Using model parameters describing affective dynamics to predict romantic relationship dissolution.

In a previous paper demonstrating the utility of intraindividual process dynamics in predicting relationship dissolution Castro-Schilo and Ferrer (2013) applied dynamic factor analysis (DFA) models to intensive multivariate data from romantic dyads. This paper helped to demonstrate the utility of using intraindividual process data to predict future, intraindividual outcomes. We revisit their example, highlighting conditions when DFA models may not be appropriate for such applications due to insufficient measurement invariance to justify comparison of intraindividual dynamics of the latent variables at the interindividual level. We then propose that the application of intraindividual network dynamics models can be employed as an alternative approach to capture multivariate dynamics while avoiding the problem of measurement invariance by focusing on manifest, rather than latent, variable dynamics. Results are framed in the context of an empirical example of romantic dyad dynamics predicting future dissolution of romantic relationships, and the broader application of incorporating intraindividual dynamics in interindividual analyses. (PsycINFO Database Record (c) 2018 APA, all rights reserved)