Dear All

I'm fitting a hurdle model (with the command churdle linear) and use the margin (margins, dydx(CategoricalVariable)). I wonder why the statistical significance is different between the hurdle linear and the margin.

Hurdle linear part

SED_wkday | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------------+----------------------------------------------------------------
SED_wkday |
speaking_wk | -13.15755 16.96346 -0.78 0.438 -46.40532 20.09022
texting_wk | -22.34044 11.64429 -1.92 0.055 -45.16282 .4819424
social_net_wk | -50.51344 8.730203 -5.79 0.000 -67.62433 -33.40256
browsing_wk | -24.73634 11.57269 -2.14 0.033 -47.41838 -2.054286


Margin

| Delta-method
| dy/dx Std. Err. z P>|z| [95% Conf. Interval]
--------------+----------------------------------------------------------------
speaking_wk | -13.60962 10.8538 -1.25 0.210 -34.88268 7.663439
texting_wk | -9.302818 7.627541 -1.22 0.223 -24.25252 5.646887
social_net_wk | -19.89934 5.714631 -3.48 0.000 -31.09981 -8.698867
browsing_wk | -.7922133 7.934503 -0.10 0.920 -16.34355 14.75913


Please could someone help me? what's going on and why the statistical significance differ?

In addition, how could I fit an interaction in hurdle model?

Thank you
Ellie