Session: Challenges in Multilevel Latent Class Analyses: Implications in School Based Prevention Science (Society for Prevention Research 25th Annual Meeting)

Session Introduction:  Increasingly, schools are becoming a primary location for prevention and intervention programming. This presents many opportunities for access to participants, but also presents a number of methodological challenges in terms of appropriately modeling individual and contextual influences on behavior and other outcomes. The goal of this symposium is to highlight challenges associated with multi-level latent class analysis and offer a modeling solution flexible enough to be applied in many contexts. This symposium represents the conference theme of innovative methods and statistics

The first paper, “Class enumeration challenges with nested data” uses simulated data of a school based study in which multiple schools provide individual level data on student behavior. This simulated data will demonstrate the biases introduced when not taking into consideration the multilevel nature of the data. Major strengths of this paper include the focus on the key challenges many school-based researchers in large intervention trials face, while presenting a gold standard methodology for class enumeration with multilevel data.

The second paper, “School level influences on individual patterns of behavior: Taking advantage of a multi-cohort, multi-site intervention study,” uses data from the Multisite Violence Prevention Project, which includes data from over 5,000 students from four locations in the United States across two cohorts. The goal of this study is to characterize student level behavior and model the impact of school climate on those behaviors. Challenges to utilizing this data set include multiple cohorts within the same schools, school climate measured at the student level as well as the school level and finally data captured from multiple sites, which may introduce additional variance that needs to be accounted for.

The third paper, “Understanding the classroom context and its influence on intervention implementation,” uses data from a school-based intervention randomized control trial testing the Good Behavior Game (GBG) versus an integration of GBG with a socioemotional skills curriculum (PATHS). The goal of this study is to determine if classroom level contexts, as defined by student behavior and teacher characteristics (e.g., teacher burn-out), influence intervention implementation. Challenges to utilizing this dataset include multiple informants of classroom context and a limited number of classrooms, which will potentially limit model specification.

Finally, a discussant will highlight commonalities among the papers, discuss implications for prevention, and moderate a discussion between the presenters and the audience.