Dynamic clustering: Classifying people through ecological momentary assessment

Time series analysis gained large popularity in psychological research. Adequate psychological time series models describe and predict individual dynamics, while maintaining the generalisability of dynamics to populations of individuals. The currently available approaches fail to cover both: It either focuses on each single individual separately –losing the population of individuals – or on all individuals jointly –losing the unique individual dynamics. We propose to bridge this methodological gap via dynamic cluster modelling. Its value for psychology will be demonstrated by answering two clinically relevant leading questions. Software will be made available for an easy application of the methodology developed.