A0914
Title: Comparing ordinal Markov random fields in two independent samples with Bayesian model selection
Authors: Maarten Marsman - University of Amsterdam (Netherlands) [presenting]
Abstract: In psychological network analysis, data are often assessed on an ordinal scale (e.g., variables assessed on a Likert scale). An ordinal Markov random field graphical model has been proposed recently to adequately analyze the network structure underlying ordinal data. Bayesian edge selection methodology is used in combination with Bayesian model averaging to assess the evidence for the inclusion or exclusion of individual edges in the estimated network structure. However, quite often, interest goes out to assess if and how the estimated network structure differs across groups (e.g., based on gender or clinical status). The ordinal Markov random field model is extended to two-group designs, and the Bayesian methodology is used to assess differences in the estimated network structure across groups.
