Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Jennifer M Keister <jkeister@ucsd.edu> |
To | "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> |
Subject | st: Problem with Seemingly Unrelated Regression (SUR) |
Date | Fri, 14 Jan 2011 13:28:12 -0800 |
I'm hoping someone might have some advice on a survey analysis problem I'm having. The analytical problem is analogous to the below: Respondents are asked to rate how much they enjoy each of the following fruits (on a 0-3 scale, with "Don't know" and "Refuse to answer" also provided as options). ORANGES: "How much do you enjoy oranges?" 0) Very little 1) Some 2) A lot 3) Don't know 4) Refused to answer Same question format for: BANANAS, APPLES, GRAPES Each record includes a "grade" for each fruit (again, on a 0-3 scale) reflecting how much that respondent dis/likes the fruit. Grades are thus not forced to be mutually exclusive. I would like to compare the "grade" each fruit receives, so that I may make comparative statements about popularity across fruit. For example "On average, respondents rate ORANGES more highly than GRAPES." Thus, I need to be able to compare responses across the "fruit variables." Seemingly unrelated regression (SUR) seems a viable option to perform this analysis. Unfortunately, the SUREG command in Stata does not seem to allow pweights or clustered standard errors, which is a problem since I'm working with survey data. Is there a way around this problem? Or an alternate analysis that does allow pweights and clustered SEs? Best regards, Jen * * 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/