Background - The Fear of Negative Evaluation scale (FNE: Watson and Friend, 1969) is the measure most commonly used to determine the degree to which people experience apprehension at the prospect of being negatively evaluated. Although the development of the FNE preceded the inclusion of social anxiety disorder (or social phobia) in the diagnostic classification system, itis widelyusedas a measure of cognitive symptoms because the feature tapped by this measure is at the core of recent cognitive models of social anxiety. According to these models, socially anxious individuals divide their attention between the internal representations of their social self (negative images and, felt sense,) and external cues that could be taken as a sign of negative evaluation by others. The FNE was validated in a student and patient population in English speaking countries. The English version demonstrates adequate empirical validity. It shows excellent internal consistency and one-factor structure. Test retest reliability is satisfying. It is proved to be sensible to change after treatment. It is well correlated with other measures of social anxiety demonstrating good convergent validity. However, divergent and discriminate validity have been a subject of controversy. The aim of the present study was to assess the psychometric properties of the French version of the FNE in order to obtain a valid instrument measuring the cognitive component of social anxiety. Method - The social anxiety group consisted of 88 patients referred to our clinic for cognitive-behavioural group therapy. All met ICD-10 criteria for social phobia (generalized subtype) as determined by the Composite International Diagnostic Interview (social phobia section). Additional axis-I diagnostic information was obtained using the Mini Intemational Neuropsychiatric Interview (MINI: Lecrubier et al., 1997) for ICD-10. The non-patient control group consisted of 80 participants who didn't respond to social anxiety CIDI criteria. All participants were rated by the assessor on the Liebowitz Social Anxiety Scale (LSAS) and completed the following self-report questionnaires: the Fear of Negative Evaluation questionnaire (FNE), the State-Trait Anxiety Inventory (STAIA-B), and the Beck Depression Inventory (BDI 13). Results - The internal consistency is very good with Kuder-Richardson 20 coefficient of 0,94 corresponding to coefficients found in the validity studies of the original version of the FNE. One factor Anova tests showed that our two groups of subjects differed significantly on the FNE [F (1 - 166) 282.26, p < 0.001]. Patients had a mean score of 25.6 with a standard deviation of 4.1 and the non patient control subjects had a mean score of 12.1 with a standard deviation of 6.1. This result confirms the empirical validity of the instrument. In the patient sample, the FNE was significantly correlated with. the LSAS total score (r= 0.55;p < 0.001). This result indicates good convergent validity. However, the FNE was also significantly correlated with the BDI-13 (r = 0. 48; p < 0.001) and the STAI B (r= 0.47; p < 0.001). However because high levels of general anxiety and depression are common among patients with social anxiety disorder, substantial correlations between measures of social anxiety and measures of general anxiety and depression should be probably expected. A multiple regression analysis shows a significant association of the FNE with the LSAS (Beta = 0.39, p < 0.001) and the STAIB (Beta = 0.28, p < 0.01) We explored the unidimensionality of the scale by using a principal component analysis of tetra-choric correlation suitable for dichotomic items. One sole factor with an eigenvalue superior to 1 emerged. We deduce therefore that our results are in favour of the unidimensionality of the French version of the FNE. Conclusion - The present study shows that the French version of the FNE has good psychometric properties and differentiates social phobic patients from the non-clinical control subjects. Given the importance that models of social anxiety attribute to cognitive processes and the importance of cognitive techniques in its treatment, we consider that the French version of the FNE is an adequate and valid questionnaire to be used in research and therapy.
