[R] Sphericity test in R for repeated measures ANOVA

Peter Dalgaard p.dalgaard at biostat.ku.dk
Tue Aug 21 07:56:00 CEST 2007

Orou Gaoue wrote:
> Hi,
> Is there a way to do a sphericity test in R for repeated measures ANOVA
> (using aov or lme)? I can't find anything about it in the help.
> Thanks
> Orou
There is for lm() with multivariate response (mauchly.test).

For lme(), you can compare models with a corSymm correlation structure 
to ones with corCompSymm. Thus is a similar test, but not quite the same.

For aov() it doesn't really make sense, partly because the 
"repeatedness" is  ambiguous in some models, partly because aov's 
internal algorithms rely strongly on orthogonality with respect to a 
particular covariance structure. If you relax assumptions, orthogonality 
no longer holds.

More information about the R-help mailing list