College Students’ Online Pornography Use: Contrasting General and Specific Structural Variables with Social Learning Variables

Article in American Journal of Criminal Justice by Danielle Cooper and Jennifer Klein.

This research partially tests Akers’ social structure-social learning theory (SSSL). The data was collected online through a self-report questionnaire and nearly half (48.8%) of the sample of 812 college students reported visiting a porn site. To better understand this self-report behavior, bivariate correlations and three binary logistic regression analyses were conducted. In Model A, participants who were male, Hispanic, had more years in college, and more inclined toward homosexuality had an increased likelihood of visiting a porn site. In Model B, again, gender, ethnicity, year in school, and the sexuality scale were significant predictors. However, race appeared as significant for the first time along with number of sex partners, and frequency of masturbation, indicating that participants who were Black, had a greater number of sexual partners, and masturbated more frequently had an increased likelihood of visiting porn site. As with the first and second models, gender, race, sexuality scale, and frequency of masturbation were significant predictors in Model C. Additionally, differential peer association, differential reinforcement, and definitions favorable were significant, indicating that participants who had greater association with peers who viewed porn, who had observed someone watching porn and decided to mimic their behaviors, and who had defined visiting porn sites favorably had an increased likelihood of visiting a porn site. Overall, Akers’ SLT variables fully mediated ethnicity, year in school, and number of sex partners, but it only partially mediated gender, race, and sexuality scale.