Interaction term of IV and control variable

[Melissa Mena]

Hallo everyone,

I have two questions.

The first is more basic.
For my first hypothesis, I use one IV.
For my second hypothesis, I test it through an interaction term of the IV from H1 and a control variable.

Is an interaction term described as an IV in the methodology part? Do I have now to describe the interaction term and the previous control variable as independent variable in my methodology part, although I am only interested in the interaction term?

Second, the DV is a dichotomous variable.
I want to get the marginal effect of the interaction term through an OLS regression, which is weakly significant.
Then, I want to plot the marginal effect on different levels of one variable through a logit regression. However these levels are not significant.

Can I use the insignificant results from logit to illustrate the significant result from OLS?

Thank you in advance!

Melissa

[Phil Bromiley]

You didn't get a quick answer. You'll increase your chances of a helpful answer by following the FAQ on asking questions; provide Stata code in code delimiters, readable Stata output, and sample data using dataex. We work better with results in the text rather than as pictures (which are often hard to read). Also, it is better to ask one or two questions rather than a long list.

When you interact an endogenous variable with another variable, you normally have to include both the endogenous variable and a calculated value for the interaction as endogenous variables [i.e., you can't use factor variable notation for the endogenous variables in ivreg and similar estimators]. The interaction is endogenous because it includes the endogenous variable.

With a dichotomous dv, you are probably better off to use an estimator that allows for such variables rather than ols. 2sls is consistent with a binary endogenous variable but not correct with a binary dv. You should look at the eregress estimators, cmp, or gsem.