Dear R users,
I have a rather knotty analysis problem and I was hoping that
someone on this list would be able to help. I was advised to try this list
by a colleague who uses R but it is a statistical inquiry not about how to
use R.
In brief I have a 3x2 anova, 2 tasks under 3 conditions, within subjects. I
also took a variety of personality measures that might influence the results
under the different conditions. I had thought that an ancova would be the
best test, but it might be the case that this would not work with a within
subjects design. I have not found anything that explicitly states whether or
not it would, but all the examples I have read are between subjects design.
I also thought of investigating a manova, but it is not really the case
that I have more than one DV, it is the same DV in 6 different combinations
of task and condition.
There were 4 personality measures and I wanted to look at the degree to
which they affected the task/ condition interaction.
I have explained this briefly here, but I can of course provied more
details to anyone who can advise me further with this.
Thanks,
Mary-Jane Anderson
Information Analyst
Platform Project
Information Services Division,
NHS National Services Scotland,
Gyle Square,
1 South Gyle Crescent,
Edinburgh,
EH12 9EB.
0131 275 7163.
_________________________________________________________________
NHS National Services Scotland Disclaimer
The information contained in this message may be confidentia...{{dropped}}
Anderson, Mary-Jane wrote:> Dear R users, > I have a rather knotty analysis problem and I was hoping that > someone on this list would be able to help. I was advised to try this list > by a colleague who uses R but it is a statistical inquiry not about how to > use R. > In brief I have a 3x2 anova, 2 tasks under 3 conditions, within subjects. I > also took a variety of personality measures that might influence the results > under the different conditions. I had thought that an ancova would be the > best test, but it might be the case that this would not work with a within > subjects design. I have not found anything that explicitly states whether or > not it would, but all the examples I have read are between subjects design. > I also thought of investigating a manova, but it is not really the case > that I have more than one DV, it is the same DV in 6 different combinations > of task and condition. > There were 4 personality measures and I wanted to look at the degree to > which they affected the task/ condition interaction. > I have explained this briefly here, but I can of course provied more > details to anyone who can advise me further with this. >This sounds like a job for a Multivariate Linear Model (assuming that you have complete data for each subject or are prepared to throw away subjects with missing values). This lets you decompose the response into mean, effects of task and condition, and the interaction effect. Each component can then be separately tested for effect of predictors, using multivariate tests, or F tests under sphericity assumptions. Have a look at example(anova.mlm); this mostly looks at cases where effects are tested against zero, but the last example involves a (bogus) between subject factor f. -- O__ ---- Peter Dalgaard ?ster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk) FAX: (+45) 35327907
Hi Mary, it sounds like you have a split-plot design, or more gruesomely, a split-subject design. The model that I infer from your description of the design can be fit using the lme() function of the nlme() package, along the lines of a similar analysis documented in section 1.6 of Pinheiro and Bates (2000). It should also be possible by using the aov() function in base R. Cheers, Andrew On Wed, Jul 11, 2007 at 10:50:57AM +0100, Anderson, Mary-Jane wrote:> Dear R users, > I have a rather knotty analysis problem and I was hoping that > someone on this list would be able to help. I was advised to try this list > by a colleague who uses R but it is a statistical inquiry not about how to > use R. > In brief I have a 3x2 anova, 2 tasks under 3 conditions, within subjects. I > also took a variety of personality measures that might influence the results > under the different conditions. I had thought that an ancova would be the > best test, but it might be the case that this would not work with a within > subjects design. I have not found anything that explicitly states whether or > not it would, but all the examples I have read are between subjects design. > I also thought of investigating a manova, but it is not really the case > that I have more than one DV, it is the same DV in 6 different combinations > of task and condition. > There were 4 personality measures and I wanted to look at the degree to > which they affected the task/ condition interaction. > I have explained this briefly here, but I can of course provied more > details to anyone who can advise me further with this. > Thanks, > Mary-Jane Anderson > Information Analyst > Platform Project > Information Services Division, > NHS National Services Scotland, > Gyle Square, > 1 South Gyle Crescent, > Edinburgh, > EH12 9EB. > 0131 275 7163. > > > _________________________________________________________________ > NHS National Services Scotland Disclaimer > > The information contained in this message may be confidentia...{{dropped}} > > ______________________________________________ > R-help at stat.math.ethz.ch 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.-- Andrew Robinson Department of Mathematics and Statistics Tel: +61-3-8344-9763 University of Melbourne, VIC 3010 Australia Fax: +61-3-8344-4599 http://www.ms.unimelb.edu.au/~andrewpr http://blogs.mbs.edu/fishing-in-the-bay/