Question

If you are comparing two samples (each sample has 20 observations), which test will give you the most conservative outcome?

	A.	the t-test with separate variances
	B.	all of the tests should give you the same outcome since the sample does not change
	C.	the Z test
	D.	the pooled variance t-test

Answer 1

Since there are 2 samples, different samples and

We don't know their population variance, so can't use z test

Their variance may be different.

We use pooled variance t test if we can assume that the two samples have same variance, which is not sure here

This it's better to use t test with different variances

In this way, we are making less assumptions which have chances of being invalid and affecting our test negatively.

Hence the answer is

A