Communities attempting to prevent delinquency, teenage pregnancy, school drop-out and substance misuse frequently find that there are few models of broad community-level planning and action that have been proved to be effective. This might indeed explain why, although there are numerous evidence-based programs (EBPs) addressing such issues, their penetration is still low.

Communities that Care (CTC) is one of the best-known models for developing community partnerships to promote the adoption of EBPs, and then support them once they are up and running. Developed by David Hawkins and Richard Catalano from the University of Washington, Seattle, nearly 20 years ago, it has been found in a recent randomized controlled trial to reduce risk factors and enhance protective factors for adolescent behavior problems.

But while that study was based on only 12 communities where CTC was running, and involved intensive support from researchers, new research indicates the model’s effectiveness when implemented on a large scale without this additional help. “Community coalitions can affect adolescent risk and protective behaviors at a population level when EBPs are utilized. CTC represents an effective model for disseminating such programs,” conclude Mark Feinberg and a team from the Prevention Research Center at Penn State University in an article in Prevention Science.

Their research made use of the implementation of CTC in 120 communities in Pennsylvania over a decade. They used data collected through anonymous student surveys conducted in schools in 2001, 2003 and 2005.

The Pennsylvania Youth Survey, conducted by the state’s Commission on Crime and Delinquency, used a well-validated, student self-report measure that covers a range of risk and protective factors for adolescent problem behaviors. These include family relations, attitudes towards school and social and physical features of the child’s neighborhood. The sample included over 59,000 young people from grades 6, 8, 10 and 12.

The research team examined changes in groups of individuals over time using multilevel models, comparing students for CTC and non-CTC communities.

They found that there was less growth in delinquency among youth in CTC communities than youth in non-CTC communities. Levels of risk factors like anti-social attitudes increased more slowly, and protective factors (like community and family cohesion) and academic performance decreased more slowly, among young people in CTC communities than those in the communities with which they were being compared. Furthermore, there was one-third less annual decline in the level of academic performance for youth exposed to EBPs through CTC, and the normative increase in delinquent behavior fell by 11 per cent a year.

The pattern of effect sizes, which were small to moderate, was consistent with the focus of EBPs implemented in CTC communities, most of which are universal adolescent programs delivered through schools. So, CTC exerted a stronger impact when measured in terms of youh’s commitment to school and their opportunities to get involved in class discussions and other activities at school, and a weaker effect when viewed in terms of the availability of drugs and firearms in the community.

The authors argue that the value of the findings from this quasi-experimental study is enhanced by the fact that it looked at a naturally occurring dissemination of the model rather than a research-initiated or demonstration project. They also suggest that the research removes the possibility that selection bias is responsible for results because each community served as its own control.

As for why there was no discernible impact on youth-reported substance use, the research team suggest that more positive results might have emerged from a less conservative design. For example, it was not possible to identify adolescents who had recently moved into CTC communities who would by definition not have been affected by programs implemented in previous years.

The study’s other limitations, acknowledged by the researchers, include the predominantly rural sample, which means that results cannot be generalized to large urban settings, and the reliance on self-report data.

Reference
Feinberg, M. E., Jones, D., Greenberg, M. T., Osgood, D. W. & Bontempo, D. (2010) ‘Effects of the Communities that Care model in Pennsylvania on change in adolescent risk and problem behaviors’, Prevention Science 11: 163-171.

Hawkins, J. D., Brown, E. C., Oesterle, S., Arthur, M. W., Abbott, R. D. & Catalano, R. F. (2007) ‘Early effects of Communities that Care on targeted risks and initiation of delinquent behavior and substance use’, Journal of Adolescent Health 43 (June): 15-22.