Dear statalist,
I have mailed before about a panel data set I am working with using 38 countries over the period 1980 to 2005. This uses age standardized mortality rates as the main outcome variable.
I notice that I get very similar results using xtgee f(p)vce(r) and xtreg, fe i(Country) vce(r), the main difference being that the xtgee has smaller standard errors than the FE model (which I understand is normal).
Could this mean that the XTGEE correlation structure does not affect my results that much, and that I could have just used the FE model instead? I do realize that XTGEE is population level model, and therefore has different assumptions... but regardless of differences in their approach they appear to bring up very similar results.
I was thinking to include both in my work to demonstrate it is robust to different techniques.
Any comments would be very welcome.
Thanks,
A
_________________________________________________________________
Download new and classic emoticon packs at Emoticon World Brought to you exclusively by Windows Live
http://windowslive.ninemsn.com.au/emoticon.aspx?
*
* 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/
