[R] aov and errors used for interaction effects

O.Langner o.langner at psych.ru.nl
Mon Mar 31 12:47:07 CEST 2008


Dear list-members,

I just switched to R, so here probably a newbie question, I couldn't find a
answer for yet:

Assume a design with three repeated measures factors (F1-F3), I tried to run
the following:
aov(rt ~ F1*F2*F3 + Error(subject / (F1*F2*F3)))

Comparing the output from here and SPSS (General Linear Model -> Repeated
measures) I found, that R does all calculations right for the main effects
of F1-F3, but seems to do something different for the interactions: Whereas
SPSS gets the F for an interaction by dividing the MS of the interaction
term (e.g. F1*F2) by the corresponding MS of the interaction with subject
(then F1*F2*subject), R just divides all interaction MS by the residual MS.

Could somebody point out to me, what I'm doing wrong or how I can get aov to
use the appropriate error terms also for the higher order interactions?

Thanks for helping!
-- 
View this message in context: http://www.nabble.com/aov-and-errors-used-for-interaction-effects-tp16395617p16395617.html
Sent from the R help mailing list archive at Nabble.com.



More information about the R-help mailing list