Which term is a measure of the probability of making a Type I error across a family of tests?

Which term is a measure of the probability of making a Type I error across a family of tests?

• A. Fixed coefficient
• B. Fisher's exact test
• C. Familywise error rate
• D. Factorial ANOVA

Answer: (C)

Multiple tests increase the chance of at least one false positive, so the measure of that overall risk across all the tests is the familywise error rate. It’s the probability that at least one true null hypothesis is rejected when you’re looking at a set of hypotheses. If the tests are independent and each uses the same per-test alpha (say 0.05), the FWER is approximately 1 minus (1 minus alpha) raised to the number of tests, showing how quickly the risk grows as you add more comparisons. The other terms don’t describe this overall error probability: a fixed coefficient is just a model parameter, Fisher’s exact test is a specific test for small-sample 2x2 tables, and factorial ANOVA is a design for examining multiple factors and their interactions.

Multiple tests increase the chance of at least one false positive, so the measure of that overall risk across all the tests is the familywise error rate. It’s the probability that at least one true null hypothesis is rejected when you’re looking at a set of hypotheses. If the tests are independent and each uses the same per-test alpha (say 0.05), the FWER is approximately 1 minus (1 minus alpha) raised to the number of tests, showing how quickly the risk grows as you add more comparisons. The other terms don’t describe this overall error probability: a fixed coefficient is just a model parameter, Fisher’s exact test is a specific test for small-sample 2x2 tables, and factorial ANOVA is a design for examining multiple factors and their interactions.