ABSTRACT
One approach for generating generalizable lessons about "what works" was to conduct statistical meta-analytic studies of collections of local evaluations. This approach allows statistical assessment of relative program effectiveness by developing effect sizes for evaluations that have comparable outcome measures, meet methods criteria, and have generally similar intervention approaches. As a study more concerned with generating generalizable knowledge to improve interventions for a public health concern, the High Risk Youth Programs (HRY) study generated less intense or contentious relations with stakeholders than more specific studies that relate more clearly to specific actions and interests. The HRY evaluation generated interest and support from policy research and public health professionals, but there were no strong interests immediately at stake. Multisite evaluations are not unusual for large demonstration grant programs intended to "seed" interventions to address education, behavioral health, or other social problems.
