Hausman test in multilevel modeling

[Romuald Landry]

I'm building a two-level multilevel model with variables at student and school level, I assume a level twon endogeneity, i.e. between school and student characteristics. I'd like to confirm or invalidate this hypothesis using the Hausman test. I enter the following command, but it gives me an error I don't understand

 mixed Nrdtlectf  diflcev1ir avoirfaim rvauxdome11ique langue1arlé Nbelvdanslécole nbsalledeclasse agedirecteur || id_ecole: , covariance(unstructured)

 hausman fixed ., sigmamore

[Daniel Schaefer]

I'm confused. In general, the Hausman test compares a model that is known to be consistent to another model which may be more efficient. I'm also not used to seeing it in the multi-level model context: I usually see it used with the panel style fixed and random effects models.

So what two models do you want to compare here exactly?