Hi Everyone,

I want to run a probability model by using high-dimensional fixed effects to understand the effect of X on Y. To do so, I picked the linear probability model since reghdfe command provides a computational advantage. However, half of my predicted values are negative. I have two questions:

i) Would it be ok under some circumstances?
ii) Does logit or probit have a similar command to reghdfe?

Thanks in advance.

Ulas