At the very least, you could perform a Hausman test, a likelihood-ratio test and Akaike's and Schwarz's Bayesian information criteria (AIC and BIC) tests. These are typically used to test between various models specifications to find the better of the models estimated.

Thanks Linden