Objectives The objectives this study were (1) to investigate correlations between measures of psychosocial workplace stress as measured in separate years by the Job Content Questionnaire (JCQ) and Effort–Reward Imbalance (ERI) scales; (2) to establish a valid measure of psychosocial job stress with its components (by identifying the individual and interactive associations of job stress components) and (3) to use the component measures to assess the risk of psychosocial strain at work on fatigue.

Methods The JCQ and ERI from the annual survey of the GAZEL cohort established in 1989 initially with 20 624 respondents were used to investigate the associations of workplace stress on mental and physical fatigue in two separate years (1998 and 2006). First, the JCQ measures from separate years (1997 and 1999) were combined to create a measure for the same year as ERI (1998). The new measure was validated for internal and external consistency. Using logistic regression, the subcomponents of stress (upper tertiles of psychological demands, physical demands, decision latitude, social support, effort, reward, ERI and overcommitment) were tested for associations with the highest reporting of mental and physical fatigue.

Results By combining JCQ responses from 1997 to 1999, we were able to increase the amount of information available on psychosocial factors in 1998. Psychometric properties of the workplace stress scales also showed expected factor loadings. Workplace psychosocial factors had greater associations with fatigue among men than women. Although psychosocial factors became less predictive of fatigue at 8 years of follow-up, associations between fatigue and psychosocial components (overcommitment, social support and rewards) remained significant.

Conclusions These analyses continue to validate the various subcomponents scales of workplace stress as measured by the JCQ and effort–reward imbalance model in GAZEL. They also highlight the importance of psychosocial work factors in the experience of overall fatigue even after an 8-year follow-up.