ABSTRACT
The 3,018 persons who were in continuous employment in a Health Authority from April 1974 to March 1976 formed the population for the study of independent variables, metric and non-metric, on the two dependent variables, number of spells and days of absence. The independent metric variables were age and length of service and the non-metric ones were sex, marital status, type of job [full or part-time], method of payment [weekly or monthly]. Because analysis of covariance could not determine whether interactions between metric and non-metric variables or between two metric variables had a significant effect, additional information was obtained from a multiple regression analysis with dummy variables. In this study, the effect of age on both the indicators of sickness absence [number of spells and days of absence] is not strong. Other variables rarely reached the 5% significance level, suggesting that any such effect is either very slight or may be due to random variations