0

In linear regression, one of its assumption is the residual should be normally distributed. Why does it in logistic regression, the assumption says that residuals do not need to be normally distributed?

Samuel Vera
  • 1
  • 3
    What’s the outcome in a logistic regression? – Dave Jul 10 '21 at 04:48
  • 2
    No, linear regression (more precisely, Ordinary Least Squares) does *not* assume that residuals are normally distributed. It is nice if they are, but the CLT will ensure that *parameter estimates* are normally distributed for sufficiently large sample sizes (which usually holds), and this is what inference cares about. – Stephan Kolassa Jul 10 '21 at 08:22
  • 2
    Think about the possible values that residuals take on here. – dimitriy Jul 10 '21 at 08:24
  • Read this two post to get a better understanding (tell us if it helps!) https://stats.stackexchange.com/questions/124818/logistic-regression-error-term-and-its-distribution and https://stats.stackexchange.com/questions/295340/what-to-do-with-glm-gamma-when-residuals-are-not-normally-distributed/302413#302413 – kjetil b halvorsen Jul 10 '21 at 23:04

0 Answers0