marginal effects in logistic regression

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

Posts: 1
Joined: Thu Aug 31, 2017 3:39 pm

marginal effects in logistic regression

Postby Podrick » Thu Aug 31, 2017 4:15 pm

Hey guys,

I hear that SPSS doesn't directly support a function for marginal effects in logistic regression. So I tried to do this manually and the results do make sense but I wanted to make sure they are actually correct.

I basically substracted the probability predicted by the model for a given individual from the predicted probability if the x was 1 unit higher. As an example if I wanted to get the marginal effect for β1 I'd do this:

individual Marginal effect of β1 = 1/(1+e^-(α+(x1+1)*β1+x2*β2))) - 1/(1+e^-(α+x1*β1+x2*β2)))

And then take the mean of all individual marginal effects. Just wanted to make sure if this makes sense and I'm computing the right thing here.

Also, am I right when I say it doesn't make sense to compute this for explanatory Dummy-Variables and I shoud use the Odds-Ratio instead?

I'm sorry for my poor English and very grateful for any comment on this.

Who is online

Users browsing this forum: No registered users and 2 guests