A new paper in Prevention Science using data from the Community Youth Development Study (Rhew et al. 2015) did not find effects of CTC on reducing problem behaviors in repeated cross-sectional samples of 6th, 8th and 10th grade students. Although drug use and delinquency generally decreased in prevalence in CTC communities for each of these grades over the course of the study, the level of decrease was not significantly different from control communities.These results are in contrast to earlier findings showing significant reductions in levels of problem behaviors in CTC compared to control communities when using a different study design that followed a longitudinal panel of youth starting from 5th grade (Hawkins et al. 2008).
While disappointing, these contrasting findings were not entirely surprising to the research team. We recognized at the outset that the longitudinal panel design would have greater ability to detect CTC effects than the repeated cross-sectional analyses of all students in a particular grade every two years. One reason for this is that collecting data from the same individuals over time allows analyses to control for individual characteristics and differences at baseline. The repeated cross-sectional samples, on the other hand, include different individuals at each wave and the composition of the samples may differ according to some important ways that may mask effects. The longitudinal panel may also be better suited to detect effects because the panel experienced very little attrition whereby youth who were exposed to CTC are not followed up at later assessments. Nor was it possibly biased by accretion whereby the inclusion of young people in the sample who are new to the community and thus not exposed or minimally exposed would offset beneficial effects among those who received greater exposure to CTC.
In summary, while CTC had significant effects on reducing alcohol use, tobacco use and delinquent and violent behavior in the longitudinal panel through grades 8, 10 and 12, the repeated cross-sectional grade-level samples did not show similar significant effects likely due to methodologic limitations of the study design. We still recommend that communities collect cross-sectional data to track their progress over time related to their prioritized risk and protective factors and prioritized outcomes. Communities implementing CTC in the trial were able to demonstrate progress toward their goals and demonstrated steeper decreases in their communities than the state-wide data overall. Using cross-sectional data for community monitoring is still recommended.
Hawkins, J. David, Catalano, Richard F., Arthur, Michael W., Egan, Elizabeth (2008). “Testing Communities That Care: The rationale, design and behavioral baseline equivalence of the Community Youth Development Study.” Prevention Science, 9(3), 178-190.
Rhew, Isaac C., Hawkins, J. David, Murray, David M., Fagan, Abigail A., Oesterle, Sabrina, Abbott, Robert D., Catalano, Richard F. (in press). “Evaluation of community-level effects of Communities That Care on adolescent drug use and delinquency using a repeated cross-sectional design.” Prevention Science.