[R] how to analyze repeated measures count data?

René Mayer mayer at psychologie.tu-dresden.de
Mon Mar 22 15:31:05 CET 2010

Dear R community,

I've data-set with reaction times and count data (answers - yes, no)  
of N subjects under conditions A, B.
For the analysis reaction time I used aov.

fit.rt = aov(rt ~ A * B + Error(subjects/(A*B)), data = m )

But how do I analyze the frequencies correctly?

example fable of frequencies from one subject:

, , = A1

       	B1 	B2 	B3
   yes   31     36    19
   no    22     27    10
, ,  = A2

       B1       B2    B3
   yes   22     27    10
   no    31     36    19

Is a generalized linear model the right method?
How do I specify the same model for the count data (frequencies) in glm?

is this right: glm(count~A*B*answer+(1|subject),family=poisson)?

Regards, René

More information about the R-help mailing list