16. If the number of subjects in a program small, : 2132719
16. If the number of subjects in a program is small, statistical tests can detect only very large program effects of differences in outcome measures between the two groups. This illustrates the problem of _____.
b. statistical conclusion validity
17. Treatment integrity is _____.
a. roughly equivalent to reliability
b. roughly equivalent to validity
c. the difference between the experimental intervention is delivered as intended
d. both reliable and valid
18. Randomized experiments for evaluation purposes require which of the following?
a. case flow must be adequate to produce enough subjects in both the experimental and the control groups
b. program staff may minimize the exceptions to randomization
c. special interventions must be consistently applied
d. maintaining the desired level of control over experimental conditions
19. In the home detention studies used in your text as examples of randomized studies, the authors concluded that _____.
a. randomization allows generalization
b. randomization allows for reliability
c. randomization does not control for variation in treatment integrity and program delivery
d. randomization does control for variation in treatment integrity and program delivery
20. What can be said about the relationship between process evaluations and impact assessments?
a. process evaluations can be used in interpreting results from an impact assessment
b. process evaluations monitor program implementation but impact assessments do not answer questions about a program’s effects
c. process evaluation cannot be used to interpret outcomes
d. there is no relationship between impact assessments and process evaluations