Re: st: Setting the base group in dummy variable interaction terms

[Phil Clayton <philclayton@internode.on.net>]

Have you tried using -fvset base-?

Phil

On 30/09/2012, at 9:04 PM, Tim Burnett (ECO) wrote:

> Dear all,
> 
> I have a question which may b trivially easy to answer but unfortunately I can't find it.
> 
> I am using a random effects probit model as follows:
> 
> xtprobit  switched ib4.service ib1.supplier ib2.duration  bundled numberservs i.supplier#i.service  i.bundled#ib1.supplier i.bundled#ib4.service  ib1.income ib1.employment i.children i.gender ib1.education ib2.age if INCLUDEVAR==1, re
> 
> Which obviously includes many dummy variables.  I am using the prefix ib#. to set the base category for the dummy variables for the standalone dummies. Unfortunately this isn't working for the dummy interaction terms which I would like to be based upon the excluded categories of the standalone variables (4 for service, 1 for supplier) since in these cases the excluded category has specific relevance in terms of understanding the results.. 
> 
> Placing, respectively, ib1 and ib4 in the command above doesn't seem to do any harm, but Stata just ignores it as follows:
> 
> note: 1.bundled#6.supplier omitted because of collinearity
> note: 1.bundled#4.service omitted because of collinearity
> 
> It ignores my base category request.
> 
> Any suggestions?
> 
> Many thanks,
> 
> Tim
> 
> --
> 
> Tim Burnett
> PhD Researcher
> Department of Economics & ESRC Centre for Competition Policy
> University of East Anglia
> 
> Email: tim.burnett@uea.ac.uk
> Phone: +44 (0) 7793 116 522
> Skype: timb318
> 
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/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/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/