As my question says, I am wondering what people mean when they say the data follows a normal distribution, does this mean the population follows a normal distribution?


A sample can’t be normal!

Samples are necessarily finite. A normal distribution can take on infinitely many values, so a sample can’t be normal.

Therefore, when we say that we have points that are distributed normally, we mean that they come from a normal population.

  • $\begingroup$ If we assume that the data is normally distributed meaning the data come from a normal distribution, then do we need to perform the normality check? $\endgroup$ – StoryMay Oct 28 '20 at 14:53
  • 1
    $\begingroup$ Your options are to: 1) proceed with the assumption 2) check if your assumption is a good one. $\endgroup$ – Dave Oct 28 '20 at 15:15
  • $\begingroup$ I am wondering if we need a normality check if we assume the data are normally distributed. $\endgroup$ – StoryMay Oct 28 '20 at 15:17
  • 1
    $\begingroup$ I would like to check if my assumption is a reasonable one. // You're still learning, so please be careful with terminology so you know what's really happening. The data can't be normal; the data could be drawn from a normal population distribution, however. $\endgroup$ – Dave Oct 28 '20 at 15:21

It is hard to guess, where your problem lies, but it seems to me that you have not yet grasped the concept of a sample. The statistical model underlying this concept is as follows:

  1. A numeric result of a random experiment follows a certain probability distribution.
  2. The random experiment is repeated independently $n$ times
  3. The list of results $x_1,\ldots,x_n$ is called a "sample".

The distribution in 1. can be a normal distribution, which is the case to which you refer.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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