Independent and dependent variables in adolescent and young adult sexuality research: Conceptual and operational difficulties

We concluded, after reviewing 130 articles related to adolescent and young adult sexuality, that efforts to understand sexuality in this population are hampered by three aspects of typical approaches to the research process: (a) examination of single classes of predictor variables; (b) differences in operationalization of dependent variables; and (c) sampling differences. The present study, in an effort to demonstrate that conceptual clarity is impeded by these factors, examined the simultaneous influence of multiple predictor variables using data collected from a sample of college students enrolled in an undergraduate human sexuality class. We examined four different classes of dependent variables; birth control, choice of sexual partners, condom use, and the combination of condom use and choice of sexual partners. Results support the strategy to conduct separate analyses for each class of dependent variables. No single predictor variable, for example, was significant in all four regression models.