L’objet de cette étude est de proposer une version française validée de l’échelle de rumination de colère – Anger Rumination Scale [12Sukhodolsky D.G., Golub A., Cromwell E.N. Development and validation of the anger rumination scale Pers Ind Diff 2001 ;  31 : 689-700 [cross-ref]

Cliquez ici pour aller à la section Références]. L’apport principal de cette échelle est de s’intéresser à la nature des contenus cognitifs supports de la rumination liée à l’expérience émotionnelle de colère. La validité factorielle de la version française a été testée auprès d’un large échantillon d’étudiants (n=617) répartis aléatoirement en deux groupes. Une analyse factorielle exploratoire (AFE) a été conduite avec les données du groupe 1, et une analyse factorielle confirmatoire (AFC) par modèle d’équation structurelle a été conduite avec les données du groupe 2. L’AFE reproduit le modèle à quatre facteurs de la version originale, et l’AFC met en évidence la bonne adéquation du modèle initial (χ²₁₄₆ : 302,14 CFI : 0,96 RMSEA : 0,058). De la même manière, les indices de consistance interne de reproductibilité et de validité externe se sont révélés satisfaisants. La validité de cet outil en langue française ouvre les champs d’étude de l’effet des processus de rumination dans la dynamique qui lie colère et comportement d’agression.

The aim of this study was to propose a French version of the Anger Rumination Scale (ARS). The main contribution of this scale is to evaluate cognitive responses involved in rumination experience related to anger. The rumination experience is assessed through four scales: angry afterthoughts, thoughts of revenge, angry memories and understanding of causes. Because anger rumination is related to anger experience, it is usually understood to be a dysfunctional or unconstructive behavior. However, because the rumination involves cognitive activity, it can be assimilated as a cognitive strategy to cope with a negative event in a specific context. The ARS seems to be useful in understanding the rumination effects both on anger experience as well as on aggressive behaviors.

The original scale was translated following scientific recommendations for cultural adaptation of questionnaires. Six hundred and seventeen voluntary undergraduate students were asked to complete the ARS-French version. Among these 617 students: 305 completed the French ARS version twice over a 1-month period to test the reliability of the French version; 361 filled out both the French ARS version and the STAXI-II; 342 filled out both the French ARS version and the Aggression Questionnaire (AQ). The Total Sample was randomly split into two groups. An exploratory factor analysis (EFA) was conducted with group I's data. The confirmatory factor analysis (CFA) was conducted with group II's data.

The EFA revealed a four factors solution close to the original version. The CFA confirmed a good model fit for the original four factors solution (χ²₁₄₆: 302.14 CFI: 0.96 RMSEA: 0.058). A one factor model was tested, but was not retained. The ARS's internal consistency coefficients replicated those of the literature: alpha coefficients ranged from 0.60 to 0.86, and test-retest correlation coefficients ranged from 0.65 to 0.70. External validity indices conformed to previous studies, both for gender differences and for relationship with STAXI-II and AQ.

Results of the present study support the validity and the reliability of the French version of the ARS. This validation provides a multifactorial scale for assessing anger rumination. This scale could be used, for example, to explore processes leading to aggressive behavior in several contexts in which overt behavior is prohibited or detrimental to performance.