Hello, I am trying to compare 4 different groups (black women, white women, black men, and white women) and their probability for leadership. Leadership is a dichotomous DV measured as (0,1) (oldleadership) and the fixed effects controls for differences in state legislatures across states. It also includes an interaction variable. This is the code I have ran and the output: xtlogit oldleader newideology female##race seniorityterms majoritymember, i(statefe) fe


Array

Array Array
Of course, if I run margins after this, the output would be nonsensical. I have no idea what else I should try. Can you not predict probabilities with panel data, a dichotomous dv, and a fixed effect? Thank you so much for your attention to this question.

Sincerely,
Jatia Wrighten