tion program to inform parents, teens, educators, and employers about the health and safety concerns of young workers to initiatives to update the state's child labor laws.
State-level data on the extent of work by young people and on work-related injuries could be helpful in targeting inspections and studying injury rates. Many states oversee health and safety inspections of workplaces, but do not have the resources to mount a surveillance system, so they need better information to target those inspections.
Information from workers' compensation reports and claims has been used to analyze injuries among workers under the age of 18 in several states, including California (Bush and Baker, 1994), Connecticut (Banco et al., 1992), Massachusetts (Brooks and Davis, 1996), Minnesota (Heinzman et al., 1993; Parker et al., 1994a), New York (Belville et al., 1993), Texas (Cooper and Rothstein, 1995), and Washington (Heyer et al., 1992; Miller, 1995). A study in Utah looked at work-related burns, using both data from a hospital burn center and the State Insurance Fund's industrial records (Hayes-Lundy et al., 1991). Emergency-department data in Massachusetts (Brooks et al., 1993), other medical records in Wisconsin (Stueland et al., 1996), and survey methodology in Minnesota (Parker et al., 1994b) have also been utilized in studying injuries of young workers (see Table 3-2, above).
The drawbacks to emergency department data have already been noted. Workers' compensation records also have some drawbacks. In no state are all employees covered; small-business employees, farm laborers, domestic servants, and casual employees are frequently excluded from workers' compensation. The self-employed are also excluded, and children working informally for family businesses are unlikely to be covered. The number of lost workdays required to qualify for workers' compensation payments differs by state, so comparisons of injuries across states are difficult. In addition, in many states, claims for medical treatment alone are not included in the computerized datasets, leaving only claims for lost worktime (indemnity claims) available for analysis. This exclusion may eliminate a large proportion of the cases. Between 1988 and 1991, for example, 76 percent of the claims for injuries to children