Quasi-experimental evidence on short- and long-term consequences of bullying victimization: a meta-analysis

Exposure to bullying victimization is associated with a wide-range of short and long-term adverse outcomes. However, the extent to which these associations reflect a causal influence of bullying victimization remains disputed. Here, we aimed to provide the most stringent evidence regarding the consequences of bullying victimization by meta-analyzing all relevant quasi-experimental (QE) studies. Multilevel random effects models and metaregression were employed to (a) estimate the pooled QE-adjusted effect size (Cohen d) for bullying victimization on outcomes and to (b) evaluate potential sources of heterogeneity. A total of 16 studies were included. We derived 101 QE-estimates from three different methods (twin design, fixed effects analysis, and propensity score matching) for three pools of outcomes (internalizing symptoms, externalizing symptoms, academic difficulties). QE-adjusted effects were small for internalizing symptoms (dadjusted = 0.27, 95% CI [0.05, 0.49]), and smaller for externalizing symptoms (dadjusted = 0.15, 95% CI [0.10, 0.21]) and academic difficulties (dadjusted = 0.10, 95% CI [0.06, 0.13]). Accounting for a shared rater effect between the exposure and the outcome further reduced the effect for internalizing (dnonshared rater = 0.14, 95% CI [0.05, 0.23]) and externalizing symptoms (dnonshared rater = 0.06, 95% CI [0.01, 0.11]). Finally, the adverse effects declined in the long-term, most markedly for internalizing symptoms (dlong-term = 0.06, 95% CI [-0.01, 0.13]). Based on the most stringent evidence available to date, findings indicate that bullying victimization may causally impact children's wellbeing in the short-term, especially anxiety and depression levels. The reduction of adverse effects over time highlights the potential for resilience in individuals who have experienced bullying. Secondary preventive interventions in bullied children should therefore focus on resilience and address children's preexisting vulnerabilities.Exposure to bullying victimization is associated with a wide-range of short and long-term adverse outcomes. However, the extent to which these associations reflect a causal influence of bullying victimization remains disputed. Here, we aimed to provide the most stringent evidence regarding the consequences of bullying victimization by meta-analyzing all relevant quasi-experimental (QE) studies. Multilevel random effects models and metaregression were employed to (a) estimate the pooled QE-adjusted effect size (Cohen d) for bullying victimization on outcomes and to (b) evaluate potential sources of heterogeneity. A total of 16 studies were included. We derived 101 QE-estimates from three different methods (twin design, fixed effects analysis, and propensity score matching) for three pools of outcomes (internalizing symptoms, externalizing symptoms, academic difficulties). QE-adjusted effects were small for internalizing symptoms (dadjusted = 0.27, 95% CI [0.05, 0.49]), and smaller for externalizing symptoms (dadjusted = 0.15, 95% CI [0.10, 0.21]) and academic difficulties (dadjusted = 0.10, 95% CI [0.06, 0.13]). Accounting for a shared rater effect between the exposure and the outcome further reduced the effect for internalizing (dnonshared rater = 0.14, 95% CI [0.05, 0.23]) and externalizing symptoms (dnonshared rater = 0.06, 95% CI [0.01, 0.11]). Finally, the adverse effects declined in the long-term, most markedly for internalizing symptoms (dlong-term = 0.06, 95% CI [-0.01, 0.13]). Based on the most stringent evidence available to date, findings indicate that bullying victimization may causally impact children's wellbeing in the short-term, especially anxiety and depression levels. The reduction of adverse effects over time highlights the potential for resilience in individuals who have experienced bullying. Secondary preventive interventions in bullied children should therefore focus on resilience and address children's preexisting vulnerabilities.