Longitudinal examination of the stability and variability of two common measures of absence

This study, conducted amongst 121 employees in two hospitals, examines sources of variability in two common measures of absence: frequency and time-lost. Time series analyses over a five-year span suggest substantial differences by month, season and year. More specifically, the analysis of variance of the frequency measure reveals significant differences between seasons and years and a significant season by year interaction. Peak frequencies are systematically recorded during the winter season, while the lowest occur in summer. These trends are less prominent with the time-lost measure, with only the seasonal differences remaining significant. The paper concludes with a discussion of the implications of these findings for the study and measurement of absenteeism.