We get questions about the normality of residuals from models, e.g., Dependent variable - bimodal? Often, as in this example, it turns out that the OP wants to perform inference on model parameters and believes that residuals need to be normally distributed for inference to be valid.

I would have expected to find a highly voted question in the tag that I could use to VTC as a duplicate, but I couldn't find it.

Where is our canonical answer to questions about normality of residuals and inference?

  • 11
    $\begingroup$ Look through these search results. $\endgroup$
    – whuber Mod
    Sep 23, 2018 at 16:10
  • 11
    $\begingroup$ @whuber: your search-fu leaves me speechless. $\endgroup$ Sep 23, 2018 at 17:23
  • $\begingroup$ Here is another search I have found useful for locating already used targets: stats.stackexchange.com/search?q=assumption*+residual*+normal*+++-logistic+duplicate%3Ayes it is indirect, a direct way for searching for already used targets would be useful! $\endgroup$ Sep 30, 2018 at 9:50

1 Answer 1


whuber pointed to a very nice search. It seems to me like the following thread is the most useful one for our purposes here: Regression when the OLS residuals are not normally distributed.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .