Answers to Practice Questions for Final:
(annotated with side comments in places)
Part II: Application:
I've also added the research question, as with the handout on Which Test? On final, you will be asked for both research question and test. If you are facing a problem like this and feel clueless, identify the variables and their nature, see if you can figure out how many samples/groups there are, and that (in conjunction with your "which test" handout) should help you narrow the choices down.
Note: You might choose a different wording for the research question--as long as your version matches the study, it will be counted as correct. There are wrong answers, however. For example, for 1, an incorrect research question might be: Does the rate of unprotected sex in students depend on what school they attend? This would miss the point of the study, which is to look at AIDS education. An incorrect question for 5 might be: Was Rob happier before the breakup with Dan, or after the breakup? This "difference" question cannot be addressed by the data described.
1. c, Does AIDS education at school reduce the percentage of time students have unprotected sex?
2. d, Does AIDS education workshop change people's condom use?
3. i, Do students value learning, having fun, or a good grade more or less in their courses, or do they value them all equally?
4. h, Does milk consumption differ based on gender and region of the country someone lives in?
5. e, Is Rob's mood associated with the time since his breakup with Dan?
6. b, Does the anti-depressant drug increase the incidence of nightmares?
Note: Identifying the research question should help you narrow down to t-test (or possible z-test). To know which one, however, you have to look more closely at how the data was collected, and whether population mean and variance are known for the comparison population. Second note: It's possible to get confused about the nature of the variables here. Nightmares are being counted, but that doesn't mean that you should select a chi-square. If each person was simply classified into "has nightmares" or "doesn't have nightmares" then you would just have a frequency count of people in different categories. Instead, nightmares is treated as a quantitative variable, which can range from 0 to something large (if people have lots of nightmares).
7. j, Does the likelihood of calling on male or female students depend on the gender of the professor?
8. f, How good will today's lecture be? [Note: for this scenario, the regression procedure is not a "test" in the same way as the other procedures are tests of hypotheses--instead it's a strategy for prediction]
9. a, Does the age of people on juries differ from the average age for voters in the county? [Note: z-test here because mu and sigma are known for one population, and you have a sample mean for a single sample. You'd get partial credit if you chose b.]
10. g, Do average grades (a measure of course difficulty) vary across the different areas of psychology?