Access Type

Open Access Dissertation

Date of Award

January 2015

Degree Type


Degree Name




First Advisor

Ty Partridge


Adolescence is a developmental period marked by much change across physical, cognitive, psychological, and social domains leading to greater vulnerability for poor decision making. As a result, adolescence is a period of increased risk taking behaviors. Prevention of risk behaviors would benefit from early intervention prior to the onset of these risk behaviors. Unfortunately, it is difficult to identify those youth who may be most at risk. Risk-taking adolescents may choose to engage in specific risk behaviors; as well, risk factors that influence risk taking may also differ as a function of the specific domains of risk behaviors. The present study assesses youth from a longitudinal trial of two HIV intervention prevention programs following the same cohort of youth from grade six to grade 12 in the Bahamas. A person-centered approach was used to examine risk behaviors and determine whether there are certain behaviors that co-occur among different subsets of youth. Latent class analysis and latent transition analysis revealed four distinct profiles of risk behavior involvement. Latent classes of grade 12 behaviors included a low risk class, a high risk class and two moderate risk behaviors classes. The patterns of the four risk statuses identified in the latent transition analysis were similar to the latent classes. Results showed that involvement in risk behaviors at a previous time point increased the probability of remaining in either the same risk status or transitioning to another risk status compared to transitioning to the low risk status. Furthermore, it was hypothesized that contextual factors including perceived peer involvement in risk behaviors, parental monitoring, and neighborhood exposure to risk behaviors and individual level factors such as sensation seeking and values orientation would predict group membership. Results showed peer involvement in risk behaviors, neighborhood risk involvement, and gender were significant predictors of latent classes and peer involvement in risk behaviors and gender were significant predictors of latent statuses.