discreteResults margeff method not tests, obsolete

[josef-pkt]

the margeff method has zero test coverage, get_margeff is the new method

is margeff obsolete ?
or does it need unit tests comparing to separated out Margins class ?

[josef-pkt]

@jseabold Why do we deprecate a broken function ?

for example Logit case here (there is PR #892 for fixing the examples)
http://statsmodels.sourceforge.net/devel/examples/notebooks/generated/example_discrete.html

logit_mod = sm.Logit(spector_data.endog, spector_data.exog)
logit_res = logit_mod.fit()
print 'Parameters: ', logit_res.params
print 'Marginal effects: ', logit_res.margeff()

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-5-761e1b345c46> in <module>()
      2 logit_res = logit_mod.fit()
      3 print 'Parameters: ', logit_res.params
----> 4 print 'Marginal effects: ', logit_res.margeff()

e:\josef\eclipsegworkspace\statsmodels-git\statsmodels-all-new2\statsmodels\statsmodels\discrete\discrete_model.pyc in margeff(self, at, method, atexog, dummy, count)
   2111                                                 self.model._derivative_exog,
   2112                                                 dummy_ind, count_ind,
-> 2113                                                 method)
   2114         # don't care about at constant
   2115         self.margeff_cov = margeff_cov[ind][:, ind]

TypeError: margeff_cov_with_se() takes exactly 10 arguments (9 given)

Optimization terminated successfully.
         Current function value: 0.402801
         Iterations 7
Parameters:  [  2.82611259   0.09515766   2.37868766 -13.02134686]
Marginal effects: 

e:\josef\eclipsegworkspace\statsmodels-git\statsmodels-all-new2\statsmodels\statsmodels\discrete\discrete_model.py:2043: FutureWarning: This method is deprecated and will be removed in 0.6.0. Use get_margeff instead
  " Use get_margeff instead", FutureWarning)

[jseabold]

I don't think that case worked before either, and I didn't feel like fixing it since it's deprecated. I deprecated it because I had demo'd it in a few talks. I'm fine with removing it though. Or just raising a NotImplementedError to use get_margeff or something.

[josef-pkt]

I think removing it would be best in that case. I think it's strange to get a deprecation warning for something that doesn't and didn't work. If it didn't work before, then there should not be any code depending on it.

One possibility would be to delegate to the new class, if we want to keep the deprecation warning.