Hi there,

I am new to Stata 15 and need help deciding on a regression model for my panel data. I have data on 400 individuals from 2016 and 2018 (just the two years). My dependent variable is whether respondents' opinion shifted from 2016 to 2018 about a wind energy project built nearby. They're ordered 0-2; 0 = no shift/same opinion both years, 1 = shift to a more negative view, 2 = shift to a more positive view.

I have a variety of independent variables from survey questions, some were asked in both years, others in just one year. Examples include what respondents felt about the fairness of the planning process (just in 2016 survey), perceptions/reality about the look of the turbines (both 2016 and 2018 surveys), etc. I'd like to include 10-15 in my regression. There are fixed effects like gender and place of residence (no one moved over the time period), and random effects, like the ones I mentioned in the previous sentence.

I have rearranged the data into Long format. I've tried meprobit, GLLAAM, and mlogit, but I am getting overwhelmed at all of the other options out there. I would love to use mlogit because it compares, but I'm not sure that's appropriate for longitudinal data. Any recommendations?

Thanks in advance.