Validation de la version française d’une échelle évaluant les pratiques disciplinaires de parents d’enfants d’âge scolaire

Abstract

Cette étude investigue la validité de la version française de l’Échelle des pratiques parentales de Arnold et al. (1993). La version originale de cette échelle a montré son utilité pour identifier les parents dont les stratégies de discipline sont contreproductives et notamment pour quantifier les effets de certaines interventions comportementales. Nous avons testé les différentes structures factorielles proposées dans la littérature avec des analyses factorielles confirmatoires, sur un échantillon de 801 parents d’enfants d’âge scolaire (4 à 13 ans). Les résultats indiquent que le modèle à deux facteurs proposé par Irvine et al. (1999) est le meilleur pour notre échantillon. La structure est invariante selon le sexe du parent et de l’enfant, tout comme selon l’âge de l’enfant. Les indices de consistance interne des deux facteurs laxisme et réactivité excessive sont satisfaisants. Nous avons pu proposer des valeurs de référence pour les mères et les pères, ainsi qu’en fonction de l’âge des enfants, qui seront utiles au clinicien pour évaluer les pratiques de parents consultants. Des liens entre les facteurs et les difficultés de l’enfant évaluées avec le Questionnaire Points forts–Points faibles (SDQ) ont aussi été mis en évidence, les corrélations les plus élevées étant celles impliquant la réactivité excessive.

Several decades of research have emphasized the importance of parenting in the development of behavioral problems in children. Prevention and intervention for these difficulties require the assessment of parental practices with reliable measures. This study investigated the psychometric properties of the French version of the Parenting Scale (PS; Arnold et al., 1993). The original version of this scale has shown its usefulness by allowing the identification of the parents whose strategies of discipline are counterproductive, and by quantifying the effects of certain parenting programs. The study included 464 mothers and 347 fathers of elementary school-aged children (between the age of 4 and 13) from Switzerland and Belgium. Parents completed the PS and a measure of child behavior, the Strengths and Difficulties Questionnaire (SDQ; Goodman, 1997). Eight different factorial structures were identified in the literature and tested with Confirmatory Factor Analyses (CFA). The results indicate that the two-factor model — laxness and over-reactivity — proposed by Irvine et al. (1999) best fitted the data. Indeed, it led to the lowest values of RMSEA and SRMR (1999). Furthermore the TLI and CFI indices were greater than .90, supporting a satisfactory fit of the data. Results from multi-group CFA analyses indicated that this factor structure did not vary across child sex and age, or parental sex. In this French version, laxness and over-reactivity had good internal consistency when analyzing either the total sample, or when separating the data according to the sex of the parents or the child's age group. It was possible to propose normative values of laxness and over-reactivity for mothers and fathers, as well as according to the age of the child, which will be useful for clinicians assessing parenting discipline practices of their clients. The results indicated that mothers rated themselves as more over-reactive than fathers (t(809) = 2.89, P < .01, d = .21), but the effect size was small. The difference was non-significant for laxness. The sex and age of the child, as well as the parents’ age did not affect over-reactivity and laxness. However, a weak link was found between education level and over-reactivity, but only for fathers (r = .19, P < .01). The association between the two factors and the child's adjustment were also evaluated. The highest correlation appeared to be between over-reactivity and the Total Score on the SDQ (r_s = .28, p < .001). Correlations between over-reactivity and externalized problems (behavior and hyperactivity) were higher than those obtained with internalized problems. As for laxness, the correlations were lower: the highest one was with behavioral disorders (r_s = .15, p < .001). Over the entire sample (R² = .09, β = .27, P < .001) and in fathers (R² = .10, β = .32, P < .001), over-reactivity was the best predictor of the child's difficulties regardless of the sex of the child. In mothers, both factors predicted the child's difficulties (R² = .08, P < .001), with over-reactivity explaining more variance of total SDQ (β = .22, P < .001) than laxness (β = .14, P < .05). Over-reactivity appears to correspond to the construct of “coercive discipline” theorized and intensively studied by Patterson and colleagues (1982), which represents parents’ tendencies to criticize their child and to engage in severe disciplinary interactions. In summary, the Parenting Scale provides a quick and inexpensive measure of dysfunctional discipline practices. In contrast to more comprehensive measures of parental attitudes and beliefs, the PS items are specific enough to target concrete behaviors regarding parenting interventions. The PS can therefore be particularly useful in planning cognitive and behavioral therapy, as well as in evaluating the effects of treatment.