Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am currently reading about the standard error. I know about central limit theorem, but I don't understand why, if my variable is normally distributed in the population, the sampling distribution (the distribution of the sample means) is also a normal distribution, no matter of sample size etc.

share|improve this question
The sampling distribution of the standard error (of the mean) is not normally distributed. Are you perhaps thinking of the sample distribution of the mean itself? Please edit your question to clarify this. – whuber Apr 16 '14 at 15:52
yes, I mean the distribution of the means. I clarified it with an edit – 00schneider Apr 16 '14 at 15:56

This is not an asymptotic result, it is a finite-sample result and has nothing to do with the Central Limit Theorem.

You start by assuming that the population is normal. The normal distribution is one of the few distributions that are closed under addition, meaning that the sum of normal random variables follows a normal distribution (although with different parameters than the components of the sum). More over, scaling a normal random variable, leaves it a normal random variable (again, affecting the parametrization of it).
The sample mean of a sample from a normal population, is a sum of normal random variables, scaled by $1/n$. Hence it too follows a normal distribution.

The Central Limit Theorem comes into effect to lead to the same result asymptotically, when the population is not normally distributed.

share|improve this answer
+1 Because statistical language can be confusing, especially to beginners, I would like to point out that scaling does change the distribution: but it does not change its shape (which remains Normal) and that of course is what you are contending. – whuber Apr 16 '14 at 19:26
So very true. Corrected. – Alecos Papadopoulos Apr 16 '14 at 19:54
+1 Some people phrase it as 'scaling doesn't change the family of distributions', because the normal isn't so much a distribution as a location-scale family of distributions. (Don't take this as a suggested change to the answer, though, which seems fine as it is.) – Glen_b Apr 16 '14 at 22:18
@Glen_b Yes I have noticed that some times what is called "distribution" in one corner, is called "family of distributions" in another. I have the impression that the latter is used more by those that are more deeply into statistics. – Alecos Papadopoulos Apr 16 '14 at 22:35
@Alecos I use them both, so I'm not even consistent, though most times it's clear enough. In a situation like that, where one wants to say "the same <something>" it does seem to reduce the chance of confusion. – Glen_b Apr 16 '14 at 22:37

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.