Cet article présente les résultats d’une première étude de validation du questionnaire Émotionnalité positive et négative (EPN-31) dans une population française de patients psychiatriques. Il s’agit d’une échelle d’analyse émotionnelle adaptée par Rolland sur la base des travaux de Diener et du modèle « tripartite » des émotions de Watson et Clark. Trois scores principaux sont fournis par ce questionnaire : les émotions positives, les émotions négatives et les émotions de surprise. Six sous-scores peuvent également être calculés. Un groupe de 400 patients psychiatriques a rempli l’EPN-31 et a été évalué par ailleurs à l’aide de l’échelle HAD (Hospital Anxiety and Depression scale). Une analyse en composantes principales (ACP) a été effectuée, de même que des analyses de consistance interne et de validation concourante. L’ACP identifie une structure à trois facteurs correspondant, après rotation orthogonale Varimax, aux émotions négatives (F1), aux émotions positives (F2) et aux émotions de surprise (F3). Les coefficients de consistance interne varient entre 0,80 et 0,95 pour les trois scores principaux, et entre 0,72 et 0,90 pour les six sous-scores. Le score d’émotions positives est corrélé négativement à la sévérité des symptômes dépressifs, et le score d’émotions négatives est corrélé positivement à la sévérité des symptômes anxieux et dépressifs. Le score de surprise paraît moins robuste. En conclusion, cette première étude de l’EPN-31 confirme la validité et l’intérêt de cet instrument de mesure des émotions en psychiatrie.

Background –This paper reports the first validation study of the EPN-31 scale (Positive and Negative Emotionality scale, 31 items) in a French psychiatric sample. This questionnaire has been adapted by Rolland from an emotion inventory developed by Diener, and is also in accordance with Watson and Clark’s tripartite model of affects. Methodological aspects –Respondents were asked to rate the frequency with which they had experienced each affect (31 basic emotional states) during the last month. The answer format was a 7-point scale, ranging from 1 « Not experienced at all » to 7 « Experienced this affect several times each day ». Three main scores were calculated (positive affects, negative affects, and surprise affects), as well as six sub-scores (joy, tenderness, anger, fear, sadness, shame). Four hundred psychiatric patients were included in this study, and completed the EPN-31 scale and the Hospital Anxiety and Depression (HAD) scale. The Global Assessment of Functioning (GAF) scale was rated, as well as DSM IV diagnostic criteria. We performed a principal component analysis, with Varimax orthogonal transformation, and explored the factorial structure of the questionnaire, the internal consistency of each dimension, and the correlations between EPN-31 scores and HAD scores. Results –The factorial structure of the EPN-31 was well-defined as expected, with a three-factor (positive, negative and surprise affects) solution accounting for 58.2 % of the variance of the questionnaire. No correlation was obtained between positive and negative affects EPN-31 scores (r = 0.006). All α Cronbach coefficients were between 0.80 and 0.95 for main scores, and between 0.72 and 0.90 for sub-scores. GAF scores were significantly correlated with EPN-31 positive affects scores (r = 0.21 ; p = 0.001) and with EPN-31 negative affects scores (r = – 0.45 ; p = 0.001). We obtained significant correlations between positive affects score and HAD depression score (r = – 0.45 ; p ≪ 0.001), and between negative affects score and HAD anxiety (r = 0.56 ; p ≪ 0.001) and depression (r = 0.45 ; p ≪ 0.001) scores. This pattern of correlation was in accordance with the Watson tripartite model of emotionality. Significantly higher EPN-31 positive affect mean scores were observed in females when compared to males (p ≪ 0.001). The third factor of the EPN-31 is less robust than the others and, the validity of the surprise score could hence be discussed. Conclusion –In all, this study confirmed the validity and the interest of the EPN-31 use in psychiatric patients. Various clinical and research applications can be considered, such as infra-symptomatic studies of emotions in affective disorders and during treatment protocols or definition of phenotypic markers in genetic or neuro-imagery studies.