9
$\begingroup$

I know this question might have been around before like in this SO thread, but maybe (hopefully) its answer has changed over time.

  • Is there any package in R or an outline how to do panel regressions with a discrete dependent variable?
  • Is there any other open source package that does it and that would help coding something in R?
Improve this question
$\endgroup$
1
  • $\begingroup$ the binary case might be a sufficient first start... $\endgroup$
    – hans0l0
    Apr 29, 2011 at 9:48

4 Answers 4

Reset to default
4
$\begingroup$

For "fixed effects" logit regression as understood in econometrics (also called Chamberlain's conditional logit), "clogit" in the "survival" package does the job.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Any confirmation that that is what this package does? From the examples, mclogit seems do do multinomial logit, not binary logit. I can't see a difference between mclogit and clogit. $\endgroup$
    – RegressForward
    Jul 9, 2016 at 17:19
  • 1
    $\begingroup$ @RegressForward Yes, that seems to be the case. I've edited my answer. Clogit in survival definitely does the job. $\endgroup$
    – Julian Schuessler
    Jul 20, 2016 at 17:32
1
$\begingroup$

What is wrong with using plm or lme4 (another link)? Particularly the glmer function?

Improve this answer
$\endgroup$
1
  • $\begingroup$ plm as far as I know does not have functions for fitting discrete data models. lme4 is a good suggestion, but should be used with care, since it assumes that all the independent variables are exogenous, which is not always the case for panel data in econometrics. $\endgroup$
    – mpiktas
    May 30, 2011 at 7:41
1
$\begingroup$

pglm is now available and for e.g. conditional logit there is a closed form estimator that should be straightforward to implement.

Improve this answer
$\endgroup$
1
0
$\begingroup$

If your dependent variable is only binary, you could try out the package "bife".

Their package description is:

"Estimates fixed effects binary choice models (logit and probit) with potentially many individual fixed effects and computes average partial effects. Incidental parameter bias can be reduced with an asymptotic bias correction proposed by Fernandez-Val (2009)"

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.