Condom Use Self-Efficacy and Peer Substance Use Norms Among South African Secondary School Students: An Application of Bivariate Growth Mixture Modeling

Introduction: In South Africa, risky individual sexual behaviors, particularly among youths, have turned into national public health concerns. Adolescent research has found that individuals using substances are more likely to have sexual encounters, engage in riskier sexual practices, and experience an adverse sexual health outcome. Having rebellious or deviant peers increases the likelihood that an adolescent will engage in deviant and risky behaviors. This study examines whether the development of sexual behavior and perceptions of peer deviance co-vary among South African high school students and whether there are subgroups of students with different developmental patterns.

Methods: We applied bivariate growth mixture modeling techniques to multiple waves of data from South African eighth graders (N=5,677; M_age=13.97). At each wave, adolescent sexual behavior was measured by a 6-item condom use self-efficacy score (CUSE:  and peer deviance was measured by a 3-item perceived peer substance use score (PSUB: . Based on a large body of literature suggesting gender influences both condom use self-efficacy and substance use, gender was included as a covariate.

Results: On average, students’ CUSE and PSUB rose between waves 1 and 8, however this masked some heterogeneity. The analysis revealed the development of CUSE in relation to PSUB was best represented using a three class linear growth mixture model. The first class had the highest initial CUSE that improved over time and moderate initial PSUB that increased the most rapidly over time (41.8% of students). The second class had moderate CUSE that improved the fastest and the highest initial PSUB that still increased over time (43.8%). The final class was characterized by low CUSE that improved the least over time and low, unchanging PSUB (14.6%).

Conclusions: Where there is higher perceived peer substance use, there tends to be higher condom use self-efficacy. Over time, people with the least improvement in condom use self-efficacy also perceived low levels of substance use in their schools. Future studies can expand on these findings by exploring the reasons behind the association between high condom use self-efficacy and high perceptions of substance use among peers. These findings suggest that, while most students respond well to existing programs and policies to help them develop safer sex behaviors, there is a subgroup that is not being reached. Understanding why some students’ condom use self-efficacy development slows with low initial self-efficacy and low perceived peer deviance can be informative for future research, prevention, and policy efforts.