[R] Bootstrapping Krippendorff's alpha coefficient

[Jim Lemon]

On 10/12/2010 08:58 PM, Łukasz Ręcławowicz wrote:
> Hi,
>
> I don't know how to sample such data, it can't be done by row sampling
> as default method on matrix in boot.
> Function  takes matrix and returns single coefficient.
>
> #There is a macro but I want use R :)
> http://www.comm.ohio-state.edu/ahayes/SPSS%20programs/kalphav2_1.SPS
> library(concord)
> library(boot)
> # The data are rates among observers with NA's
> nmm<-matrix(c(1,1,NA,1,2,2,3,2,3,3,3,3,3,3,3,3,2,2,2,2,1,2,3,4,4,4,4,4,
> +  1,1,2,1,2,2,2,2,NA,5,5,5,NA,NA,1,1,NA,NA,3,NA),nrow=4)
>
> sample.rates<-function(matrix.data,i){
> #mixed.rates<-sample individual rates and put back in new matrix (?)
> return(kripp.alpha(mixed.rates)$statistic[i])
> }
> to.get<-boot(nmm, sample.rates, R=1e4, stype="i")
>
Hi Lukasz,
First, switch to the kripp.alpha function in the irr package. concord is 
no longer maintained. The SPSS code would take some time to decipher and 
translate into R, so I'll see if I can locate the algorithm. Professor 
Krippendorff once wrote to me how he did it, so it must be available 
somewhere.

Jim