Difference-in-difference regression Fixed effects or OLS or Random effects

[Akios Majoni]

Dear All

May you kindly assist me with this basic question. If you are running a difference-in-difference regression, I have seen papers that run it on OLS model but with year/firms/industry fixed effects. My question is, (1) can you run a difference-in-difference regression on a Fixed effects model or Random effects model. If, yes (2) Is it applicable to use the Hausman test to choose among the three alternative models. I am asking because the output of my OLS model changes when you add interaction terms (i.e the results of the variable of interest changes sign and significance)

Your assistance will be greatly appreciated

Thank you.

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data.

I think you can run a d-i-d with fixed or random effects. The standard tests probably apply. You might look at http://econ.lse.ac.uk/staff/spischke/ec533/did.pdf

[Akios Majoni]

Thank you very much Phil, you have answered my question.