Suppose you are interested in learning whether academic achievement levels are d
ID: 3135277 • Letter: S
Question
Suppose you are interested in learning whether academic achievement levels are different for girls who attend private versus public schools, and who are in single- versus mixed-sex classrooms. You have standardized test scores for girls in four types of classrooms (private, mixed-sex; private, single-sex; public, mixed-sex; public, same-sex). You know that you can perform several t-tests to compare scores from each type of classroom to every other type of classroom (a total of 6 comparisons), but you are concerned about the possibility of making a Type I error by performing a large set of comparisons on the same variable. What is the familywise error rate in this situation, assuming =.05? Select one:
a. There is a 19% chance that at least one t-test will be statistically significant
b. There is an 30% chance that at least one t-test will be statistically significant
c. There is a 20% chance that at least one t-test will be statistically significant
d. There is a 26% chance that at least one t-test will be statistically significant
Explanation / Answer
Suppose you are interested in learning whether academic achievement levels are different for girls who attend private versus public schools, and who are in single- versus mixed-sex classrooms. You have standardized test scores for girls in four types of classrooms (private, mixed-sex; private, single-sex; public, mixed-sex; public, same-sex). You know that you can perform several t-tests to compare scores from each type of classroom to every other type of classroom (a total of 6 comparisons), but you are concerned about the possibility of making a Type I error by performing a large set of comparisons on the same variable. What is the familywise error rate in this situation, assuming =.05? Select one:
a. There is a 19% chance that at least one t-test will be statistically significant
b. There is an 30% chance that at least one t-test will be statistically significant
c. There is a 20% chance that at least one t-test will be statistically significant
d. There is a 26% chance that at least one t-test will be statistically significant
There are 6 comparisons
Family wise error rate = 1 – (1 – )6= 1 – (1 – .05)6
= 0.2649 or 26%
Related Questions
drjack9650@gmail.com
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.