Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | "Aurélie P. Harou" <aph2142@columbia.edu> |
To | statalist@hsphsun2.harvard.edu, Alfonso Sanchez-Penalver <alfonso.statalist@gmail.com> |
Subject | Re: st: cmp command; mvtobit |
Date | Thu, 20 Mar 2014 14:23:54 -0400 |
Thank you Alfonso, I am trying to run: ******************** cmp setup set more off #d;cmp (duration_pre_04_yr = age age_sq sex school_primary school_middle school_higher hh_n hhn_work yrs_in_village sni Pokrom Darmang Oboadaka ph om ph_sq) (duration_post_04_yr = age age_sq sex school_primary school_middle school_higher hh_n hhn_work yrs_in_village sni Pokrom Darmang Oboadaka ph om ph_sq),
ind($cmp_left "cond(duration_post_04,$cmp_left,$cmp_right)") technique(dfp) nonrtolerance difficult ; #d cr ******************* with the mvtobit command: ****************** #d;mvtobit (duration_pre_04 = age age_sq sex school_primary school_middle school_higher hh_n hhn_work yrs_in_village sni Pokrom Darmang Oboadaka ph om ph_sq) (duration_post_04 = age age_sq sex school_primary school_middle school_higher hh_n hhn_work yrs_in_village sni Pokrom Darmang Oboadaka ph om ph_sq),
cluster(villageid); #d cr ******************** Thank you! -Aurelie Quoting Alfonso Sanchez-Penalver <alfonso.statalist@gmail.com>:
Salut Aurelie,The exact -cmp- command that you're trying to run would be helpful to try and figure out what may be going wrong.Alfonso Sanchez-PenalverOn Mar 20, 2014, at 2:14 PM, "Aurélie P. Harou" <aph2142@columbia.edu> wrote:To Whom It May Concern: I am trying to estimate a bivariate tobit model with left censoring in one model and left- and right-censoring in the other. I believe the stata command mvtobit only allows for left-censoring at 0. I have tried using the cmp command, but I get "Warning: regressor matrix for duration_post_04_yr equation appears ill-conditioned" and do not get any results for the full model. I get this even when I reduce the number of RHS variables so as to get a condition number of 700. Furthermore, when I specify the cmp command to estimate the same model as the mvtobit command (i.e., left censoring at 0), I still get thesame warning message and the full model is not estimated (whereas it is estimated using mvtobit).Why/how is this happening? How can I get the cmp command to converge? Thanks in advance for any help! Best, Aurelie --------------------------- Aurélie P. Harou, PhD Earth Institute, Post-doctoral Fellow Agriculture and Food Security Center Columbia University, Lamont Campus 61 Route 9W, Lamont Hall, 2F Palisades, NY 10964 c: 954-614-8161 aph2142@columbia.edu * * 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/
--------------------------- Aurélie P. Harou, PhD Earth Institute, Post-doctoral Fellow Agriculture and Food Security Center Columbia University, Lamont Campus 61 Route 9W, Lamont Hall, 2F Palisades, NY 10964 c: 954-614-8161 aph2142@columbia.edu * * 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/