Adjusting confidence intervals for paired t-tests of multiple endpoints

Erich Studerus, Psychiatrische Uni-Klinik <erich.studerus at bli.uzh.ch> [Sat, Feb 28, 2009 at 09:23:05PM CET]:
Thanks, but as far as I know the multcomp package and the tukeyHSD 
function can only handle one dependent variable, whereas I have 60 
dependent variables.
Couldn't you re-phrase your model by including timepoint as a continuous 
regressor and scale as a factor?
Johannes H?sing               There is something fascinating about science. 
                              One gets such wholesale returns of conjecture 
mailto:johannes at huesing.name  from such a trifling investment of fact.                
http://derwisch.wikidot.com         (Mark Twain, "Life on the Mississippi")
Couldn't you re-phrase your model by including timepoint as a continuous 
regressor and scale as a factor?
Well, the measuring time points are 70 170, 300 and 1440 minutes after drug 
intake. Since the influence of time is cerainly not linear, I think it's 
better to treat it as categorical factor. In fact, I already calculated 
mixed effects models with the four-level factor time and the two-level 
factor treatment and random effects factors subjects (nested in studies) and 
studies for each scale. I used mixed effects models, because I'm doing a 
meta-analysis on the raw data of 8 studies. Thus my dataset has not only a 
grouping structure, but is also unbalanced (not all studies did 4 
measurements). I'm not sure, if I could extend my model by including another 
factor for scale, since these scales measure quite different things. I 
guess, It would be also very difficult the set up appropriate contrasts to 
compare each scale on each measuring time point.
Hi Erich,

Look at the SimComp package, which handles multiple endpoints, i.e. columns
of response variables. Ratio-comparisons are also handled. Good ready-made
range of comparisons; you can also set up your own.

Regards, Mark.
  Johannes Huesing <johannes at huesing.name> wrote:
Couldn't you re-phrase your model by including timepoint as a continuous 
regressor and scale as a factor?
Well, the measuring time points are 70 170, 300 and 1440 minutes after
drug 
intake. Since the influence of time is cerainly not linear, I think it's 
better to treat it as categorical factor. In fact, I already calculated 
mixed effects models with the four-level factor time and the two-level 
factor treatment and random effects factors subjects (nested in studies)
and 
studies for each scale. I used mixed effects models, because I'm doing a 
meta-analysis on the raw data of 8 studies. Thus my dataset has not only a 
grouping structure, but is also unbalanced (not all studies did 4 
measurements). I'm not sure, if I could extend my model by including
another 
factor for scale, since these scales measure quite different things. I 
guess, It would be also very difficult the set up appropriate contrasts to 
compare each scale on each measuring time point.

______________________________________________
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.

View this message in context: http://www.nabble.com/Adjusting-confidence-intervals-for-paired-t-tests-of-multiple-endpoints-tp22247598p22283385.html
Sent from the R help mailing list archive at Nabble.com.
Thanks, this comes close to what I need, but the problem is that I have
repeated measures in each comparison and SimComp does not seem to handle
this. In my experiments, each subject received both placebo and the drug in
a randomized order. So, I want to do multiple paired t-test corrected for
multiple comparisons. I'm primarily interested in the corrected confidence
intervals because I want to plot the mean differences with it.

Regards,
Erich

-----Urspr?ngliche Nachricht-----
Von: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] Im
Auftrag von Mark Difford
Gesendet: Montag, 2. M?rz 2009 08:34
An: r-help at r-project.org
Betreff: Re: [R] Adjusting confidence intervals for pairedt-tests of
multiple endpoints

Hi Erich,

Look at the SimComp package, which handles multiple endpoints, i.e. columns
of response variables. Ratio-comparisons are also handled. Good ready-made
range of comparisons; you can also set up your own.

Regards, Mark.
  Johannes Huesing <johannes at huesing.name> wrote:
Couldn't you re-phrase your model by including timepoint as a continuous 
regressor and scale as a factor?
Well, the measuring time points are 70 170, 300 and 1440 minutes after
drug 
intake. Since the influence of time is cerainly not linear, I think it's 
better to treat it as categorical factor. In fact, I already calculated 
mixed effects models with the four-level factor time and the two-level 
factor treatment and random effects factors subjects (nested in studies)
and 
studies for each scale. I used mixed effects models, because I'm doing a 
meta-analysis on the raw data of 8 studies. Thus my dataset has not only a
grouping structure, but is also unbalanced (not all studies did 4 
measurements). I'm not sure, if I could extend my model by including
another 
factor for scale, since these scales measure quite different things. I 
guess, It would be also very difficult the set up appropriate contrasts to
compare each scale on each measuring time point.

______________________________________________
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.

View this message in context:
http://www.nabble.com/Adjusting-confidence-intervals-for-paired-t-tests-of-m
ultiple-endpoints-tp22247598p22283385.html
Sent from the R help mailing list archive at Nabble.com.

______________________________________________
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.