A descriptive study was done of a microcomputer-based modeling approach to retention analysis used at the University of Hartford (Connecticut) which relied on an information system which integrates multi-year admissions, financial aid, and academic data with surveys of college student adjustment, involvement, and reasons for leaving. Analyses were performed with an interactive decision support system that enabled users to test hypotheses about high and low risk student subpopulations using a simple system of menus and screens. An executive information system allowed decisionmakers to navigate their way through retention reports and graphs in order to achieve a better understanding of comparative and longitudinal aspects of retention. Using this system, longitudinal and comparative studies of retention were prepared from a multi-year database for all student cohorts entering the University of Hartford between 1984 and 1989. Separate analyses of attendance patterns and degree attainment were prepared for selected student subpopulations and for freshmen and transfers in each of the University's eight colleges. Results of the analyses for retention research are presented for academic prediction studies, student adaptation to college, reasons for leaving, and student involvement. Included are 6 figures and 12 references. (JB)