Quantitative Methods and Results From the Second Year of Alabama’s Math, Science, and Technology Initiative Study

• In Event: Evaluating Alabama’s Math, Science, and Technology Initiative: Results of a Three-Year, Statewide Randomized Experiment

Abstract

Although AMSTI is a two-year program, the confirmatory analyses address the effect of the program after the first year. The effect of AMSTI after the full intervention was implemented (that is, after two years) cannot be estimated without additional assumptions because the control group entered the program after its first year and was no longer a pure control group in the second year. Bell and Bradley (2008, 2009) developed a method to estimate the full effect of an intervention after the control group receives the intervention in a randomized controlled trial. Their method simulates the experimental framework that would have existed had the control group not received the intervention. The method provides an unbiased measure of the effect of the intervention in the second year if the impact of AMSTI during the first year of its implementation is stable over time—that is, if AMSTI affected the original control group schools the same way it affected treatment group schools a year earlier. This presentation will review the theory, assumptions, analytic methodology and results of the Bell-Bradley approach.

Authors

• Steve Bell, Abt Associates Inc.

• Andrew P. Jaciw, Empirical Education Inc.