Dear STATA users,

I am running a model with a logit regression (binary DV), in which I include my IV and the sqaure of my IV testing for a quadratic relationship. How can I calculate and discuss the effect size here? in a normal logit, usually I calculate the average marginal effects, but here with the square terms what should I do?

Next, I also interact the IV and the sqaure of IV with a moderator, testing for an interaction. Once again, how can I calculate and discuss the effect size of such result?

Basically the model looks like:

Y = a1 + b1 + b1*b1 + a1*b1 + a1*b1*b1 + controls

Thanks in advance!

best