This article analyzes the "too few cases/too many variables" problem, suggests a research agenda for a third generation of implementation researchers, and outlines several research designs, including ...