Lisa Wang wrote: > I would like to test the agreement amongst 6 raters for nominal data on > a scale from 1-4, and conduct a latent class analysis in R. How should > the data be formatted and what code should I use? This kind of agreement test seems to call for Krippendorf's alpha. See http://www.asc.upenn.edu/usr/krippendorff/webreliability2.pdf