Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | James McJoseph <james.mcjoseph@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: Heckprob - postestimation graphs |
Date | Wed, 1 Dec 2010 12:12:53 +0200 |
Thnks, Maarten. I have replicated the graph successfully with my data. The tricky part is that one of the variable that I allow to change, (holdind the rest at means), has interaction. Now the interaction is left at means. I suppose, in this setting the graph is only telling half the story (better than nothing). The next question is, how can I include a command in there, to instruct stata to keep the dummy variables at their observation value, while maintaining means for the continous variables. Much, thanks, James. On Wed, Dec 1, 2010 at 11:18 AM, James McJoseph <james.mcjoseph@gmail.com> wrote: > Much thanks for this, Maarten. You sure understood my needs. Quick > confession "I am a newbie". I did replicate this code with my data, > it gives me a straightline. Obviously it must be plotting the mean. > Plse make me understand this part of the code. How do I specify the > "variable that changes". Say, loginc is the one that "changes" ( and > it assumes its sample values). Plse help me further. > > // set remaining variables at their mean > foreach var of varlist logptax loginc { > sum `var' if e(sample), meanonly > replace `var' = r(mean) > } > * * 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/