To determine whether any of the structural paths, indirect effects, and covariances varied as a function of gender, race/ethnicity, or living arrangement, we employed a multiple group model approach. We first permitted paths to vary across gender (then race/ethnicity, and then living arrangement), and then constrained them to be equal across gender. Paths did not vary either by gender [Δχ2(35) = 26.256, p > .05] or race/ethnicity [Δχ2(59) = 73.56, p > .05]. For living arrangement, we conducted a multiple group model comparing adolescents who lived with two parents to those living with anyone else (e.g., one parent, step-parent, grandparents), and found no differences in this regard [Δχ2(36) = 32.36, p > .05]. We attempted to run a multiple group model comparing adolescents who lived with two parents, one parent and a step-parent, one parent only, or “other” (i.e., grandparents). However, this model would not converge due to the small number of adolescents in the “other” group.
