When different types of policy interventions are available, there is an understandable desire to evaluate all alternatives using common metrics so scare resources can be allocated in the most efficient manner. However, systems that display significant lags in their response to some interventions can confound such an empirical approach. This paper provides a parsimonious mathematical representation of some of the challenges confronted when trying to evaluate upstream interventions on lagged systems to help clarify when it is and when it is not practical to expect those interventions to meet the same standard of proof as downstream interventions. Implications for drug policy and delinquency prevention are elaborated.
Adobe .pdf files require the Adobe Acrobat Reader.
Download and install the Adobe Acrobat Reader.