R help - Mar 2010 - transposing data

Hi. I have repeated measures data of the form where each observation  
is a trial, and trials are grouped by subject, and variables encode  
whatever level of a factor was present during that trial, and the  
dependent variable is response time (RT). I want to transpose the  
data to a form suitable for MANOVA such that there is one observation  
per subject and RT is recoded across many variables, the totality of  
which represents all the combinations of all the levels of all the  
independent variables.

For example, exp1.r5 is my data frame, Subject is the grouping  
variable, I want to go from long to wide, the dependent variable is  
RT, and I want to get RT as observed at each combination of each  
level of all the variables listed in the timevar argument:
exp1.r5.wide <- reshape(exp1.r5, idvar="Subject", direction =
"wide",
sep = "", v.names="RT", timevar = c("CannonAngle",
"CannonOriB",
"CannonOriR", "nRedPellets", "TargetColor",
"tbearing"))

Instead what I get is Subject plus six variables that are some sort  
of conjunctions of all the levels of each of the variables I want.  
Instead what I want is Subject plus up to 1,008 variables encoding RT  
at every combination of CannonAngle, CannonOriB, CannonOriR,  
nRedPelelts, TargetColor, and tbearing that occurred. Is that  
something the reshape function can give me or should I be looking  
elsewhere?

Sincerely,

Frank Tamborello, PhD
W. M. Keck Postdoctoral Fellow
School of Health Information Sciences
University of Texas Health Science Center, Houston


	[[alternative HTML version deleted]]

Frank -
    I think you need to create a composite time variable
to do what you want to do:

exp1.r5$key = with(exp1.r5,paste(CannonAngle,CannonOriB,
                   CannonOriR,nRedPelelts,TargetColor,tbearing,sep='.'))
exp1.r5.use = subset(exp1.r5,select=-c(CannonAngle,CannonOriB,
                   CannonOriR,nRedPelelts,TargetColor,tbearing))
exp1.r5.wide <- reshape(exp1.r5.use, idvar="Subject", direction =
"wide",
        v.names="RT", timevar = 'key')

 					- Phil Spector
 					 Statistical Computing Facility
 					 Department of Statistics
 					 UC Berkeley
 					 spector at stat.berkeley.edu





On Fri, 5 Mar 2010, Frank Tamborello wrote:
> Hi. I have repeated measures data of the form where each observation
> is a trial, and trials are grouped by subject, and variables encode
> whatever level of a factor was present during that trial, and the
> dependent variable is response time (RT). I want to transpose the
> data to a form suitable for MANOVA such that there is one observation
> per subject and RT is recoded across many variables, the totality of
> which represents all the combinations of all the levels of all the
> independent variables.
>
> For example, exp1.r5 is my data frame, Subject is the grouping
> variable, I want to go from long to wide, the dependent variable is
> RT, and I want to get RT as observed at each combination of each
> level of all the variables listed in the timevar argument:
> exp1.r5.wide <- reshape(exp1.r5, idvar="Subject", direction =
"wide",
> sep = "", v.names="RT", timevar =
c("CannonAngle", "CannonOriB",
> "CannonOriR", "nRedPellets", "TargetColor",
"tbearing"))
>
> Instead what I get is Subject plus six variables that are some sort
> of conjunctions of all the levels of each of the variables I want.
> Instead what I want is Subject plus up to 1,008 variables encoding RT
> at every combination of CannonAngle, CannonOriB, CannonOriR,
> nRedPelelts, TargetColor, and tbearing that occurred. Is that
> something the reshape function can give me or should I be looking
> elsewhere?
>
> Sincerely,
>
> Frank Tamborello, PhD
> W. M. Keck Postdoctoral Fellow
> School of Health Information Sciences
> University of Texas Health Science Center, Houston
>
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> 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.
>