Many studies have shown that the strategies used to cope with chronic pain play a very important role in the adjustment to the pathology and to its effects (emotional distress, physical and psychosocial impairment, and quality of life). Among the methods assessing coping with pain, the most widely used instrument at present is the coping strategies questionnaire (CSQ) developed by Rosenstiel and Keefe, [Pain 17 (1983) 33-44]. This questionnaire is composed of 48 items distributed in eight subscales each including six items: diverting attention, reinterpreting pain sensations, coping self-statements, ignoring pain sensations, praying and hoping, catastrophizing, increasing activity level, and increasing pain behaviour. Most studies examining the factor structure of the CSQ have used the scores of its eight prior theoretically derived scales rather than the 48 items. Three studies, Tuttle et al. [Rehab Psychol 36 (1991) 179-187], Swartzmann et al. [Pain 54 (1994) 311-316; Robinson et al. [Clin J Pain 13 (1997) 43-49] have examined the factor structure of the CSQ from the 48 original items on the questionnaire and have yielded five or six factors. A structural confirmatory analysis showed the superiority of the six-factor model [Clin J Pain 13 (1997) 156-162]: distraction, catastrophizing, ignoring pain sensations, distancing from pain, coping self-statements and praying. The present study aimed at measuring the internal consistency and the construct validity of the French version of the CSQ. The CSQ was translated into French with the forward and backward translation procedure. To evaluate internal consistency, Cronbach's alphas were computed. Construct validity of the questionnaire was estimated through confirmatory factor analysis (CFA) in a sample of 330 chronic pain patients (71% of women): 40.3% suffered from low back pain, 33.6% from headaches and 26.1% from neuropathic pain. The three factor structures previously proposed in the literature were tested using the LISREL 8.3 structural equation-modelling program developed by Jöreskog and Sörbom, [Lisrel 8: user's reference guide, Chicago: Scientific Software International, 1993]. The CFA performed on the three models of factor structures of the CSQ previously reported confirms the best fit of the six-factor model by Robinson et al. [Clin J Pain 13 (1997) 43-49] in our sample. However, the coping self-statements factor, whose internal consistency was too weak in our analysis (Cronbach's alpha=0.57), was eliminated. We therefore retained only five factors in our sample. The French version of the CSQ (CSQ-F) is composed of 21 items belonging to five factors: distraction, catastrophizing, ignoring pain sensations, reinterpreting pain sensations, and praying. The present study indicates that the internal consistency and the construct validity of the French version of the CSQ were adequate, and contributes to demonstrate the stability of the factor structure of the CSQ across samples. The 21-item French adaptation of the CSQ (CSQ-F) appears to be a very interesting tool because it facilitates the use of this questionnaire, not only for research but also in the clinical assessment of the patients suffering from chronic pain.