Abstract: | [Correction Notice: An erratum for this article was reported in Vol 98(3) of Psychological Bulletin (see record 2009-20912-001). An unnecessary σ was added to the second line of Equation 32 on page 555. The second line should read: + 1/???](1 + ?a?b)/(1 – ?a?b), with the rest of the equation correct as printed.] Presents a model for the analysis of dichotomous data generated under a Markov chain model. Methods for estimating, testing, and comparing proportions and for analyzing dyadic interactions are described and compared with the standard techniques derived under the assumption of serial independence. It is shown that the regular test statistics must be adjusted by a correction factor that is a simple function of the within-S serial dependence parameter. Eigenvalues of the transition matrix are appended. (39 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved) |