Abstract for:Understanding Time-varying Correlation in Dynamic Systems
We explore the use of correlation in the context of system dynamics models. In particular, we use time-varying correlation to illustrate aspects of the dynamics of systems that are not evident using other mechanisms. In a study of performance measurement dynamics, we find that the correlation between the true goal of a system and its associated metric for performance varies over time. The identification of the usability of time-varying correlation coefficients and associated underlying causal mechanisms has potential implications for social scientists and researchers. Areas such as health, corporate policy, security, and any type of intervention-based study may benefit from the insights derived from this paper.