What We Don’t Know about Evidence-based Programs
In recent years both government and private funders have called for greater use of evidence-based programs and practices to serve children and youth. This is generally a good thing. We should fund programs that have at least some potential to improve the lives of children and youth. The good news is that there is a significant—and growing–number of programs with demonstrated positive outcomes. For example, Child Trends maintains a “What Works” database of more than 600 experimentally evaluated programs, which includes programs that have demonstrated positive, mixed, and negative impacts on children’s and youth’s outcomes.
The bad news is that it is far from clear what distinguishes effective from ineffective programs. In other words, we don’t know much about why programs work or don’t. In 2007, the late Douglas Kirby wrote Emerging Answers: Research Findings on Programs to Reduce Teen Pregnancy and Sexually Transmitted Diseases. In this update from his 1997 report, he emphasized how much the number of rigorous impact studies had increased over ten years. However, throughout the report he noted how little we know about program implementation. Indeed, one of Kirby’s recommendations for the field is: “Provide much more complete program descriptions in published articles, as well as more informative process evaluations, to help reviewers ascertain why some programs were effective and others were not.”
Unfortunately, although progress continues to be made in terms of impact studies, relatively little progress has been made in studying program implementation across a variety of programs, including teen pregnancy prevention programs. For example, Dr. Mary Terzian, a colleague at Child Trends, has been reviewing teen pregnancy prevention programs that work for Latino adolescents (forthcoming). She found very little available information on implementation in the evaluations. Recently, I was reviewing a comprehensive list of reproductive health programs from Child Trends’ database of programs that work—and don’t –to improve children’s and youth’s outcomes. In reviewing those studies it was clear that, although there are a few clear conclusions about what works and what doesn’t work, many strategies work in one program but not in others. Similarly, there are both short-term “light-touch” and long-term intensive programs that have evidence of effectiveness, but there are also both short- and long-term programs that do not work. Without detailed information about program operations, disentangling features that matter greatly from those that don’t is not possible.
Interestingly, current replications of evidence-based programs place a priority on ensuring that the programs are being implemented with fidelity to the program model. This is done to help improve the chances that the program effects can be replicated in other settings. This is important, but we have missed an important step: If we don’t understand what it is about the program that made it effective in the first place, then it is challenging to replicate the effects that made the program desirable in the first place. And the evidence is clear: program replications tend to show smaller effects. A better understanding of program implementation will improve the chances for successful program replications.