其他摘要：Policies that call for members of the public to change their behavior fail if people don't change; predictions of whether the requisite changes will take place are needed prior to implementation. I propose to solve the prediction problem with Factorial Forecasting, a version of functional measurement methodology that employs group designs. Aspects of the proposed new policy are factorially manipulated within scenarios, and respondents typical of those whose behavior would need to change are asked to project how they would react. Because it is impractical to validate the projections by seeing if they correspond to what eventually happens, I advocate evaluating validity by invoking a coherence criterion.
