Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Sara Head <sara.head@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: generating predicted probabilities, setting variables constant at mean or mode |
Date | Mon, 29 Aug 2011 14:13:14 -0400 |
Hi Statalist, I am graphing some predicted probabilities from a Generalized Ordered Logistic Regression model. I am using Long & Freese's SPost commands (i.e. prgen and praccum) and Williams' gologit2 (in Stata IC 11.2). My model has both categorical and continuous predictors. When I run praccum, I use the x() option to assign specific values to some variables and then I use rest(mean) to set other variables constant at their mean. Rest(mean) obviously makes sense for continuous predictors, but it isn't intuitive for categorical variables. I can't find SPost examples where, after specifying x(), users set categorical predictors constant at their mode. Could someone provide insight into why I would set categorical predictors constant at their mean and not their mode? Thanks, Sara * * 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/