[R] question about anova

Waverley waverley.paloalto at gmail.com
Fri Oct 19 20:13:32 CEST 2007

I have a question regarding two way anova.
I have a data sheet as following:
ID -- representing the same sample gets several repeated measures at
the same time point
group -- treated vs untreated
time -- different time point (3, 9, 15 hours)
signal -- measurement signal

I use the aov function and formula = signal ~ (group*time)  + Error


1. is this the right way Error (ID/(group*time)
 to handle multiplicated repeated measures of the same sample (by ID)
at the same time point?  I have checked on the web and it looks like
there is some controvercial discussion.  Can you educate me here?

2.  In the result of the model.tables (aov.study, "means"), the means
value under "group:time" of treated and untreated vs time point is
different from what I have calculated by hand. It is fairly close to
that value.  Can someone explain why I have discrepancies here?

3. a R program tech question: I need to extract the p value from the
summary(aov.study) but not sure how to do it.  looks like the summary
is not a data frame or list.  I need to take the p value and continue
to use it in the R program for some other reports.

Thanks much in advance.

Waverley @ Palo Alto

More information about the R-help mailing list