Using Computer Models to Predict Prevention Policy Outcomes

P policy by nature is prospective (i.e., it antici­ pates future events); research, however, is primarily retrospective (i.e., it analyzes past events). Public policy to reduce alcohol­involved problems is best formu­ lated when decisionmakers are aware of the potential cost and future effects of each prevention strategy alternative. Projecting policy outcomes is a difficult task, one that must be informed by the best research. The challenge of developing scientifically based prevention policy is par­ ticularly difficult at the local level, because most alcohol