Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Tammy Leonard <tcl051000@utdallas.edu> |
To | <statalist@hsphsun2.harvard.edu> |
Subject | st: Margins with VCE Jackknife |
Date | Thu, 26 May 2011 16:57:41 -0500 |
Hello, I am trying to estimate marginal effects for an ordered logit regression where jackknife weights are specified via svyset. I issue the command: margins, dydx(*) predict(outcome(1)) vce(unconditional) And it returns the error message ³vce(jackknife) is not supported². Is there a way to obtain marginal effects when the survey design specifies jackknife weights? (I¹m using HINTS data.) In a further attempt to obtain marginal effects, this time at the sample mean (so vce(unconditional) is not necessary) I issue the command: margins, dydx(*) predict(outcome(1)) atmeans Which returns the result that the marginal effects are not estimable. However, using the old mfx command (which defaults to the atmeans calculation), marginal effects are quickly estimated. What difference in the two algorithms is causing this discrepancy? Are the results of mfx valid? Thanks for any and all help! Best, Tammy * * 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/