log pseudo likelihood and wald test in logit model

[Dorgyles Kouakou]

Hi all,

Why does stata report log pseudo likelihood (not log likelihood) and wald test (not LR test) when we use the robust option in logit model estimation?

Can anyone help me!!!

Thanks a ton.

[Joao Santos Silva]

Dear Dorgyles Kouakou,

You ask for robust standard errors because you assume the likelihood is not correctly specified and therefore the problem becomes a pseudo MLE estimation and the LR rest is not veported because it would not be valid.

Best wishes,

Joao