0

I'm looking for guidance on how to calculate the 95% confidence intervals for Krippendorf's alpha coefficients in R. I've already calculated the coefficients: 

key2 <- data.frame("rater1"=as.vector(t(rater1[2,])), "rater2"=as.vector(t(rater2[2,])))

key2 <- t(as.matrix(key2))
kripp.alpha(key2, method="ordinal")

Results: n=30, 2 raters, Alpha = 0.767

Previous forums suggesting conventional bootstrapping techniques don't seem to be the right course of action:

Not sure if this is also helpful: http://web.asc.upenn.edu/usr/krippendorff/boot.c-Alpha.pdf

Thomas
  • 43,637
  • 12
  • 109
  • 140
Erika L
  • 1

1 Answers1

0

Welcome newbie! A couple pointers:

1) Always include a reproducible example, if you can. You can set a seed with set.seed(), then simulate some data, and post all your code here.
2) StackOverflow is more for coding questions, the companion site CrossValidated is for the theory questions - e.g. SO is where you'd ask how to do bootstrapping for this, and CV is where you'd ask if bootstrapping is appropriate.

I'd say bootstrapping is called for, Krippendorf actually wrote a short paper about it: http://web.asc.upenn.edu/usr/krippendorff/boot.c-Alpha.pdf

There's also this page for some implementations in R, if you haven't come across it already: http://agreestat.com/r_functions.html

kng229
  • 473
  • 5
  • 13