Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
| From | CARLOS MAGNO SANTOS CLEMENTE <carlosmagno_sc@ig.com.br> |
| To | statalist@hsphsun2.harvard.edu |
| Subject | Re: st: RE: Propensity Score Matching |
| Date | Sat, 12 May 2012 12:19:00 -0300 |
Thanks, I'll try again 2012/5/12 Fernando Rios Avila <f.rios.a@gmail.com>: > Hi Carlos > So one thing you need to know is that there is no RULE to structure > the model for a propensity score matching setup. And unless one is > working with your data, its going to be impossible to just give you a > model that will solve your problem. > As you may know, propensity score matching (PSM) relies on finding > observations who are treated and untreated who are similar enough to > be used as counter factual. The error you reported just means that > based on -pscore- analysis, after estimating the propensity score and > matching, some characteristics are sill statistically different. > What it is suggested is to try a different specification, which > basically means you have to either include more variables as > covariates, or include either higher terms (square cubes) and or > interactions. > Again, there is no magical rule, you just need to keep trying until > one of the specification works for you. > Bests. > Fernando > On Sat, May 12, 2012 at 11:01 AM, CARLOS MAGNO SANTOS CLEMENTE > <carlosmagno_sc@ig.com.br> wrote: >> I am using the following command: >> >> "pscore binaria pop_2010 pib_09 s_inad_10 analf_10 medi_ind_ >> decl_med_2 decl_med_2 dista_capi, pscore(mypscore) blockid(myblock) >> comsup level (0.005) logit" >> >> >> binaria = dummy >> pop_2010 pib_09 s_inad_10 analf_10 medi_ind_ decl_med_2 decl_med_2 >> dista_capi = covariables >> >> Could help me in structuring the model? >> >> Thank you >> >> >> >> >> 2012/5/11 Sarah Edgington <sedging@ucla.edu>: >>> Carlos, >>> The glib answer is that the message you're getting already tells you what >>> procedure to take "Try a different specification of the propensity score." >>> As noted in the Stata FAQ >>> (http://www.stata.com/support/faqs/res/statalist.html#question) you should >>> specify what command you're actually using, and where it comes from if it >>> isn't an official Stata command. Without that information and at least a >>> summary of what you're trying to do, it's going to be very difficult to >>> offer any more specific advice. >>> >>> Have you read the article "Estimation of average treatment effects based on >>> propensity scores" by Becker and Ichino, which appears in the Stata Journal >>> (2002 v2 #4)? http://www.stata-journal.com/sjpdf.html?articlenum=st0026 >>> I found that discussion helpful when implementing an analysis with >>> propensity scores. >>> >>> Good luck. >>> >>> -Sarah >>> >>> -----Original Message----- >>> From: owner-statalist@hsphsun2.harvard.edu >>> [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of CARLOS MAGNO >>> SANTOS CLEMENTE >>> Sent: Friday, May 11, 2012 3:41 PM >>> To: statalist@hsphsun2.harvard.edu >>> Subject: st: Propensity Score Matching >>> >>> PSM the following error occurs. >>> >>> he balancing property is not satisfied >>> >>> Try a different specification of the propensity score >>> >>> What procedure to take? >>> -- >>> Att, >>> >>> Carlos Magno Santos Clemente >>> >>> Geógrafo, Mestrando pela Universidade >>> >>> Estadual de Montes Claros – UNIMONTES >>> >>> Atuação: Geotecnologias >>> >>> * >>> * 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/ >>> >>> >>> * >>> * 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/ >> >> >> >> -- >> Att, >> >> Carlos Magno Santos Clemente >> >> Geógrafo, Mestrando pela Universidade >> >> Estadual de Montes Claros – UNIMONTES >> >> Atuação: Geotecnologias >> >> * >> * 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/ > > * > * 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/ -- Att, Carlos Magno Santos Clemente Geógrafo, Mestrando pela Universidade Estadual de Montes Claros – UNIMONTES Atuação: Geotecnologias * * 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/