Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | josephine gakii <jogakii@yahoo.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: Interaction terms |
Date | Thu, 7 Apr 2011 01:55:38 -0700 (PDT) |
Thanks Marteen and the rest for the replies to these thread. yes i used the official Stata program -mprobit-. Because my error terms are correlated i settled on -mprobit- which has four categories in the dependent variable . i concur with Marteen that i have to use the delta method which gives the difference in the predicted probabilities. Unfortunately i am not conversant with the proposed -asmprobit-. could anyone please explain to me the -asmprobit-? again, does this imply i cannot use the delta method after running the -mprobit- command? i am also interested in understanding how the -predictnl- and -nlcom- commands work given that when i run my -mprobit- (assuming it is justified to use it) i get results in three categories ( the dependent variable has four categories); since the predictnl is on each variable, how can i extract the predictions of each variable in each of the three different categories of my dependent variable? thanks Josephine --- On Mon, 4/4/11, josephine gakii wrote: > I am using multinomial probit with an interaction term in > it. I would like to compute standard errors for meaningful > interpretation of my results that have an interaction > term. > > I have tried using the -inteff- command in STATA but it > seems to support only the binary logit and probit. i have > also tried the CLARIFY command in STATA but it does not > support multinomial probit > > Is there a command that can handle interaction terms in > multinomial probit? I am assuming you used the official Stata program -mprobit-. In that case there isn't much added value of using -mprobit- over using -mlogit-, as the error terms are assumed to be uncorrelated. If this does not worry you, than I would go for -mlogit- and interpret the relative risk ratios as discussed in: M.L. Buis (2010) "Stata tip 87: Interpretation of interactions in non-linear models", The Stata Journal, 10(2), pp. 305-308. If you want to include correlated error terms, then you first need to think about identification, e.g. through alternative specific variables as implemented in -asmprobit-. After that you'll probably need to compute the appropriate marginal effects for the interaction terms manually using either -nlcom- or -predictnl-. Those programs will than compute the standard errors (using the delta method) for you. Hope this helps, Maarten -------------------------- Maarten L. Buis Institut fuer Soziologie Universitaet Tuebingen Wilhelmstrasse 36 72074 Tuebingen Germany http://www.maartenbuis.nl -------------------------- * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/