The paper first reviews how interrelationships among adolescent risk behaviors are typically studied; discusses proposed explanations for why risk behaviors are interrelated; and explains why such accounts matter to theory, research, and related intervention efforts. The authors then describe an aspect of structural equation modeling (SEM) related to the meaning of explaining variance in a latent risk-behavior factor that is often misinterpreted by risk-behavior research, but which has substantive implications. The paper then demonstrates a simple and easily interpreted method for directly examining hypotheses regarding why risk behaviors are linked. The authors then propose a novel inquiry procedure that is an important nexus for adolescent risk-behavior theory, research, and intervention efforts. They propose that explanatory factors should be evaluated in terms of whether they predict individual risk behaviors and account for the linkages among risk behavior. Affirmative evidence for both criteria would indicate that such predictors could be considered possible common predictive factors and possible sources of the linkages among multiple risk behaviors, i.e., potential linkage factors. 3 tables and 41 references