Gabriel:
The post hoc comparison tests that you reference are of doubtful validity or
utility in anything but balanced designs with simple covariance structures.
With missing data there are two critical issues: why are the data missing
and how do they need to be handled as a result? -- just ignoring them may
produce biased results if it's "informative" missingness, and
inference is
even more of a headache(it's difficult, unknown, or unresolvable depending
on the details). I strongly suggest you consult a local statistical expert.
Bert Gunter
Genentech Nonclinical Statistics
-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
On
Behalf Of Kaufman Gabriel
Sent: Friday, April 16, 2010 12:32 PM
To: r-help at r-project.org
Subject: [R] Multiple comparisons on Anova.mlm object
I would like to perform multiple comparisons or post-hoc testing on the
independent variable in an Anova.mlm object generated by the Anova function
of the car package. I have defined a multivariate linear model and
subsequently performed a repeated measures ANOVA as per the instructions in
section #3 of the following comprehensive tutorial on the subject from the
Gribble lab at UWO:
http://gribblelab.org/2009/03/09/repeated-measures-anova-using-r
Unfortunately, since my data has missing values,I can't seem to use the
classical univariate approaches of aov() or lme() (suggested in sections #1
and #2 of the tutorial linked to above).
The relevant portions of the R console output are copied below (redacted
somewhat for intellectual property considerations). In sum, I am stuck at
the Anova.mlm object, as I cannot seem to apply any of the standard multiple
comparisons functions such as pairwise.t.test or p.adjust....
Thank you in advance for your help.
Gabriel Kaufman
Orthopedic Molecular Biology Laboratory
Centre de recherche CHU Sainte-Justine
Montreal, Quebec
--------------------------
R console output
> # define Treatment group as the factor defining the intra-subject model
> as.factor(Treatment)
> ## define repeated measures linear model
> # define repeated-measures data as matrix vector
> RM <- cbind(repeated_measure_1, repeated_measure_2, repeated_measure_3,
repeated_measure_4, repeated_measure_5)> mlm <- lm(RM ~ Treatment, data = RMdata.file)
> # load required package car
> library(car)
> ## Define Anova model object for repeated-measures ANOVA
> # define idata data frame
> idata <- data.frame(RM = factor(1:5))
> # define Anova object
> mlm.aov <- Anova(mlm, idata = idata,idesign = ~RM, type =
"II")
> # display class of Anova object
> class(mlm.aov)
[1] "Anova.mlm"> # display session information
> sessionInfo()
R version 2.10.1 (2009-12-14)
i386-apple-darwin9.8.0
locale:
[1] en_CA.UTF-8/en_CA.UTF-8/C/C/en_CA.UTF-8/en_CA.UTF-8
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] car_1.2-16
______________________________________________
R-help at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.