Visualizing holistic person-specific dynamic systems: a case example with self-efficacy and self-care behavior in emerging adults with type 1 diabetes

Common ways to test associations between two repeatedly measured constructs have two primary limitations. Studies often report the average effects and ignore the heterogeneity. Independently interpreted autoregression and cross-lagged coefficients (i.e. local effects) may not match the holistic dynamic patterns (i.e. considering all coefficients simultaneously). Our paper aims to address the limitations by introducing vector plots to visualize holistic person-specific dynamic patterns. We utilized a case example of 14-day daily diary data of diabetes self-efficacy and self-care from 200 emerging adults with type 1 diabetes. A dynamic structural equation model was used to generate person-specific coefficients. Vector plots and eigenvalues were generated to visualize person-specific holistic patterns. We found heterogeneity in both local and holistic dynamic patterns. Most participants (N = 178) had mismatching local and holistic patterns. Our study provided important evidence that failing to capture person-specific holistic dynamic patterns might result in incomplete interpretations of dynamic associations.