t and z test

[suspense]

Why does the cfa book prefer t-test over z-test when sample size >30

 

[S2000magician]

It’s slightly more conservative, and more accurate. The differences, however, are slight. They will not give you a problem with a large sample size where using one test will lead to the correct answer choice but using the other will lead to a wrong answer choice.

 

[edupristine]

Theoretically, in situations where the distribution is non-normal with sample size more than 30, t-test should be used.
In these situations, z-test can also be used. With increasing sample size, z-test gives very similar results just like with t-test. Technically speaking, you should prefer t-test over z-test.
But when you talk about CFA, you need not worry too much. As already mentioned by s2000magician, both tests would lead to correct answer choice.