Evidence-based policy prioritizes interventions that have been shown to be effective through rigorous experimental studies, with the expectation that these evidence-based programs will lead to better outcomes. Despite this expectation, local decision makers cannot know the impact an intervention will have in their jurisdiction and must rely on the available evidence—usually from other places—to make a best guess about whether they should adopt the intervention. Stuart and colleagues will leverage six national datasets consisting of multisite randomized controlled trials and use statistical modeling to simulate whether youth in each site would have benefited had local decision makers adopted the program based on the results from the other sites. The team will use a leave-one-site-out method to estimate how well decision makers can predict the impact of the intervention for a particular site based on evidence from the other sites in the multi-site evaluation. The team will also identify which pieces of evidence about the site and population served from the multi-site evaluations might lead to more reliable decision making and better youth outcomes. This project is among the first that the Foundation has funded to test the assumption that using high-quality research evidence leads to improvements in youth outcomes.
What is the potential for local decision makers to improve youth outcomes by adopting programs that have yielded positive effects in other sites?