It is a bad idea to write me privately. First, because there is some advise on the FAQ not do so, and it's clearly explained why it's a bad idea (in this case, while I know what -ssm- _ which incidentally you can download from SSC _ does, I never used it). Second, because I asked my e-mail to treat as spam any email but from [email protected], and you are lucky it took only 3 days to read it (it could have taken more than an entire life).
Therefore I am cc:-ing Statalist in the hope that someone could help you.
Nicola
P.S. I'll NOT receive/read any email but the Digest.
________________________________
Da: Clara Velasquez Garzon [mailto:[email protected]]
Inviato: ven 19/12/2008 13.40
A: Nicola Baldini
Oggetto: help about ordered probit models in stata
Dear Nicola,
I read (in statalist) an aswer from you about estimating the likelihood function of the ordered probit model with sample selection, you advised to use the ssm comand in STATA. I would like to use this command to estimate an ordered model to compute the probability of finding a job matching the level and type of studies, considering three alternatives: no matched, matched only for level or field and fully matched. But as we work only with young people with a level of education obove secondary we believe that we must take into account the possible problem of sample selection, the question is:
We have in our ordered probit model the dependent variable "match" and 22 dummy and 2 continuous independent variables. In the selection equation we have the dependent variable "Z" and 7 dummy and 1 continuous independent variables. To run the program we get the error: "conformability error. I have checked the command many times and apparently I am doing it as the manual says, could you tell me if there is something that I am not taking into account, or maybe the problem is that we have too many dummy variables?
Here it is my code.
ssm match sexo fpomedio fposup diplom letras tecnicasmi csaludfq /*
*/fpm_letras fpm_tecnicasmi fpm_csaludfq /*
*/fps_letras fps_tecnicasmi fps_csaludfq /*
*/ dip_letras dip_tecnicasmi dip_csaludfq /*
*/ unipa bachipa secunpa unima bachima secunma tca_paroa, /*
*/ s( Z = sexo tedadfe unipa bachipa secunpa unima bachima secunma pasalariado tca_paroa)q(8) f(binom) link(oprobit) sel
Could you help in finding the mistake?
Thanks,
Clara Velasquez
Universidad del Pais Vasco
Bilbao - España
*
* 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/
