Using Latent Class Analysis and the Social-Ecological Framework to Predict Membership in Latent Classes of Bullying, Victimization, and Bystander Behavior and Examining Associations with the Derived Classes Among High School Students

Abstract

Bullying in schools across the United States is highly prevalent and is considered an Adverse Childhood Experience (ACE) by the Centers for Disease Control and Prevention. In this quantitative study, I used Latent Class Analysis (LCA) to identify and describe underlying classes of bullies, victims, and bystanders among a sample of high school students in a mid-western state in the United States. Further, using the social-ecological model (SEM), I examined the individual-, family-, peer-, and school-level factors associated with bullying as they relate to the identified classes. To my knowledge, no research, up to the time of conducting this study had specifically used LCA with a sample of high school students in United States to examine bullying victimization, perpetration, and bystander behaviors concurrently. In addition, researchers have seldom systematically examined the associations of individual-, family-, peer-, and school-level factors with different patterns of bullying, victimization, and bystander classifications. My study is anticipated to fill this gap in literature by employing LCA to explore the subgroups of bullies, victims, and/or bystanders, and the unique risk and/or protective contextual factors for each subgroup. I used data from the Dane County (Wisconsin) Youth Assessment Survey of 2018 for this study. I conducted a multilevel LCA to identify underlying classes of bullies, victims, and/or bystanders in this sample of high school students (Grades 9-12). The sample size was 16,288 students across 24 schools. Three latent classes were identified - “uninvolved” (68.7%), “high verbal victimization” (23.6%), and “high perpetration and victimization” (7.7%). Furthermore, studies using LCA to study bystander behavior in bullying are sparse even though bystanders can play an important role in alleviating bullying incidents and defending behavior has become an important factor in antibullying interventions. This gap in bystander literature was addressed in my study by shedding light on bystander behavior by examining both positive and negative bystanders of bullying. My study suggested that the “high perpetration and victimization” class had a higher probability of being both positive (0.76) and negative bystanders (0.6), whereas the “high verbal victimization” class had a higher probability of being positive bystanders (0.61) but not negative bystanders (0.01). The “uninvolved” class had low probabilities of being both positive (0.17) or negative bystanders (0.01). In the next step of the analysis, the three identified latent classes were used as a dependent variable in a multilevel multinomial logistic regression analysis to examine how the classes were associated with several risk and protective factors of the social ecological framework. Several individual-, family-, peer-, and school-level factors were associated with the latent classes of bullying behavior. High perpetration and victimization behaviors were predicted by male gender (RRR=0.34, for females), younger grades, disability status (RRR=1.4), anxiety (RRR=1.42), self-harm (RRR=1.24), suicidal ideation (RRR=1.54), alcohol use (RRR=1.55), tobacco use (RRR=1.78), lower levels of parental structure (RRR=0.84), exposure to ACEs, high deviant peer affiliation (RRR=1.31), and lower connection to school (RRR=0.56). High verbal victimization behaviors were predicted by male gender (RRR=0.79, for females), younger grades, white race, disability status (RRR=1.48), depression (RRR=1.42), anxiety (RRR=1.66), suicidal ideation (RRR=1.45), alcohol use (RRR=1.12), higher levels of parental structure (RRR=1.23), exposure to ACEs, and lower connection to school (RRR=0.63). Overall, almost all factors indicated that the “uninvolved” class had fewer risk factors than students in other profiles of bullying and victimization thus suggesting that students who are involved in bullying either as perpetrators, victims, and/or bystanders are at a higher risk for other negative outcomes as well. It is therefore urgent and necessary to identify and intervene for youth involved in bullying. My results provide implications for tailored prevention and intervention strategies to reduce adolescent bullying and victimization through a multi-contextual holistic approach.