Predicting Employee Absenteeism Using Tobit Regression and OLS Models

The results of predicting employee absenteeism based an Ordinary Least Square (OLS) has been criticized. The present research compared the results from predicting absenteeism based on four varibales of job involvement, ab- senteeism compared to others, stress from life events, and number of children with the OLS and Tobit regression mo- dels. A sample of 197 male employees of factories in Esfa- han city mean age: 28, SD: 10 responded to the Life Eve- nts Stress Inventory (Thoits, 1981), The Absenteeism Com- pared to Others Inventory (Baba, 1990), and the Job In- volvement Inventory (Lodahl & Kejner, 1965). Tobit mo- del analysis predicted employee absenteeism more precis- ely and explained a furhter 13% of the variance (R2 = 42.1) than did OLS.