|
|
Part I: Conceptual questions
1) What is the distribution of sample means? What is the standard error? How is the standard error (or estimate of) used in hypotheis tests?
2) How are one sample t and z tests related? How are they different?
3) How are one sample t tests and related sample t-tests related?
4) Why do we pool variances when perfoming an independent samples t-test?
5) What factors affect stastical power (and describe in what way)?
6) How can effect size and statistical power be used in planning a study?
Part II: Sample problems
7) An organizational psychologist predicts that assembly workers will hav a somewhat higher level of job satisfaction if they are given a new kind
of incentive program (that is he predicts a medium effect size). On a standard job satisfaction scale, for
assembly workers in this company overall, the distribution is normal, with a mean of 82 and a standard deviation of 7.
The psychologist plans to provide the new incentive program to 25 randomly selected assembly workers.
a) What is the power of this study (assume that they will use an alpha level = 0.01)?
Answer the follwoing questions using the majors.sav datafile
8) Perform t-tests to compare high school english scores, high school math scores, and high school science scores (do three separate
t-tests: hss vs. hse, hss vs. hsm, hse vs hsm). State your conclusions about each (assume two-tailed tests).
9) Perform t-tests to compare men and women on high school math, science, and english scores.
10) Compare three regression models predicting SATm scores.
| Model 1 | high school math scores |
| Model 2 | high school math & english scores |
| Model 3 | high school math, english, & science scores |
Describe the models, which model is "best" and why?
11) Compare three regression models predicting SATv scores.
| Model 1 | high school english scores |
| Model 2 | high school science & english scores |
| Model 3 | high school science, english, & math scores |
Describe the models, which model is "best" and why?
Explain why the beta weight for hse is significant in Model 1, but not in model 2.
|
