A comment on methodological concerns: A rejoinder

This rejoinder to Professor Cohen's Comment is intended to elaborate on approaches we utilized in our original paper. Some of our specification was necessitated by data availability (e.g., the specification of the sanction process). We attempt to clarify our approach in regard to the intervention process. We provide an explanation of the modeling that could underly a test of the hypothesis that the control effect is a product of the arrest process—a test that we had regarded as informative but peripheral to our main points. As noted, a significant part of that test utilizes the full range of control variables available to us. Finally, we enumerate what we felt to be the main points of our paper—points we are afraid might otherwise be lost in this discussion of methodology.