Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
| From | Adam Cheung <adam_kalok@yahoo.com.hk> |
| To | statalist@hsphsun2.harvard.edu |
| Subject | st: qreg and reghv with pooled cross-sectional data |
| Date | Tue, 22 May 2012 15:20:28 +0800 (SGT) |
statalist@hsphsun2.harvard.edu. Dear all Statalisters, I would like to compare the distribution (probably variance and quantiles)of an interval variable X across two or more waves of cross-sectional data. The variable is not normally distributed, so I prefer looking at the quantiles. Perhaps I will need to control another variable too. Can I just simply pool the repeated waves of data and run a -sqreg- like: sqreg X wavevar controlvar, q(.1 .25 .5 .75 .9) Is this method problematic because of the possible autocorrelation of the pooled data? Similarly, Is it problematic if I use -reghv- to use wave as a predictor of log-variance with a hetereoscdastic regression? Any suggestion? Thanks very much. Regards, Adam * * 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/