I'm doing ordinal logit/probit only to analyse the direction of causality (e.g. if some variable makes it more likely to observe a low scale or a high scale). No interpretation is needed beyond this.

I can only report either (i) odds ratios, (ii) coefficients, in my table.

Is there a standard one I should be reporting?

The application is empirical economics

  • 1
    $\begingroup$ To me, odds ratios are much more intuitive. But which you should report probably depends on where you are submitting/presenting this. $\endgroup$ – Peter Flom Oct 1 '12 at 11:01
  • 1
    $\begingroup$ Both are pretty poor summaries of the marginal effect of this variable. Odds ratios are the least bad option of the two though. $\endgroup$ – Zach Mar 18 '13 at 20:51
  • 3
    $\begingroup$ @Zach : Marginal effects are just a linear model estimated on top of a non-linear model. If you want to report the results of a linear model then the best way to get those is to directly estimate a linear model and cut out the middle man. So to me marginal effects are poor summaries of a linear model; poor in the sense of hiding all the problems associated with linear models rather than solving them. $\endgroup$ – Maarten Buis May 15 '13 at 15:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy