BUG summary2 doesn't propagate/use use_t

[josef-pkt]

I'm not sure yet how this works, but I don't see how use_t ("t" or "z") is propagated or used in summary2

found while fixing #1812

It's a bug. See discrete_choice_example.py
print(affair_mod.summary2()) show "t" instead of "z"

[vincentarelbundock]

The coef/se/t/ci panel of your ascii summary is a dataframe that is (typically) constructed using a helper function. You need to make sure that your result class' summary method passes the use_t argument down to the summary_params function:

https://github.com/statsmodels/statsmodels/blob/master/statsmodels/iolib/summary2.py#L300

[vincentarelbundock]

But now that I look at summary_params(), I expect you might want it to do something else...

[josef-pkt]

@vincentarelbundock I changed it here 4b9f7ee
In add_base

[vincentarelbundock]

Cool. I've been debating (internally, in my head) whether having add_base is a good idea. Seems like a wrapper around few things. Might be better to just do it explicitly. Oh well.

[josef-pkt]

It doesn't really matter for this where we do define the argument for use_t.

I'm reusing the paramstable for other things like t_test where we need the use_t option. For the results summary I want to start to hardcore the use of the use_t results attribute.

[vincentarelbundock]

Good idea.