Is there a Stata program/ command for cross-difference calculation of interactions in logits (following Ali & Norton 2003)?

[Rosa Blau]

Hi,

I'm calculating two logit models, and want to compare estimators in the first model, which are not part of on interaction, to estimators of interactions in the second model. Following Mood (2010) I'm using AME to interpret the first model (the one without interaction effects). But I know that you can't accurately calculate AMEs for interaction effects in logits. I'm now trying to follow Ali & Norton (2003) suggestion of calculating a cross-difference. My question is: has a Stata program / command been developed for this yet?

Best,
Rosa

[Rosa Blau]

I've now found this extremely helpful thread: https://www.stata.com/statalist/arch.../msg00293.html. As one of my interaction terms is categorical and the other is binary, I have decided to plot the average marginal effect of by dummy at all three intervals of my categorical. Can you point me to a citable publication that suggests this use of plots to understand the effect of an interaction of two categorical independents on a categorical dependent?

[Richard Williams]

This might do what you want (and it is an excellent paper to read regardless):

https://journals.sagepub.com/doi/10....81175019852763

Abstract

Many research questions involve comparing predictions or effects across multiple models. For example, it may be of interest whether an independent variable’s effect changes after adding variables to a model. Or, it could be important to compare a variable’s effect on different outcomes or across different types of models. When doing this, marginal effects are a useful method for quantifying effects because they are in the natural metric of the dependent variable and they avoid identification problems when comparing regression coefficients across logit and probit models. Despite advances that make it possible to compute marginal effects for almost any model, there is no general method for comparing these effects across models. In this article, the authors provide a general framework for comparing predictions and marginal effects across models using seemingly unrelated estimation to combine estimates from multiple models, which allows tests of the equality of predictions and effects across models. The authors illustrate their method to compare nested models, to compare effects on different dependent or independent variables, to compare results from different samples or groups within one sample, and to assess results from different types of models.

[Maxence Morlet]

Rosa Blau you might find these commands (user-written) interesting: inteff (there might also be inteff2 or inteff3, I'm unsure). inteff definietely exists though.