University of Florida’s Survey Research Center. He has taken an
interdisciplinary track for his degree in anthropology with a specialization in quantitative
methods. He received his Master of Arts degree in 1999.
David has conducted fieldwork in Latin America for several research projects. He
conducted master’s fieldwork in Oaxaca City, Mexico, during the summer of 1997. He
conducted fieldwork in Honduras on several occasions since 1998. In July 1998, he
worked on a team that conducted a preliminary investigation of epilepsy in Honduras.
From July 2000 until October 2000, he collected data for a study of the nutritional aftereffects of hurricane Mitch on children under the age of 5 years old. From July 2000 until
June 2001, he collected the data and conducted preliminary analyses for his dissertation
in Catacamas, Olancho, Honduras.
David now lives in Durham, North Carolina, with his wife, Marie, and his dogs,
Mavis, Mystere, and Milka. He works as a post-doc fellow at the University of North
Carolina’s Carolina Population Center.
291

Similar documents

×

Report this document