Which type of tests typically have greater power: planned contrasts or post hoc tests?

Which type of tests typically have greater power: planned contrasts or post hoc tests?

• A. Planned contrasts.
• B. Post hoc tests.
• C. Polychotomous logistic regression.
• D. Posterior distribution.

Answer: (A)

Power is higher when you use planned contrasts because they test specific, theory-driven hypotheses about the group means and allocate the alpha level to only a small, focused set of comparisons. By targeting a few predefined linear combinations of means, you maximize the likelihood of detecting the hypothesized differences, since you’re not spreading your alpha across many possible tests. Post hoc tests, used after seeing the data to explore many possible differences, must protect the overall error rate across numerous comparisons. This requires adjustments (e.g., Tukey, Bonferroni) that make each test more conservative, reducing the chance of finding true differences, especially when there are many groups or the effects are modest. So planned contrasts are typically more powerful for confirming predefined hypotheses. Polychotomous logistic regression and posterior distribution refer to different analytic frameworks and are not the typical basis for comparing power of contrasts in ANOVA.

Power is higher when you use planned contrasts because they test specific, theory-driven hypotheses about the group means and allocate the alpha level to only a small, focused set of comparisons. By targeting a few predefined linear combinations of means, you maximize the likelihood of detecting the hypothesized differences, since you’re not spreading your alpha across many possible tests. Post hoc tests, used after seeing the data to explore many possible differences, must protect the overall error rate across numerous comparisons. This requires adjustments (e.g., Tukey, Bonferroni) that make each test more conservative, reducing the chance of finding true differences, especially when there are many groups or the effects are modest. So planned contrasts are typically more powerful for confirming predefined hypotheses.

Polychotomous logistic regression and posterior distribution refer to different analytic frameworks and are not the typical basis for comparing power of contrasts in ANOVA.