Sun 12 Dec, 2010 09:52 pm
When running an OLS with regression with robust standard errors, do I also need to run a Hausman endogeneity test? My data is at the student level and clustered by college. Thanks.
In case anyone ever stumbles upon this, I've been in touch with my stats teacher and he gave me the answer. He said it's not necessary to run the Hausman test, because by using a ols regression with robust standard errors you're already assuming there's little between variation between groups (check your unconditional ICC first though).