Methodological and theoretical issues in cross-national comparative studies of psychosocial job hazards: From questionnaire item to social class

There has been little concern, among international work and health researchers, for methodological and theoretical issues that are critical for the validity of international comparison studies of perceived psychosocial job hazards. In this study, two methodological issues were addressed: internal and external consistencies of the five psychological demands items of the Job Content Questionnaire (JCQ) as a scale for practical use; differential understanding of JCQ items among different cultures and its impact on the scale-level cross-cultural mean comparisons. One important theoretical debate was also addressed: the relationship between socioeconomic status (SES) and job strain of Karasek's Demand-Control model. Multidimensional structures of the five items were identified. In some sub-samples of a dataset, several significant correlations of the components of the psychological demands scale with job dissatisfaction and life dissatisfaction were obscured by the full scale. On average, 34.0% to 39.5% of the total tested items turned out to be understood differentially between cultures. The psychological demands and supervisor support scales of the JCQ were the most affected. The impact of the differential item functioning on the scale-level mean comparisons was minor generally, but some of the findings have implications for a robust international comparison. The orthogonal relationship between SES and job strain was generally supported, if not in all of the datasets tested. The proportion of variance in minor mental disorders explained by job strain (on average, 7% for males and 5% for females) was much higher than that by SES (less than 1%). However, significant indirect effects of SES on mental disorders via psychosocial job hazards were also identified in some female samples. A more robust theoretical framework of work demands should be developed. A stricter translation process needs to be employed for effective cross-cultural adaptation of JCQ items. More diverse SES indicators and more objective psychosocial job characteristics/health outcomes in representative and longitudinal datasets need to be employed for the debate. Nonetheless, this study implies that work quality policy could bring major improvement of daily worker's mental health in the working populations of industrialized countries.