Skip to main content
Cross Validated

You are not logged in. Your edit will be placed in a queue until it is peer reviewed.

We welcome edits that make the post easier to understand and more valuable for readers. Because community members review edits, please try to make the post substantially better than how you found it, for example, by fixing grammar or adding additional resources and hyperlinks.

5
  • 2
    $\begingroup$ I disagree that there is no error term. The observed binary response minus the conditional expectation is the error term, and its variance is as I stated in my answer. $\endgroup$ Commented Jan 1, 2021 at 15:30
  • $\begingroup$ Any comments on the part of the question asking about why you would need Robust Standard Errors for logistic regression? $\endgroup$ Commented Apr 20, 2021 at 14:10
  • 1
    $\begingroup$ Logistic regression need not be thought of as having an error term. And robust standard errors sometimes are less precise than ordinary estimates. $\endgroup$ Commented Sep 16, 2023 at 11:12
  • $\begingroup$ Heteroscedasticity is to linear regression what scale effects are to logit models. Like in the linear regression case, it depends on what predictors and functional forms go into the regular formula (mean or location structure). However, testing scale effects in logit models can be fragile due to the "complete separation" problem: Not uncommon to provide a huge coefficient of the scale effect, singular Hessian, all standard errors are NA, or algorithm cannot converge. But they are worth examining. Williams wrote Stata packages for these methods. See my answer. $\endgroup$ Commented Apr 8 at 11:42
  • $\begingroup$ @BigBendRegion that is an artificial construct in this setting and doesn’t help the discussion. $\endgroup$ Commented Apr 8 at 14:57