Hi Statalist,

I am currently working on a probit model in which the dependant variable is labor force participation (1 if the person participates and 0 if not) and the independent variables are the age group (0 to 2), the birth cohort (0 to 9) and other discrete and continuous variables.

I did a simple probit and then used the margins command to obtain the marginal effects at the means. For the latter, I separated the results for age group:

Code:
probit lfp i.age_group i.cohort other_variables
margins, dydx(i.cohort) atmeans over(age_group) post
The thing is that, by doing this, I get the results comparing the change from the base level (the output says "Note: dy/dx for factor levels is the discrete change from the base level."). In this case, It would be the marginal probability of being in the labor force for an average person of the age_group=1 and cohort=1, compared to the same average person of the age_group=1 but cohort=0.

My doubt is how can I get the discrete change not from the base level (always comparing with cohort=0), but from the previous level (cohort 1 vs cohort 0; cohort 5 vs cohort 4, etc)?.

Thanks in advance,
Isidora.