Comparing sequential associations within a single dyad

We present a new application of sampled permutation testing to examine whet her two sequential associations are different within a single dyad (e.g., a teacher and a student). A Monte Carlo simulation with the same (i.e., 100 vs. 100) or a different (100 vs. 400) number of event pairs was used to sim ulate designs that. use time-based (typically producing equal-length compar isons) and event-based (typically producing different-length comparisons) d ata, respectively. For these pairs of simulated data streams, we compared t he Type I error rates and the kappa for agreement on significance decisions , using the sampled permutation tests and the more traditional asymptotic l og linear analysis. The results provide the first evidence relevant to eval uating the accuracy of log linear analysis and sampled permutation testing for the purpose of comparing sequential associations within a single dyad.