We describe a sequential qualitative ➔ quantitative mixed-method procedure used to construct conceptually grounded quantitative metrics of interpersonal behavior from continuous spatiotemporal data. Metrics were developed from data collected during an experiment in which racially diverse participants interacted with self-resembling avatars at social events hosted in the virtual world Second Life. In the qualitative stage, the researchers conceptualized four distinct patterns of movement from overhead video recreations of participants interacting during the social events. In the quantitative stage, these patterns of movement were operationalized into metrics to reflect each type of observed interpersonal behavior. The metrics were normalized through a series of transformations, and construct validity was assessed through correlations with self-report measures of intergroup behavior. Finally, the metrics were applied to an analysis of the virtual-world study examining the influence of resource competition on racial group interactions. The findings contribute to our understanding of the influence of resource competition on Blacks', Asians', and Whites' group dynamics. Applications of these metrics for the future of the psychological study of interpersonal behavior are discussed.
Keywords: Dynamic spatiotemporal data; Group dynamics; Mixed methods; Race.