The research focused on 2,533 postal workers in Boston and on the average costs for the postal service in Boston and nationwide. Results revealed that drug screening would have saved the postal service $162 per applicant hired. However, these results were sensitive to the assumptions in the model. If the prevalence of drug use in the population screened were 1 percent rather than 12 percent, the program would lose money. Similarly, if the cost per urine sample screened were $95 rather than the $49 assumed, the program would lose money even if the prevalence of drug positives was as high as 9 percent. Findings indicated that because of the sensitivity of the analysis to changes in its underlying assumptions, any company considering preemployment drug testing should carefully assess the costs and benefits in its own industry. Tables and 14 footnotes