Examining Rural Adolescents' Risk and Protective Profiles to Inform Youth Violence Prevention

Background/Purpose:  Violence prevention science and positive youth development fields suggest youth who experience a particular constellation of protections and risks across multiple domains likely benefit from different prevention strategies than youth who experience another constellation. However, little research has been done on risk and protective factors for early adolescents in rural settings; further, existing studies generally do not have diverse samples and often use a deficit orientation or cumulative risk approach. As a result, the variation and heterogeneity in strengths and risks has been neglected. Person-oriented research has demonstrated that latent profile analytic procedures can identify unique combinations of strengths and risks that capture this variation and also improve prediction of developmental outcomes. The current study expands existing literature by using a person-centered approach to identify unique constellations of risk/protection among racially/ethnically diverse, rural early adolescents. In addition, the study provides important stakeholders such as school social workers and practitioners with a typology to guide violence prevention efforts in ways that recognize both risks and strengths.

Methods: A diverse sample of rural 6th-8th graders (N=4,321) responded to items from the School Success Profile. Using Mplus6, Latent Profile Analysis was used to classify students into distinct subgroups based on seven continuous indicator variables representing risk and protective factors across various domains (Neighborhood; Peers; Family; Individual). Using a random subsample (n=2,160), solutions specifying various numbers of groups were run iteratively and compared based on substantive meaningfulness, Bayesian Information Criteria, Lo-Mendell-Rubin, entropy, posterior probabilities, and class size. Gender and race/ethnicity were examined as possible covariates. The final model was replicated with a second random subsample (n=2,161), and validated based on group differences on substantively and empirically related variables (e.g., school conduct/behavior) using one-way ANOVAs with Bonferroni post hoc comparisons in SPSS 17.0.

Results: Findings from the study suggest a typology with four distinct subgroups of students. Profile 1 (19%) was distinguished by high emotional distress and parent/peer conflict. Profile 2 (6%) was characterized by moderate problems related to neighborhood safety and delinquent friends. Profile 3 (72%) was characterized by low risk/broad protective experiences across multiple domains. Profile 4 (3%) was characterized by substantial problems related to neighborhood safety and delinquent friends. Gender significantly predicted group membership; race/ethnicity did not. ANOVAs revealed significant group differences in perceived support and school conduct/behavior.

Conclusions/Implications: Study results offer preliminary evidence suggesting four latent profiles of rural, early adolescents’ risk/protective experiences in domains relevant to youth violence.  Results offer a nuanced understanding of the experiences of early adolescents – a key stakeholder in youth violence prevention efforts. Results can inform allocation of limited resources by guiding selection of potential solutions to more directly target adolescents’ specific risk and protective profiles. For example, our results suggest students in Profile 1 may benefit from interventions targeting emotional distress, interpersonal skills, and family conflict. Practice implications include recommendations to embed specific strategies within universal violence prevention programs to capitalize upon specific strengths (e.g., parent support) within the context of specific risks (e.g., neighborhood safety; negative peer behavior) experienced by distinct subgroups.