Co-Principal Investigator: Farkas, George

Purpose: This project will use social ecological theory to investigate modifiable factors in family, classroom, school, and neighborhood contexts that are associated with or predictive of bullying victimization in 3rd-5th grade, the extent to which these social contextual factors function as mediators or moderators of other explanatory factors (e.g., family sociodemographic level, children's K–2nd grade academic or behavioral functioning), and whether and to what extent bullying victimization is related to children's academic, behavioral, and socioemotional functioning at the end of 5th grade. Despite meta-analyses indicating that elementary school is the optimal time period for bullying prevention interventions, the available empirical work has mostly analyzed samples of adolescents, used cross-sectional or short-term (e.g., one year) longitudinal designs, relied on single-item or -reporter bullying victimization measures that do not include specific subtypes, used no or very limited statistical control for potential confounds, and/or do not report on risk or protective factors as early as the primary grades. Consequently, the field's knowledge base about the early risk and protective factors for bullying victimization, how victimization and perpetration inter-relate dynamically over time across the elementary grades, and the extent to which bullying victimization is related to children's academic, behavioral, executive, and socioemotional functioning at the end of elementary school is currently very limited. Particularly limited is rigorously derived knowledge about the modifiable factors in children's social ecological contexts (i.e., families, classrooms, schools, and neighborhoods) that constitute promising targets of bullying prevention interventions and how these contextual factors may mediate or moderate other modifiable factors including children's early academic or behavioral functioning.

Project Activities: The researchers will analyze data from the Early Childhood Longitudinal Study-Kindergarten Cohort of 2011 (ECLS-K: 2011). The ECLSK: 2011 is a multiyear population-based cohort of U.S. schoolchildren followed from kindergarten entry to the end of 5th grade. Students, parents, and teachers completed multi-item measures of bullying and victimization during 3rd-5th grade. These measures allow for analyses of specific types of bullying victimization (i.e., physical, verbal, relational, and social). The measures also allow us identify students who are bullies, victims, or bully-victims. The project's analyses will include OLS regressions, structural equation models (SEM), tests for mediation and moderation, and random-intercept cross-lagged panel models (RI-CLPM) that rigorously evaluate the over-time dynamic between victimization and perpetration.

Products: The researchers will disseminate the project's findings at national refereed conferences and through peer-reviewed publications in high-impact journals. They will also disseminate the project's findings to educational researchers, policymakers, and practitioners through non-technical op-eds, briefings, social media, and national media reporting. They will upload working papers and preprints onto non-paywalled sites.

Structured Abstract

Setting: The ECLSK: 2011 includes a nationally representative sample of 968 U.S. elementary schools.

Sample: This includes a nationally representative sample of U.S. schoolchildren who entered kindergarten classrooms in 2010–2011 (N = 21,400) and who were then repeatedly and individually assessed at least yearly as they attended elementary schools until the end of 5th grade.

Factors: Factors examined in relation to bullying victimization include student, family, classroom, school, and neighborhood characteristics, particularly those modifiable in educational contexts (e.g., student-level academic achievement or behavior, school-family communication, classroom orderliness, teacher experience and certification, classroom and school size and sociodemographic composition, school climate and safety). Factors examined in relation to student level functioning by the end of elementary school include bullying victimization in analyses adjusted for other modifiable but potentially confounding factors (e.g., prior behavior).

Research Design and Methods: Students, teachers, school administrators, and parents were individually surveyed during kindergarten, 1st, 2nd, 3rd, 4th, and 5th grade. Multi-informant and multi-item measures of both bullying and victimization were administered in 3rd, 4th, and 5th grade. The researchers hypothesize that students attending less-resourced classrooms and schools in more dangerous neighborhoods will be at greater risk for bullying victimization. They also hypothesize that students who are bullies, victims, or bully-victims will be at greater risk of experiencing academic, behavioral, executive functioning, and socio-emotional difficulties by 5th grade including in analyses controlling for potential confounds.

Control Condition: Due to the nature of this project, there is not a control condition.

Key Measures: The measures in this study assess the frequency of physical, verbal, relational, and social victimization and perpetration. Additional student-level measures include: (a) student and family socio-demographics (e.g., family SES, race/ethnicity, language spoken at home, disability status); (b) parenting and home environment characteristics including parental stress, use of harsh discipline (e.g., spanking, hitting), parental warmth, economic insecurity, depression and substance use, and social support; (c) student academic, behavioral, and socio-emotional functioning using individually administered achievement measures, teacher behavioral ratings, and student self-report; and (d) self-regulatory as well as externalizing and internalizing behaviors. Teachers reported on classroom characteristics including: (a) size; (b) socio-demographic composition; (c) classroom behavior; (d) time spent on classroom discipline and addressing disruptions. Teacher also reported on their: (a) experience levels, certifications, and professional development; (b) communication with families; and (f) feelings of self-efficacy. Teachers and school administrators reported on school characteristics including: (a) size; (b) socio-demographic composition; (c) safety and orderliness including the level of student misbehavior; and (d) the frequency of fighting, bullying, vandalism, disorder as well as to what extent absenteeism, aggressive or disruptive behaviors, or teacher turnover were occurring. School administrators reported on the school's neighborhood characteristics including the degree of: (a) drug use or excessive drinking in public; (b) gangs; (c) crime and violence; and (d) vacant houses and buildings. Parents reported on their neighborhood's safety.

Data Analytic Strategy: The data analyses will include multivariable Ordinary Least Squares (OLS) and multinomial logistic regressions with interactions. The research team will also conduct multilevel structural equation modeling (SEM) to identify moderators and mediators, and random intercept cross-lagged panel models (RI-CLPM) including with time-varying measures to provide evidence of potential causality of the over-time dynamic between victimization and perpetration. Analyses will include sampling weights, missing data procedures, and clustered standard errors to provide generalizable and unbiased estimates as well as to account for clustering.

Related IES Projects: ADHD: Population-Based Estimates of Diagnosis, Treatments, and School Outcomes (R324A120331); Who Receives and Benefits From Special Education in the U.S.? Analyses of Three Nationally Representative Datasets (R324A200166)