The age decrement model has traditionally been a fairly popular model of the human life span. This model has the basic premise that with increasing age there is a corresponding decrease in a wide range of abilities. Not all research has agreed with the unqualified age decrement model. This study examined the relationship between an information-processing based test battery, job performance, and age using a sample of petroleum transport drivers (N=71). Only the Auditory Selective Attention Test (ASAT) and the Group Embedded Figures Test (GEFT) were valid predictors in the information processing battery; they were significantly correlated with the number and cost of driving accidents and the cost of spills and blends. Older drivers performed worse than younger drivers on the ASAT and the GEFT, the two valid predictors. However, there were no significant age differences in the costs of both driving accidents and spills and blends. The results suggest that an increasing number of driving accidents may have resulted from factors other than age. The results also suggest that when age differences in job performance occur, they are better explained by performance on valid predictors than by age per se. Thus the results bring into question the use of age as a surrogate predictor of job performance. (Author/ABL)