Article révisé par les pairs
Résumé : Introduction Several studies have evaluated the effectiveness of school-based anti-bullying interventions (SBABI). However, their evolution over time has not been studied. The SET-Bullying project aimed to further explore this question by performing a secondary data analysis on data used for the evaluation of the effectiveness of 2 SBABIs. Methodology Scores of pupil self-reported frequencies of being bullied and bullying others were analysed using mixed effect models, taking into account pupil gender, school type (primary vs secondary), and the hierarchical and longitudinal structure of the data. Three models were fitted: (1) making no assumption on the evolution of effectiveness over time (time as a categorical variable), (2) assuming a linear evolution (time as a continuous variable with a linear term), and (3) assuming a minimal evolution at early study measurements and a sharper evolution later (time as a continuous variable with both linear and quadratic terms). Results The three models gave similar results. The model with the quadratic term seemed to fit the data slightly better compared to the other two. The magnitude and shape of the evolution differed by score, school type, and pupil gender. The observed patterns were either a linear evolution of effectiveness over time or a “delayed effect” after 24 months since baseline. Conclusions Taking into consideration the possibility of a “delayed effect” when designing and evaluating SBABIs allows to select the appropriate SBABI components and evaluation schedules. Thus, both SBABIs can become more efficient and evaluation practices can allow for the effectiveness to be observed.

Documents en relation

DI-fusion