Date:         Thu, 8 Jul 1999 14:43:54 EDT
Reply-To:     La8arrival@AOL.COM
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         La8arrival@AOL.COM
Subject:      Error in Mixed Model F test
Comments: cc: nichols@spss.com
Content-Type: text/plain; charset="us-ascii"

I am using SPSS graduate pack 8.0.  When I run a GLM two factor ANOVA with A
fixed and B random, SPSS computes the F observed for B as MSb / MSab instead
of MSb/MSe as it should be.  SPSS correctly computes Fobs for A as MSa/MSab.
The variance component table SPSS prints out appears quite correct so this is
a mystery.  Any thoughts?  It seems too obvious for it to be an SPSS error
but I don't know what else it could be.

Mike Conley