[R] intra-class correlation? coherence among multiple ordinal responses

Daniel Malter daniel at umd.edu
Sat Aug 22 00:15:09 CEST 2009

I have a quick statistical question and hoped somebody has a tip for me
without me having to go to the local statistician on Monday. 

I assess 4 statements from 90 subjects. Each of the 4 statements receives
one of three responses (say -1, 0, or 1). I can use Cramer's V or Spearman
correlations to assess the correlation between each pair of statements, but
I am looking for a measure of coherence across/among all 4 statements. To
me, this seems to be intraclass correlation. So I was wondering whether I
can apply it here (specifically, ICC(2,1) or ICC(3,1)) or whether I should
use an alternative procedure given that my responses are ordinal.

I am grateful for your tips,

"Who has visions should see a doctor," 
Helmut Schmidt, German Chancellor (1974-1982).

More information about the R-help mailing list