[R-sig-ME] Origins of difference in significance between lmer/pvals.fnc and aov?

S Ellison S.Ellison at LGCGroup.com
Fri Oct 5 14:54:05 CEST 2012

> I have a data set which seems amenable to analysis either 
> through lmer or aov, so I am comparing their results. I'm 
> finding small p-values for the fixed effect of interest in 
> either case, but they differ quite a bit

One contributing factor is that if I read your output correctly, aov() is doing an F test using some quite low degrees of freedom (1?) for the top line, while lmer is testing a coefficient's t-statistic using (iirc) only the density degrees of freedom which is typically, if not always (?) the same as the residual dof. 

S Ellison

This email and any attachments are confidential. Any use...{{dropped:8}}

More information about the R-sig-mixed-models mailing list