Le harcèlement scolaire est un problème de santé publique aux conséquences graves, particulièrement dès l’école primaire. Les données quantitatives étant rares pour les 7–11 ans en Wallonie, cette étude poursuit un double objectif : évaluer les propriétés psychométriques du questionnaire agresseur/victime révisé d’Olweus et dresser un premier état des lieux du phénomène dans le Hainaut.

Une étude prospective observationnelle a été menée auprès de 77 élèves de 7 à 11 ans. Le questionnaire agresseur/victime révisé d’Olweus, un auto-questionnaire anonyme, a été utilisé pour évaluer les statuts de « victime » et « agresseur » après approbation du comité d’éthique. La validité de construit a été testée par Analyse des Correspondances Multiples et la fiabilité par le coefficient alpha de Cronbach.

Le questionnaire présente une fiabilité modérée à acceptable. L’analyse des correspondances multiples révèle une structure cohérente pour l’échelle « victime » (α = 0,689) mais plus dispersée pour l’échelle « agresseur » (α = 0,628), suggérant un biais de désirabilité sociale. L’item « racisme » se distingue comme problématique ; son retrait améliore la cohérence de l’échelle « agresseur » (α = 0,66). L’état des lieux révèle une prévalence alarmante de « victimes » (49,4 %), avec 19,5 % de « victimes-agresseurs » contre seulement 1,3 % d’« agresseurs » purs. Le harcèlement psychologique est la forme la plus subie, le physique la plus exercée.

La discussion souligne la pertinence du questionnaire comme outil de dépistage malgré ses limites. La prévalence élevée et la complexité des profils (victime-agresseur, témoins) appellent à dépasser une vision binaire du harcèlement et à renforcer la prévention précoce. Ces résultats ouvrent des perspectives pour la pratique clinique et pour des recherches longitudinales.

School bullying is a major public health concern with profound consequences, affecting one in four European children from an early age, according to UNESCO data. Defined by intentionality, repetition, and a power imbalance, its impacts are severe, ranging from mental health disorders and academic decline to lasting neurobiological alterations of the stress axis. Chronic stress induced by bullying is known to deregulate the hypothalamic-pituitary-adrenal (HPA) axis, potentially leading to structural changes in key brain regions like the hippocampus and the amygdala. Despite the severity of the phenomenon, a significant research gap exists regarding quantitative data for children aged 7–11 in Wallonia (French-speaking Belgium). Most large-scale studies (like HBSC) focus on adolescents. This study addresses this gap with two primary goals: To assess the psychometric properties (construct validity and internal consistency) of the French version of the Olweus Bully/Victim Questionnaire-Revised (rBVQ) for this younger population; To provide a descriptive overview of bullying prevalence, roles, and forms in primary schools in the Hainaut province.

A prospective, observational study was conducted in three primary schools in Mons, Belgium. The final sample included 77 students aged 7 to 11. Data collection utilized the rBVQ, an 18-item anonymous self-report questionnaire. To ensure comprehension for younger children (grades 2 to 6), the questionnaire was administered in small groups with standardized oral instructions and a full read-aloud of each item. Statistical Approach: Construct validity was explored using Multiple Correspondence Analysis (MCA). This choice was dictated by the qualitative nominal nature of the data (frequency levels 0, 1, and 2). MCA allows for a non-linear mapping of response patterns, visualizing the proximity between specific items and student profiles. Reliability was assessed using Cronbach's alpha coefficient.

Psychometric Validation : Psychometric analyses supported the rBVQ's utility for this age group: Victim Scale (α = 0.689): MCA revealed a coherent spatial structure where items clustered into three distinct groups: “not involved”, “intimidated”, and “victims”. This indicates that children as young as 7 are capable of reporting their victimization consistently. Bully Scale (α = 0.628): The structure was more scattered, likely reflecting a social desirability bias. Children appear more reluctant to admit to aggressive behaviors, causing a “clustering” of responses toward the center of the MCA plot. The “Racism” Item: This item was identified as a problematic outlier. Its isolation in the MCA and its negative correlation with other aggression items suggest it measures a different construct – likely related to social prejudices rather than the repetitive bullying dynamic. Removing it improved the bully scale's reliability (α = 0.66). Prevalence and Situational Analysis: The prevalence analysis revealed an alarmingly high rate of victimization (49.4%). Gender and Age: While victim status was evenly distributed between genders, roles involving aggression (pure bullies and bully-victims) were predominantly male. A significant “age shift” was observed at age 10: victimization rates dropped from 61.4% (7–9 years) to 33.3% (10–11 years), while the “Neutral” status became predominant. Forms of Bullying: Psychological bullying (exclusion, rumors) was the most frequent form sub-scale (44%), corroborated by strong correlations between “rumors” and “exclusion” items in the correlation matrices. Conversely, physical bullying was the most common form perpetrated, especially by boys, reflecting immature executive functions and emotional regulation in this developmental stage.

As child psychiatrists, we argue that these results necessitate a shift in clinical practice. The “Bully-Victim” Complexity: Nearly 20% of the sample occupied this mixed role. These children are the most vulnerable, combining internalized symptoms (anxiety, depression) with externalized behavioral issues. They represent a significant therapeutic challenge and a high risk for long-term psychiatric comorbidities. Mobilizing “Neutrals”: The increase in “Neutral” students (29.9%) after age 10 presents a prevention opportunity. These bystanders are not passive; they regulate the school climate. Transforming them into “defenders” can reduce the overall neurobiological stress in the classroom. Screening and Intervention: The high prevalence among 7–9 year-olds suggests that the rBVQ should be used systematically in clinical consultations as a mediation tool to objectify a child's experience that they might not spontaneously verbalize due to shame or fear.

This study confirms that school bullying in Wallonia is a massive and complex phenomenon starting in early primary education. It validates the rBVQ as a useful tool for detection and research in this context. The findings call for early, evidence-based prevention strategies that address the multifaceted roles – moving beyond the simple victim/bully binary – to mitigate the severe bio-psycho-social consequences of bullying on a child's developmental trajectory.