# About the standard nonnormal distribution's 'n' size

The book says, if the ‘n’ size is large enough, we can normally standardize nonnormal distribution.

I had known that the large means ‘n≥30’, but the book have not used the z-distributin.

Why I ask this question is the Kaplan 2015 lv1 note 297p’s example uses not z-distributin but t-distribution.

How much is enough?

The t-distribution is more conservative, but when n ≥ 30, the z-distribution is acceptable.

On the exam, you’ll use whichever distribution they give you.

My pleasure.

Same problem with the size of the sample

in B. the sample size is 40 which is greather than 30, and the t-statistic should be applied here according to the table below, am i right ?

I suspect that you’re missing the point of the question.

You would use a t-statistic if you had data about the earnings themselves. But here _ you don’t _ have data about the earnings themselves, you have data about analysts’ estimates of the earnings. The point is that you cannot create a confidence interval for some data when all you have is a bunch of guesses on those data (even if they’re educated guesses).

it’s all clear now

thank you

Good to hear.

My pleasure.