injury and illness rates (Federal Register, 1997). This information is similar to that required by the BLS; however, the new rule requires incidence rates rather than just absolute values and also requires that the employer be identified. BLS reporting is, on the other hand, confidential. OSHA sought this surveillance capacity in order to target its enforcement actions to the most hazardous work sites.

Through its Office of Training and Evaluation, OSHA had a training budget of $2.4 million in FY 1997. This office administers training grants to safety and health organizations, employer associations, labor groups, and educational institutions. Grants are geared to employers and employees who are in industries or establishments with significant injuries or hazards. In addition, OSHA offers safety and health courses through the OSHA Training Institute and Education Centers. The Training Institute has outreach education centers in each OSHA region of the United States. The Training Institute fulfills part of OSHA's responsibility to oversee all aspects of health and safety programs for federal employees.

OSHA's broad standard-setting authority leaves the agency with a great deal of discretion. Not surprisingly, the agency's approach has varied with changes in presidential administrations, and the efficacy and cost-effectiveness of OSHA health and safety standards have been subject to considerable debate. Critics have claimed that the existing standards tend to regulate trivial risks with unnecessary specificity in a one-size-fits-all mode, with little significant impact on injury rates. They argue that OSHA regulations should be subject to cost-benefit analysis (Viscusi, 1983; Mendeloff, 1979) or that safety standard setting should be abandoned in favor of an ''injury tax," leaving it to employers to take safety precautions to minimize their tax liability (Nichols and Zeckhauser, 1977). OSHA's defenders argue that neither of these proposals will adequately protect worker safety and that OSHA should be strengthened so that it has adequate resources to regulate more effectively (McGarity and Shapiro, 1993).

Outcome studies of OSHA regulation have been inconclusive. Overall, the most reasonable assessment is that OSHA safety regulation has probably had a modest effect in reducing workplace injuries against the general backdrop of a long-term decline (Dewees et al., 1996). Evaluation of OSHA reform proposals lies outside the scope of the present study. However, a key element of any reform strategy is to strengthen the information systems on which employers and the agency rely to identify hazards and make decisions about conducting inspections, proposing corrective action, and standard setting. The challenge is to