Research That Matters (January 17 - 20, 2008)

Background and Purpose: Residential treatment programs deliver services to youths in group-based settings. By aggregating troubled youths, concerns have been expressed that exposure to deviant peers in group settings may increase youth problem behaviors. Peer contagion is a common criticism leveled at group-based interventions for troubled youths. Among youths in out-of-home care settings, about 20% are in group-based settings, living with other troubled youths. However, little is known about peer influences in this service setting. This study focused on three objectives: 1) to identify trajectories of youth externalizing behaviors during group care to assess evidence for increasing problem behaviors over time; 2) to examine characteristics of youths in each trajectory group, including the proportion of deviant peers in the group care living unit to develop profiles of youths who are most at risk for increasing problem behaviors and most susceptible to peer influence; and 3) to assess whether individual behavior trajectories during care were associated with youth outcomes at discharge from care as well as six-month follow-up.

Methods: Secondary administrative data from a large residential treatment program was utilized in this study (N=744). Data were collected at intake to group care, discharge, and six months after leaving care. Each day during placement, a youth's serious externalizing behaviors are documented by live-in staff. The frequency of serious externalizing behaviors each month in group care was the outcome of interest in the trajectory analysis. Latent class growth analyses using Mplus was conducted to develop externalizing behavior trajectories for youths during 12 months in group care. Multinomial logistic regression using SAS 9.1 was utilized to identify characteristics of youths and their proximal peer group to identify differences across trajectory group. In addition, individual trajectory group membership was used to predict discharge outcomes (favorable discharge, return home) as well as outcomes at the six month follow-up (legal/ criminal involvement, placement stability, and positive relationships with adults/caregivers).

Results: Results provided some support for peer contagion. Five trajectory groups emerged, with one group showing an increase in problem behaviors during time in care. The proportion of proximal peers with diagnosed disruptive behavior disorders was a consistent predictor of an individual youth's trajectory. In addition, younger White males were identified as the subgroup of youths most susceptible to peer contagion and increasing problem behaviors. Individual trajectory group membership was a strong predictor of three outcomes (favorable discharge, return home, and placement stability).

Conclusions and Implications: Using innovative methods, this study identified evidence for peer contagion in group care settings. While most youths in the study maintained low or decreasing problem behaviors during residential treatment, a subgroup of youths did not fare as well. This study demonstrated that it may be possible to predict peer contagion and therefore contain or prevent it. Further, the association between behavior trajectory during care and subsequent outcomes underscores the importance of youth experiences in out-of-home placement.