- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
I am still new in SAS and trying to modelize my first probit model. I am using the following sequence of command in order to obtain my probit estimate:
proc logistic data=pop_final;
class SEX (REF="M") AGEGR (REF="3") NON_EE (REF="0")/PARAM=REF;
model SCIENCE = SEX AGEGR NON_EE / link=probit technique=newton;
ods output parameterestimates=prbparms;
output out=outprb prob=p xbeta=xbpr;
run;
However, I would now like to replicate the margins command in STATA (http://www.stata.com/manuals13/rmargins.pdf). It allowed me to estimate margins of responses for specified values of covariates of a previously fit model. For example, following my probit model that estimated, in particular, the effect of changing industry of employment following a layoff, I was able to use the margin command to estimate the probability of a workers being laid-off from a specific industry to change industry in his following job.
Would someone know a command in SAS that could replicate what I was able to do with STATA?
Thank you in advance
Accepted Solutions
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Sounds like you want to estimate so-called marginal effects which are the derivative of the event probability with respect to a predictor of interest. See this note.
- Tags:
- marginal effects
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
I think you might be looking for the, "LS-Means" option??
Also, here is this Proc GENMOD Logistic Regression . GENMOD will let you use the LS-Means statement.
Alternatively, Proc GLIMMIX . Even though you may not ever use the random intercepts or slopes. I tend to go with Glimmix, because the "NLOPTIONS" statement lets you be fairly flexable if the model has a difficult time converging.
I think each one of those procedures has information on the LS-means functionality.
Good luck!
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Sounds like you want to estimate so-called marginal effects which are the derivative of the event probability with respect to a predictor of interest. See this note.
- Tags:
- marginal effects
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Yeah, I totally misunderstood. What @StatDave said is where you need to be.
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Sorry for the delay in the reply, thank you for your help!