Perceived psychosocial and biomechanical exposures, individual factors and pain in the shoulder and neck were recorded in two groups of female service workers (healthcare and shopping centre workers). The jobs investigated were characterized by 'much' direct human relations, 'little' sitting and 'much' standing, and were light work by physiological or biomechanical criteria but potentially psychosocially demanding. A screening survey (n = 400 females) was the basis for the selected sample (n = 66 females), which was the object of the main investigation of this study. Reliability of the questionnaires was tested in a separate group of female healthcare workers (n = 29). Heart rate recordings through the work day estimated workload. There was a high prevalence of shoulder and neck pain (> 50%) for both work groups. In the two populations it proved difficult to explain shoulder and neck pain by reported physical and psychosocial exposures or individual factors, except by the variable 'perceived general tension', which clearly differentiated workers with and without pain. The findings in this study indicated, first, that perceived general tension might be an independent risk factor for muscle pain and, second, that this might be related to personality factors. However, this putative relationship must be verified in a longitudinal study. As no variable describing exposures in the working environment was associated with shoulder and neck pain, the question is posed whether such complaints can be considered work-related. Alternatively, the variables used to describe mechanical and psychosocial exposures in this study may have low specificity in characterizing work-related risk factors for service workers with customer relations.