Session Summary: Peter H. Rossi Award Lecture
November 13, 2012 09:31 AM
The Peter H. Rossi Award was presented to Thomas D. Cook for his lifetime of achievement in program evaluation. Cook, a professor at Northwestern and a faculty fellow in the Institute for Policy Research, has received numerous awards over the years for both his theoretical work and his practice of program evaluation. Introducing and presenting the award, Douglas Besharov stressed Cook’s penchant for taking on the tough questions that required looking for new techniques, often quasi-experimental to make causal inferences. Also worth noting is Cook’s attention to disadvantaged populations in much of his research.
For his lecture, Cook discussed the Interrupted Time Series (ITS) method, which has seen a resurgence in use in large part to Cook’s work. ITS often uses administrative data and in its simplest form compares the differences in means or slopes between pre-intervention time series data and post-intervention time series data. Its logic is akin to that of the differences in differences method and its primary threat to validity is a treatment that affects one group more than another. Cook discussed more complicated versions of the method that are more robust; he also gave conditions under which ITS is best used.
Concluding his talk, Cook said that we should expect to see more ITS. While the Randomized Control Trial (RCT) method remains the favored approach, there are many cases in which an RCT is simply not feasible.
Three discussants, Judith Gueron, Robinson Hollister, and Rebecca Maynard, followed Cook’s talk. All expressed their admiration for him and his work. Some caution was voiced by Hollister because RCT is already a hard sell and he worried about settling for quasi-experimental methods. Another sentiment expressed was that it would be unfortunate if a wave of graduate students began using ITS as a method looking for problems. Overall though, there was optimism for ITS playing a larger role in policy research, particularly in plugging the holes that other methods have been unable to fill.
Contributed by Brent Gibbons, University of Maryland