Re: st: Stumped...xtmixed and ANOVA F-stats not agreeing for balanced design

["Joseph Coveney" <jcoveney@bigplanet.com>]

David Airey wrote: 

Math is helpful! I think I remember reading the mixed model routine in JMP 9
allows negative estimates of variance components. I don't quite understand the
choice between SAS and StataCorp statisticians on this point.

--------------------------------------------------------------------------------

I get the impression that the NOBOUND option of PROC MIXED is used more often
than not for diagnostic purposes for a nonpositive-definite G matrix, so it
might not be so much a difference between statisticians.

Joseph Coveney

Math would be even more helpful if I could do it:

> MS_e = 0.00273899 = sigma2_e
> MS_s#a = 0.012825848 = sigma2_e + 2 * sigma2_s#a
> MS_s#b = 0.014614037 = sigma2_e + 3 * sigma2_s#b
> MS_s = 0.02026831 = sigma2_e + 6 * sigma2_s + 2 * sigma2_s#a + 3 * sigma2_s#b
> 
> sigma2_s#a = (0.012825848 - 0.00273899) / 2 = 0.00504343
> sigma2_s#b = (0.014614037 - 0.00273899) / 3 = 0.00395835
> sigma2_s = (0.02026831 - 0.01008686 - 0.01187505) / 6 = -0.00028227

sigma2_s = (0.02026831 - 0.00273899 - 0.01008686 - 0.01187505) / 6 = -0.00073876


*
*   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/