Thomas Cook

There is a continuing and pressing need for valid causal research to support education decision-making. At the same time, it is often difficult, and sometimes impossible to conduct a randomized experiment to address causal questions in education. The purpose of this series of annual workshops was to train education researchers in the design and analysis of quasi-experimental studies. Specifically, these workshops provided training in four specific areas of quasi-experimental design and analy...

Well executed randomized experiments provide the strongest evidence about causal effects of educational interventions, products, and services. But randomized experiments may not always be feasible or successful. As a result, quasi-experimental designs are often used in education research for causal purposes. However, the quality of quasi-experimental designs and their application varies widely. In addition, knowledge of the better quasi-experimental designs that provide stronger evidence for...

This project extended work on improving four quasi-experimental methods that have potential for providing unbiased or minimally biased causal inference when random assignment is not possible. These four methods include: regression discontinuity designs, propensity score matching methods, short interrupted time-series designs, and pattern matching.

The purpose of this project was to improve the yield from cause-probing research in education. Considerable professional consensus already exists regarding which quasi-experimental designs represent acceptable alternatives when an experiment is not feasible; these include regression discontinuity, case matching with emphasis on propensity scores, short interrupted time series, and pattern matching. This project focused on improving research using these four design or analytic principles. Alt...