Dyadic pattern analysis using longitudinal Actor-Partner Interdependence Model with Markov chains for unique case analysis

Abstract

Understanding the dynamics of interactions between two individuals requires special conceptual and statistical models. The Actor-Partner Interdependence Model (APIM) is the classical conceptual framework for standard dyadic designs, capturing the interdependence between dyad members by identifying the mechanisms of interaction through actor and partner effects. To analyze the temporal dynamics of dyadic interactions, the longitudinal APIM extends the classic model, often employing categorical variables to capture behavior. To analyze such data considering its categorical nature, specific statistical models are required. Markov chain is a powerful approach considering the longitudinal and categorical aspects of the data. This article describes how to adapt Markov chains in the categorical longitudinal dyadic case. It additionally offers a tutorial to model and identify the pattern of interaction using this method for unique case approach to maintain a simple and focused level of analysis. Codes in R language are provided.

Keywords

• dyadic intensive longitudinal categorical data
• actor-partner interdependence model
• markov chains
• r