The Longitudinal Interplay between Social-Emotional Competence and School Bullying: A Cross-Lagged Panel Network Analysis

Background and Purpose: School bullying remains a critical public health and social justice issue impacting children’s well-being. While social-emotional competence (SEC) is often theorized as a protective factor, the temporal precedence between SEC and involvement in bullying (both victimization and perpetration) remains underexplored in longitudinal contexts. Furthermore, previous research has largely relied on traditional variable-centered approaches that treat SEC and bullying as unitary constructs. This approach obscures the specific, item-level interactions that drive their relationship, thereby hindering the development of more precise and targeted interventions. Consequently, this study aims to utilize Cross-Lagged Panel Network (CLPN) analysis to explore the dynamic, longitudinal interactions between specific social-emotional competencies and bullying behaviors.

Methods: With approval at the school and district level, students in Grades 4 and 5 from two primary schools in China were invited to participate. With parental/guardian consent and student assent, a total of 1,041 students completed assessments across two waves with a six-month interval. At Wave 1, the average age of participants was 10.74 years (SD = 0.78, range = 9–12), with 51.4% being female. Data were collected via self-report questionnaires. SEC was measured using the WCSD Social and Emotional Competency Short-Form Assessment, comprising 17 items across five dimensions: self-awareness, self-management, social awareness, relationship skills, and responsible decision-making. Bullying involvement was assessed using the Olweus Bully/Victim Questionnaire (6 items for each). Two separate CLPN models were constructed to examine the longitudinal dynamics between SEC and bullying victimization, and SEC and bullying perpetration, respectively.

Results: The results demonstrate a unidirectional negative predictive effect from bullying involvement to SEC over time for both victimization and perpetration. However, the specific longitudinal dynamics differ between the two networks. (1) For the victimization network, “being mocked because of accent” exerted the greatest influence on subsequent SEC nodes. Specifically, it was the strongest negative predictor of “persisting through frustration” over time. Additionally, “knowing ways to calm down” and “getting along with classmates” were the nodes most significantly impacted by prior victimization. (2) For the perpetration network, “excluding or ignoring others” had the greatest negative impact on later SEC; specifically, it emerged as the strongest negative predictor for “knowing when feelings affect focus.” Furthermore, the nodes “knowing ways to calm down” and “thinking before deciding” were most influenced by prior bullying perpetration.

Conclusions and Implications: These findings suggest that anti-bullying strategies should move beyond broad programs to prioritize addressing discriminatory verbal victimization (e.g., accent-based mockery) and relational perpetration (e.g., social exclusion). Furthermore, interventions should emphasize improving specific emotional knowledge and regulatory competence for both victims and perpetrators to mitigate developmental damage. By identifying these precise intervention targets, practitioners can more effectively prevent school bullying and safeguard the long-term well-being of at-risk youth.