[R] basic handling of data.frame

[hadley wickham]

On 3/27/07, Henning Meyer <tutmann at gmail.com> wrote:
> Hello,
>
> I'm new to R but wan't to use it to compute the statistics of my medical study.
> The study includes several parameters for a number of patients. Each
> parameter was assessed by a number of readers, once with a special
> condition, once without.
> Now I have a data.frame with colums like:
>
> PatientID, ReaderID, SpecialCond(yes/no), Parameter1, Parameter2.....
>
> the rows are not sorted, and the table is not complete (e.g. some
> readers didn't read all cases under all conditions).
>
> What I would like to do now is for example compute Cohen's Kappa
> (ckappa from package(psy)) for Parameter1 under special condition
> against no special condition.
> Therefore I need a table with colums:
> Parameter1WithSpecial, Parameter1WithoutSpecial
> And the data should be matched by PatID and ReaderID. How do I get
> this kind of table?
>
> Also how would I compute Kappa for Reader1 against Reader2?

Have a look at the reshape package, http://had.co.nz/reshape.  If your
data.frame is called df, then code something like the following should
get the data in the form that you want

dfm <- melt(df, id=c("ReaderID", "PatientID", "SpecialCond"))
cast(dfm, ... ~ variable + SpecialCond)

Regards,

Hadley