Psychometric properties of the French version of the Effort-Reward Imbalance model

Background: Two main models are currently used to evaluate psychosocial fac tors at work: the Job Strain model developed by Karasek and the Effort-Rewa rd Imbalance model. A French version of the first model has been validated for the dimensions of psychological demands and decision latitude. As regar ds the second one evaluating three dimensions (extrinsic effort, reward, an d intrinsic effort), there are several versions in different languages, but until recently there was no validated French version. The objective of thi s study was to explore the psychometric properties of the French version of the Effort-Reward Imbalance model in terms of internal consistency, factor ial validity, and discriminant validity. Methods: The present study was based on the GAZEL cohort and included the 1 0 174 subjects who were working at the French national electric and gas com pany (EDF-GDF) and answered the questionnaire in 1998. A French version of Effort-Reward Imbalance was included in this questionnaire. This version wa s obtained by a standard forward/backward translation procedure. Results: Internal consistency was satisfactory for the three scales of extr insic effort, reward, and intrinsic effort: Cronbach's Alpha coefficients h igher than 0.7 were observed A one-factor solution was retained for the fac tor analysis of the scale of extrinsic effort. A three-factor solution was retained for the factor analysis of reward, and these dimensions were inter pretated as i) esteem, ii) job promotion and salary and iii) job instabilit y. The factor analysis of intrinsic effort did not support the expected fou r-dimension structure. The analysis of discriminant validity displayed sign ificant associations between measures of Effort-Reward Imbalance and the va riables of sex, age, education level, and occupational grade. Conclusion: This study is the first one supporting satisfactory psychometri c properties of the French version of the Effort-Reward Imbalance model. Ho wever, the factorial validity of intrinsic effort could be questioned. Futh ermore, as most previous studies were based on male samples working in spec ific occupations, the present one is also one of the first to show strong a ssociations between measures of this model and social class variables in a population of men and women employed in various occupations.