2

I have two related questions:

1- Does Pearson correlation coefficient of two random variables, implicitly assume that random variables (X,Y) are Normally distributed (according to normal distribution)?

2- If we have this knowledge that two random variables X and Y are distributed according to a distribution other than Normal, e.g., Beta, is it still acceptable to measure the association between these two random variables with Pearson?

Amin Kaveh
  • 65
  • 5
  • [Pearson's or Spearman's correlation with non-normal data](https://stats.stackexchange.com/questions/3730/pearsons-or-spearmans-correlation-with-non-normal-data) – user2974951 May 28 '21 at 10:08

1 Answers1

4

No, it's simply the standardized second mixed moment (covariance).

If you imbue it with distributional assumptions, however, you can derive statistics for hypothesis tests. That's where normality assumptions will often show up.

Firebug
  • 15,262
  • 5
  • 60
  • 127