This study reports on research which looks for employee and job characteristics which correlate with absenteeism. A large cross-sectional national probability sample of workers employed for at least 20 hr per week is analyzed (n = 1308). The dependent variable is the number of self-reported absences during the past 14 days. Thirty-seven independent variables are considered. Ordinary Least Squares (multiple regressions), two-limit Tobits, and two-part models are used to assess the statistical and practical significance of possible covariates. Statistically significant predictors included health variables such as being overweight, complaining of insomnia, and hazardous working conditions; job characteristics such as inflexible house; and personal variables such as being a mother with small children. Variables reflecting dangerous working conditions appear to be the strongest correlates of absenteeism. Notable variables which do not predict absenteeism include age, race, wages, and job satisfaction. Future research should direct attention toward workers' health and working conditions as covariates of absenteeism, since they are strongly significant in this study and have been neglected by most absenteeism investigators.