Robust Estimators in Logistic Regression
Hello —
I am new to using JASP and have discovered that the documentation leaves a lot of detail to the imagination. I was wondering if anyone can tell me exactly what the "robust standard errors" are for the logistic regression module. Specifically I would like to know which robust estimates are being made, and, if possible, I would like a specific citation to the algorithm used. Any guidance is appreciated! Thanks!
Comments
Hi mdt,
Point taken, but logistic regression is actually relatively well documented. Granted, details on the robust standard errors appear to be missing. I'll ask the team member who implemented this!
Cheers,
E.J.
Hi mdt,
these standard errors come from a sandwich estimator for the variance-covariance matrix of the coefficients. They are "default" Eicker–Huber–White standard errors for GLM.
The precise implementation in R used for JASP, should you want to look at it, can be found at this link.