Tackling Longitudinal Round-Robin Data: A Social Relations Growth Model

The social relations model (SRM) is commonly used in the analysis of interpersonal judgments and behaviors that arise in groups. The SRM was developed only for use with cross-sectional data. Here, we introduce an extension of the SRM to longitudinal data. The social relations growth model represents a person’s repeated SRM judgments of another person as a function of time. We show how the model’s parameters can be estimated using restricted maximum likelihood, and how the effects of covariates on interindividual and interdyad variability in growth can be computed. An example is presented to illustrate the suggested approach. We also present the results of a small simulation study showing the suitability of the social relations growth model for the analysis of longitudinal SRM data. The SRM is a mathematical model that aims to disentangle important components of interpersonal judgments and interpersonal behaviors (Kenny, 1994; Kenny, Kashy, & Cook, 2006). It is used in almost all psychological disciplines including personality, social, developmental, educational, organizational, clinical, and cognitive psychology. Personality and social psychologists, for example, examine whether interpersonal attraction between unacquainted individuals is affected by their personality traits (e.g., Küfner, Nestler, & Back, 2012). Educational psychologists investigate the performance of pupils in cooperative learning groups (e.g., Horn, Collier, Oxford, Bond, & Dansereau, 1998), and developmental psychologists study the complex interdependencies in families (Branje, Finkenauer, & Meeus, 2008). Organizational psychologists investigate teamwork perceptions in working groups (e.g., LeDoux, Gorman, & Woehr, 2012), clinical psychologists examine how participants in a group psychotherapy interrelate and influence each other (e.g., Marcus & Kashy, 1995), and finally, cognitive psychologists study how individual’s social memory is influenced by being the memory agent and the memory object (Bond, Dorsky, & Kenny, 1992). In all these cases, researchers assess groups of individuals such that every member of the group judges and is judged by all other group members. For example, groups of individuals may be asked to interact with each other and to judge each other on liking, performance, or cooperativeness. The resulting data are called round-robin data, and the SRM can be applied to such data. Traditionally, the SRM has been used to analyze cross-sectional round-robin data (a list of articles can be downloaded from David Kenny’s homepage: http://david-kenny.com/srm/srm.htm). However, researchers are becoming increasingly interested in how interpersonal judgments and behaviors change over time. Personality psychologists, for example, might want to know how liking changes when individuals become acquainted with each other, and how these changes can be explained (see e.g., Leckelt, Küfner, Nestler, & Back, 2015). There is hence a growing interest in how to analyze longitudinal round-robin data (Nestler, Grimm, & Schönbrodt, 2015). Here, we extend the SRM to longitudinal data by applying the idea of modeling individual growth trajectories to the SRM. Our approach allows researchers to examine questions of stability and change in round-robin data (e.g., what component of the interpersonal judgment or behavior changes across time) and to examine which variables affect these changes. The present article is organized as follows: We start with a brief description of the round-robin design, the SRM, and the estimation of SRM effects. This is followed by a discussion of earlier treatments of longitudinal round-robin data and a description of the social relations growth model (SRGM). We then outline how the parameters of the suggested growth model can be estimated using a restricted maximum likelihood approach, and how the effect of covariates can be examined. Thereafter, we present an illustrative example and finally, report the results of a small simulation study.