The study of absenteeism by the partial correlation and multiple regression methods

Abstract

A study of morbidity with temporary work incapacity as recorded over a period of 2 yr among 2,000 workers of a textile factory is reported. An analysis of the days and causes of incapacity failed to demonstrate any significant difference between empirical distribution and the negative binominal distribution. Factorial analysis was used to define uniform groups of explanatory variables. The explanatory variables studied were: yl, number of days of absence per subject per yr and y2, number of instances of absence per yr per subject. Factorial analysis brought to light 4 'factors', i.e. groups of variables of similar nature and with close internal correlations. They are: the demographic factor: age, total duration of employment and the actual place of employment in the factory; the family load factor: number of children under 15 yr of age, marital status, number of persons in the family and family income per capita; housing conditions factor: average salary per employee, number of rooms, kind of building and the individual feeling on the housing conditions. The explanatory variables failed to show linear correlation with the sick absenteeism indices but a statistical correlation of another type is likely and it is also possible that the correlation would be made evident by larger samples.