This panel will explore evaluation designs that use data to predict the future. At one end of the scale are big data methods, which rely on the skillful mining of large databases to maximize the chances of identifying potentially relevant information. At the other end are minimalist methods, which seek the smallest set of data that provide the best prediction of target outcomes. Both may be cheap and fast, but for different reasons. Neither is simple to do well. The panelists describe their work at opposite ends of the scale and provide examples of how predictive methods can strengthen evaluation designs.
Director, Measurement, Evaluation and Organizational Performance,
The Rockefeller Foundation