잠재계층분석을 이용한 한국 청소년 사이버불링 가해행태 분석

초록

This study aims to categorize adolescent cyberbullying perpetration using a person-centered approach, as opposed to the conventional variable-centered method. The goal is to create an academic foundation for tailoring cyberbullying prevention strategies that align with unique perpetration patterns shown by individual adolescents. To achieve this goal, the study utilized data from the 「Korean Children and Youth Welfare Panel Study 2018」conducted by the Korea Institute of Child and Adolescent Welfare. Using Latent Class Analysis (LCA), we identified latent groups of adolescents with comparable cyberbullying perpetration patterns. These groups were then studied to uncover distinctions in demographic charateeristics (gender), mental health factors (aggression, depression, grit), antisocial behaviors (delinquency), and relational factors (teacher relationships, friendships, parental caregiving attitudes). The key findings are as follows: First, a statistically significant model has been derived, which includes two to five latent groups, each displaying unique patterns of cyberbullting perpetration behavior. Second, considering model fit, types of perpetration behavior, and sample sizes, a four-group latent model has been determined as the most appropriate for achieving the research objectives. These four latent groups are labeled as: 'Low-Level Perpetration Group (n=2,087)', 'Verbal Aggression · Cyber Prison · Flaming Concentrated Group (n=36)', 'Verbal Aggression Concentrated Group (n=74)', and 'High-Level Perpetration Group (n=5)'. Third, there are no significant differences observed in the characteristics of the members within each latent cyberbullying perpetration group. However, unexpected outcomes are noted concerning delinquency, teacher relationships, and positive parental caregiving attitudes (warmth, autonomy support, structure). Consequently, the need for further research on categorizing types of adolescent cyberbullying perpetration and conducting detailed analyses for each type is proposed.

키워드