ABSTRACT
Youth develop in a nested world. For example, adolescents are part of classrooms that are nested in schools which are nested in neighborhoods which are nested in communities. This nested world calls for an integrative theoretical framework which the authors briefly describe. In addition, it poses two main problems for statistical analysis. First, if some participants share the same higher-order system (e.g., students in classrooms), their data can be statistically dependent such that any statistical test is biased. This problem can be easily resolved by adjusting the standard errors in the statistical tests. Second, higher-order units can have an impact on lower-order units, for example, the effects of the proportion of immigrants in a classroom on immigrants’ versus non-immigrants’ popularity among their classmates. In this case, multilevel regression models can be used. The authors provide a non-technical introduction into the two problems and their solutions, using examples from their research on immigrant youth adaptation.
