I am estimating the whether floods affect electoral outcomes and I am using a difference-in-difference estimator. I have run my probit model with reelected party as the dependent variable (1 if reelected, 0 if not) and I have my group variable (1 if flooded), time variable (before and after) and the treatment effect. When I run this regression everything is fine, however as soon as I try to add any other independent variable STATA says it omits it due to collinearity.
I have quite a few dummy variables in my sample, for example when I try to interact large or small flood with the treatment effect it says there's collinearity but it also says this for other variables that are not dummy variable such as the total area flooded. How do I fix this? It happens everytime I try to interact anything with the treatment effect but I think I need to interact my other independent variables with it becuase, for example flood size only affects the treatment group. Pleae help!
