Using Structural Equation Modeling to Test the Theory of Reasoned Action: Predicting Sexual Behavior

Purpose: This study examined proximal, intrapersonal factors associated with delaying first sexual intercourse in a national sample of adolescents. Specifically, this research investigated the utility of the Theory of Reasoned Action (TRA) as a plausible model for predicting whether an adolescent engages in sexual intercourse. Methods: This analysis used data from Waves I and II of the National Longitudinal Study of Adolescent Health to test the applicability of the TRA (n = 3,395). Structural equation modeling (SEM) tested the TRA constructs simultaneously and accounted for measurement error. Results: With the exception of the path from intention to behavior at Wave II, the standardized regression weight and all probit regression coefficients were statistically significant (p<.05): path between attitude and subjective norm = -.301, from attitude to intention = .328, subjective norm to intention = -.415, intention to behavior at Wave I = -.557, and behavior at Wave I to behavior at Wave II = .854. The modification indices suggested the addition of a path from subjective norm to behavior at Wave I (.186). Conclusion: The TRA may be useful for predicting whether an adolescent engages in sexual intercourse. All measured variables for subjective norm concerned one's perception of parental approval or disapproval of engaging in sexual intercourse. Researchers should continue to explore the importance of perceived norms, particularly parental perceptions. Educators and other practitioners should continuously reevaluate theoretical models on which programs and interventions are based, and SEM is a useful statistical technique for examining the practical utility of theory.