Les objectifs de cette étude sont de tester de nouveau les relations entre d'une part le stress intrinsèque et extrinsèque et d'autre part, l'absence et l'assiduité au travail et de déterminer la nature des relations entre le stress au travail et différentes mesures de fréquence des absences et de temps perdu.
An earlier study by two of the authors of this paper had shown intrinsic stress to be negatively related to absence frequency while extrinsic stress is positively related to absence behavior (Arsenault and Dolan, 1983a). Whereas stress and absence patterns has been already demonstrated in a number of studies, the unique phenomenon of «assiduity» has not received sufficient attention in the literature. Moreover, some of the research, has been plagued with methodological difficulties, namely associated with the operational definition of absenteeism.The major objectives of this study are two fold: a) to confirm (test-retest) the relationships between intrinsic and extrinsic stress on «assiduity» and absence patterns respectively; and b) to determine the relationship between job stress and a number of alternative measures of absence.Daily attendance records and absence by cause for an entire year (1984) were collected for a sample of 162 regular hospital employees. Extrinsic and intrinsic job stress indices were calculated based on linear addition of multi-item multidimensional measures derived from questionnaires. Full psychometric description of all instruments have been detailed elsewhere (Arsenault and Dolan 1983a, 1983b). A number of different measures of absenteeism have been studied. They include five different measures of «frequency» and two measures of «time-lost». Multiple Correlations and multiple regression analyses were used to test the hypotheses.Results confirm previous findings to the extent that intrinsic job stress is significantly related to an increase in assiduity and extrinsic job stress to an increase in absenteeism. The study of differents measures of absence has shown that job stress is significantly related to frequency indices, but not to time lost.
