Open Access
Maas, Megan Keely
Graduate Program:
Human Development and Family Studies
Doctor of Philosophy
Document Type:
Date of Defense:
July 20, 2016
Committee Members:
  • Jennie G Noll, Dissertation Advisor/Co-Advisor
  • Jennie G Noll, Committee Chair/Co-Chair
  • Eva S Lefkowitz, Committee Member
  • Bethany Cara Bray, Committee Member
  • Patricia Barthalow Koch, Outside Member
  • social media; teen dating violence; sexual assault; female adolescents; maltreatment; sexting; pornography; HIV risk; substance use
Adolescents spend a substantial amount of time on internet-connected devices (Refuel Agency, 2015), yet little is known about how adolescents use this new online context for sexual purposes or how individual differences in usage predict later outcomes. In addition, female adolescents have a unique experience of sexual socialization in both online and offline contexts compared to male adolescents, as there are contradictory cultural norms which encourage female adolescents to prove their sexual attractiveness but shame them for engaging in sexual behavior (Bay Cheng, 2015). To further complicate the issue, maltreated female adolescents face even more difficulty during adolescent sexual development as they process past trauma or navigate their world without sound parental guidance (Noll et al., 2013). Consequently, the internet becomes a space for sexually developing female adolescents to learn as well as practice sexual scripts, or cultural rules about sexual behavior and attitudes. Much of the research that examines online sexual behaviors among adolescents takes a variable-centered approach which lacks complexity to capture the multidimensionality of online sexual behavior, as these behaviors do not occur in isolation. Thus, this dissertation used a person-centered approach, latent class analysis (LCA), to identify patterns of a broad range of online sexual experiences (referred to as classes) to explore offline sexual behavior and substance use correlates among female adolescents specifically. Then, established classes were used to differentially predict later HIV risk, teen dating violence (TDV), and sexual assault to examine how class structure, distribution, and prediction differs between maltreated and non-maltreated female adolescents. Female adolescents (N = 312, mean age = 15.21, 45% Caucasian) who were participating in the cross-sequential Female Adolescent Development Study (FADS) filled out questionnaires across 2-5 years from approximately 14-19 years old. Nearly half had experienced substantiated maltreatment that was verified through child protective services. In study 1, a LCA was performed in Latent Gold 5.0. Based upon fit statistics and selection criteria, a four-class model was selected with the following classes: Online Abstinent, Online Inclusive, Attractors, and Seekers. Maltreated participants were more likely to have engaged in most online and offline sexual behaviors than non-maltreated participants. The same four classes were observed in both maltreated and non-maltreated participants. However, maltreated female adolescents were more likely to be members of the Online Inclusive class than any other class. In study 2, established latent classes from Study 1 differentially predicted HIV risk, TDV, and sexual assault one year later. The Attractors class was more likely to engage in HIV risk behavior and to experience TDV and sexual assault one year later compared to the Online Abstinent class. Maltreatment status moderated the prediction of sexual assault by class membership such that maltreated female adolescents in the Online Inclusive class were more likely to be sexually assaulted than non-maltreated female adolescents in the Online Inclusive class. Taken together, these studies suggest that interventions that target online sexual experiences should (1) focus on characteristics that make individuals vulnerable to online experiences instead of approaching online risk as a global issue for all adolescents, (2) address maltreatment as a unique risk factor for online and offline sexual experiences, and (3) tailor messages and programming as different adolescents are deferentially at risk for future outcomes.