L’objectif de notre étude était d’analyser la structure factorielle de la version brève du Questionnaire de Personnalité Schizotypique (SPQ-B) dans un format Likert. Une analyse en composante principale de la version française de l’échelle a été réalisée dans un échantillon représentatif de la population générale (n=233). La consistance interne et l’influence de l’âge et du sexe sur le score total et les sous-scores dimensionnels ont également été étudiées. L’échantillon était constitué de 131 femmes et de 102 hommes avec un âge moyen de 52,7±18,1ans et un score moyen du SPQ-B de 43,6±13. L’analyse factorielle a montré une solution à 3 facteurs expliquant 47,7 % de la variance, avec un facteur « désorganisation » incluant les items du « comportement bizarre », du « discours bizarre » et de « l’anxiété sociale », un facteur « interpersonnel » incluant les items de la dimension négative et de la « méfiance », et un facteur « cognitif-perceptif » incluant les principaux items de la dimension positive. La consistance interne (coefficient α de Cronbach) pour les 3 sous-échelles et pour le score total était respectivement de 0,81, 0,81, 0,77 et 0,88. Nous n’avons pas retrouvé d’influence de l’âge et du sexe sur les différents scores de schizotypie. Notre étude confirme la structure tridimensionnelle de la schizotypie et montre que la version française du SPQ-B format Likert présente de bonnes qualités psychométriques. Cet outil apparaît comme un instrument fiable et facile d’utilisation pour mesurer l’intensité des traits schizotypiques en population non malade.

The main objective of the study was to explore the factorial structure of the French version of the Schizotypal Personality Questionnaire-Brief (SPQ-B) in a Likert format, in a representative sample of the general population. In addition, differences in the dimensional scores of schizotypy according to gender and age were analyzed. As the study in the general population of schizotypal traits and its determinants has been recently proposed as a way toward the understanding of aetiology and pathophysiology of schizophrenia, consistent self-report tools are crucial to measure psychometric schizotypy. A shorter version of the widely used Schizotypal Personality Questionnaire (SPQ-Brief) has been extensively investigated in different countries, particularly in samples of students or clinical adolescents, and more recently, a few studies used a Likert-type scale format which allows partial endorsement of items and reduces the risk of defensive answers.

A sample of 233 subjects representative of the adult population from an urban area near Paris (Créteil) was recruited using the “itinerary method”. They completed the French version of the SPQ-B with a 5-point Likert-type response format (1=completely disagree; 5=completely agree). We examined the dimensional structure of the French version of the SPQ-B with a Principal Components Analysis (PCA) followed by a promax rotation. Factor selection was based on Eigenvalues over 1.0 (Kaiser's criterion), Cattell's Scree-plot test, and interpretability of the factors. Items with loadings greater than 0.4 were retained for each dimension. The internal consistency estimate of the dimensions was calculated with Cronbach's α. In order to study the influence of age and gender, we carried out a simple linear regression with the subscales as dependent variables.

Our sample was composed of 131 women (mean age=52.5±18.2 years) and 102 men (mean age=53±18.1 years). SPQ-B Likert total scores ranged from 22 to 84 points (mean=43.6±13). Factor analysis resulted in a 3-factor solution that explained 47.7% of the variance. Factor 1 (disorganized; 10 items) included items related to “odd behavior”, “odd speech”, as well as “social anxiety”, one item of “constricted affect” and one item of “ideas of reference”. Factor 2 (interpersonal; 7 items) included items related to “no close friends”, “constricted affect”, and three of the items of “suspiciousness”. Factor 3 (cognitive-perceptual; 5 items) included items related to “ideas of reference”, “magical thinking”, “unusual perceptual experiences” and one item of “suspiciousness”. Coefficient α for the three subscales and total scale were respectively 0.81, 0.81, 0.77 and 0.88. We found no differences in total schizotypy and the three dimensions scores according to age and sex.

Factor analysis of the French version of the SPQ-B in a Likert format confirmed the three-factor structure of schizotypy. We found a pure cognitive perceptual dimension including the most representative positive features. As expected, “Suspiciousness” subscale is included in both positive and negative dimensions, but mainly in the negative dimension. Surprisingly, “social anxiety” subscale is included in the disorganized dimension in our analysis. The SPQ-B in a Likert format demonstrated good internal reliability for both total and subscales scores. Unlike previous published results, we did not find any influence of age or gender on schizotypal dimensions.