In this study, simulation models were developed to explore the effects of decentralization and interdependence on individuals working in research groups. We conducted the study by simulating clinical health-services research with large repositories of data. First, we modeled the standard research process, which we called "centralized activity," in which a highly-paid researcher accomplishes all tasks. In our second model, which we called "distributed activity," tasks were performed in parallel by specialized technicians at lower wage scales. The results showed that centralized activity is slower (21.5 weeks vs. 17.7 weeks) and more expensive ($69,700 final cost vs. $34,600) than distributed activity, and places too great a load on the lone researcher. We concluded that while the distributed model may initially seem more complex, it results in shorter completion times and lower overall costs.