SPPA 619, Study Guide for Exam 1
SPPA 619 Study Guide
Notes: (1) unless you are very good at things like long division, a calculator will probably help, (2)
be sure to look over the z score problems and any other problems that were handed out in class, (3)
the formulas that you will need to know are given at the bottom of the study guide, (4) bring the z
score tables that were handed out in class, (5) as always, the study guide does not cover
everything that we talked about in class and is not intended as a substitute for studying your notes.
1. What is the difference between descriptive and inferential statistics?
2. List and describe the four types of measurement scales. What are the arithmetic properties of
each scale? Give an example of each.
3. Why is it that an average cannot be calculated for ordinally scaled numbers?
4. What is a frequency distribution? List and briefly define three fundamental numerical
characteristics of frequency distributions.
5. List and briefly define three measures of central tendency.
6. What is bimodality? (Give both the technical definition and the one that is in common use.)
7. Why is the mean deviation a poor measure of dispersion? What is the standard deviation and
how does it get around the limitation associated with the mean deviation?
8. In what cases would the median be a better measure of central tendency than the mean?
9. We have worked with two different methods for computing percentiles. One method calculates
the percentile directly, and the second method calculates a z score first, then converts the z
score to a percentile through a table look-up. Assuming that the proper assumptions are met,
why is the z score method superior?
10. Compute the mean, median, variance, and standard deviation for the following scores:
Answers: mean = 68.909, variance = 258.277, stdev = 16.071, median = 69
11. A researcher claims that 80% of the variance in lip-reading ability is accounted for by visualevoked response latency? What does this mean? Does this mean that there is a causal
relationship between these two variables? Assuming that the researcher is correct, what is the
"residual" or unexplained variance? What is the Pearson r for this relationship?
12. Does correlation imply causality? Give an example (one of your own, not one used in class)
that illustrates whatever position to take on this question.
SPPA 619, Study Guide for Exam 1
13. Draw scatter plots and regression lines showing: (a) a strong positive correlation, (b) a weak
positive correlation, (c) a strong negative correlation, and (d) no correlation.
14. A researcher measures the relationship between grade-point average and (1) GRE scores, and
(2) hat size for a group of recent master's degree recipients. (a) For which of these relationships
would the Pearson r be larger? (b) For which of these relationships would the standard error of
estimate be larger? (c) For which of these relationships would the coefficient of determination
15. For the GRE-GPA relationship above, assume that the standard error of estimate for GPA
scores predicted from GRE scores is 0.2 points. Assume also that a student receives a GRE
score of 750, which predicts a GPA of 2.8. (a) What is the probability that the actual GPA score
will be between 2.6 and 3.0? (b) What is the probability that the actual GPA score for this
student will be between 2.4 and 3.2?
16. The facts:
A large-scale test-retest reliability study is conducted for IQ. The study involves testing
subjects twice at 2-week intervals on equivalent forms of the same test.
The standard error of estimate for the test-retest relationship is 3 IQ points.
An individual scores 115 on the IQ test. Being responsible professionals, we are sensitive to
the fact that the score of 115 is not the "true" IQ for this individual, but an estimate of his
Given these facts, you are approximately 95% confident (+-2 s.e.e.) that this individual's
true IQ is between what two IQ scores?
Formulas and Statistics You Should Know
percentile (direct-score method and indirect method based on z scores)
z (standard score)
Pearson r (know to interpret it, not how to calculate it)
coefficient of determination (RSQ)
standard error of estimate (know to interpret it and how it is calculated conceptually, but not a
formula per se)