6

When should I use weights when performing a logistic regression? The weights I'm referring to are sampling weights from a survey? Or should I just use the unweighted data?

StasK
  • 29,235
  • 2
  • 80
  • 165
user22119
  • 599
  • 4
  • 10

1 Answers1

3

The sampling weights are designed to account for the non-simple random sample nature of your sample. Therefore, they are just as needed in one form of regression as another. Exactly how to do this may be complicated; e.g. in SAS there is PROC SURVEYLOGISTIC to deal with various sorts of samples. In R there is the survey package which I think does similar things (but I have not used it).

Peter Flom
  • 94,055
  • 35
  • 143
  • 276
  • Not for any form of regression. Weights in linear regression are necessary only if we think errors are not independent by $X$: ([pdf](https://ojs.ub.uni-konstanz.de/srm/article/view/47)). – Federico Tedeschi Dec 18 '18 at 11:11
  • This paper: ([pdf](http://unionstats.gsu.edu/9220/Solon_WhatWeightingFor_JHR_2015.pdf)) adds other reasons why one should weight, that are however related to violations of the standard hypotheses: i.e.heteroskedasticity, endogenous sampling and identifying average partial effects. – Federico Tedeschi Dec 18 '18 at 11:20