Longitudinal data from RAND's Health Insurance Experiment were used to test the hypothesis that provider continuity can be modeled as one behavioral consequence of patient satisfaction. Bivariate and multivariate analysis (controlling for sociodemogr...
