Skip_hessian True now works in ConditionalLogit

[Davidvandijcke]

The skip_hessian option did not work in ConditionalLogit.fit() because the ConditionalResults function tried to retrieve cov_params() from LikelihoodModel.fit() results.

[pep8speaks]

Hello @Davidvandijcke! Thanks for opening this PR. We checked the lines you've touched for PEP 8 issues, and found:

• In the file statsmodels/discrete/conditional_models.py:

Line 124:1: W293 blank line contains whitespace

[josef-pkt]

the fix looks correct to me, but I'm not familiar with the internals of these models.

Thanks for finding this and the PR

In general, skip_hessian was mainly introduced for internal use to avoid computing the hessian or cov_params more than once. I guess it's never directly unit tested.

[Davidvandijcke]

That makes sense. For researchers, it can be a useful option for inspecting coefficient values in large models without having to compute the Hessian every time -- which can take many times longer.