Predicting adolescent AIDS-related risk behavior from psychosocial factors: A path analysis

This study had two purposes. The first was to develop a model for relating the unique and combined contributions of knowledge about AIDS, attitudes toward AIDS, and several psychosocial characteristics to the self-reported sexual activity of adolescents. The second purpose was to validate the proposed causal paths by studying a sample of ninth-grade adolescents and using the results of a path analysis to revise the a priori model. The model validation study was conducted among 179 ninth-grade adolescents. A questionnaire involving knowledge, attitudinal, psychosocial characteristics, and self-report behavioral information was administered (spring, 1989). The instrument used combined a questionnaire developed by Centers for Disease Control to obtain knowledge, attitude, and behavior assessments with a questionnaire that assessed self-esteem, susceptibility to peer pressure, and locus of control which was developed by T. E. Dielman. Respondents were divided into two groups, those having had sex and those not yet sexually active. A causal path model relating all variables to sexual activity was proposed; the model was evaluated using correlations, multiple regression, and path analysis. Multiple regression analyses failed to support further analysis of the data by gender subgroup membership. No significant differences were found in the sexual activity between males and females; therefore, a path analysis was constructed for the combined sex group. In the path analysis, susceptibility to peer pressure had the strongest influence on sexual behavior, followed by locus of control and perceived vulnerability. Results indicated that those adolescents who were sexually active have lower internal control, are more subject to peer pressure, and have lower self-esteem. Attitude toward AIDS, while significantly affected by knowledge, was not shown to have a significant effect on sexual activity. The results lend support to earlier findings addressing psychosocial characteristics and adolescent drug use, which suggest that differences in these characteristics may affect individuals\u27 health behavior more than knowledge or attitudes. A revised model based on significant findings was presented for future research and for curriculum development