EXERCISES FOR CHAPTER 15
Exercises 1–7 and 19 pertain to the nursing home study of Section 220.127.116.11. List three alternative explanations for the claimed improvement of the experimental group in the nursing home experiment of Section 18.104.22.168. Suppose the experiment had been replicated using the first and third floors in the same way as was done for the second and fourth floors. What advantage would accrue?3. An alternative method would be to assign the two treatments at random to individual residents within each separate floor. What are the advantages/disadvantages of within floor randomization?4. An alternative design for the nursing home study would have given no specific talk or other treatment to second floor residents, but still make the same measurements as for the fourth floor residents. However, the control condition actually used was needed to control a confound in the experimental condition given the fourth floor residents that would affect the theoretical interpretation. What was this confound?5. The main data were change scores between measures for each resident before and after the treatment.
|1. ||What is the advantage of this change score?|
|2. ||What is the disadvantage of this change score (see Chapter 13)?|
6. “Locus of control” is one of the few aspects of personality that has shown substantial correlations with performance in a variety of tasks. Locus of control refers to the degree to which people feel that they control external events or that external events control them. How might a preliminary measure of locus of control for each resident have been used in the nursing home study to:
|1. ||Decrease error variability? How much decrease would you expect?|
|2. ||Increase substantive significance?|
7. For a follow-up design for the nursing home study, assuming comparable facilities are available, what two changes would you consider most important?
8. This exercise aims to make concrete some isses in random factor analysis. The following experimental situation is typical of one class of situations in which random factor analysis has sometimes been misapplied.
Subjects' responses may depend on personality characteristics of some person that are confounded with the treatment. Studies of social influence, for example, may use a confederate who has some essential role in the treatment. Similarly, effectiveness of marriage counseling will depend on the personality of the counselor. Again, expression of emotion in animals may depend on how