I am doing some multivariate regression. When I plot QQnorms to check my model assumptions (errors normally distributed), I get enter image description here

Question: How do you interpret the above? Are the normality assumptions completely off? What's with those "breaks" near the ends of some of the plots? Is this caused by outliers?


marked as duplicate by Glen_b Mar 22 '17 at 6:42

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • $\begingroup$ Look at this answer for how to interpret visual aspects of qq-plots. But are you sure that, with the way you got these plots, you were comparing the residual distributions against a normal distribution? Looks like you might be comparing among variables against each other. $\endgroup$ – EdM Mar 22 '17 at 0:41
  • $\begingroup$ Given some people plot QQ-plots the "wrong way around" (to my way of reckoning), you should be clear about which axis is which $\endgroup$ – Glen_b Mar 22 '17 at 3:10
  • $\begingroup$ If the indicated duplicates aren't sufficient to answer your question, please edit your question to ask about the specific issues not covered in the linked answers $\endgroup$ – Glen_b Mar 22 '17 at 6:49

Browse other questions tagged or ask your own question.