Question

Which of the following conditions would require a LARGER sample size? CHOOSE ALL THAT APPLY.

A) A beta level of 0.10 (vs a bet level of 0.20)

B) An a priori alpha of 0.05 (vs an a priori alpha level of 0.01)

C) a new drug will be compared to an existing drug (vs comparing it to a placebo)

D) a test of difference (vs a test of superiority)

E) Use of an atomic scale with high accuracy (vs use of a standard laboratory scale with moderate accuracy)

Answer 1

A) When we decrease type II error then we need to increase sample size.

So here we need to take larger sample.

B) Here we use larger type I error (0.05 insted of 0.01) so we can use smaller sample size to allow some extra error .

C) We don't know more information about the new drug than the standard drug. So we need to used more sample size.

D) Here we need larger sample size, because the standard deviation of the difference is larger than the standard error of test of superiority.

E) Here the error is small and so the standard error is small. So we need small sample size.

So the correct choices are A, C, and D.