A study is investigating the effect of different methods of therapy-cognitive behavioral, Gestalt, and psychoanalytic-on self-reports of symptom improvement. Improvement is rated on a 0-10 scale, where 10 represents complete improvement. Researchers are deciding what type of statistical analysis to use, a one-way ANOVA or three t-tests. What should they be most concerned about when making this choice?
A)Type I error
B)The number of subjects in each group
C)Violations of homoscedasticity
D)Method of symptom improvement rating

Question

Quizplus · Accepted Answer

Researchers should be most concerned about a type I error when making this decision. Whatever the researchers have determined the alpha is, it has to be multiplied by the number of tests used for the overall experiment alpha. So if the experiment alpha is .01, it will remain at that value if one ANOVA is used, but will increase to .03 if three t tests are used. The higher the alpha of an experiment, the more likely a type I error is.

A Study Is Investigating the Effect of Different Methods of Therapy-Cognitive