This study will develop standardized methods to measure the impacts of workplace employment and earnings that can be used in a variety of settings where different data are available. These measures allow more effective targeting of prevention resources. The study also will quantify the impact on earnings when employers promote rehiring of injured workers, and it will measure the differential impact of workplace injuries and illnesses on men and women and on younger and older workers. Finally, it will develop methods for comparing average lost earnings among states that differ in important ways - with disparate industrial mixes, unemployment rates, and so on. This should improve our understanding of how interstate variation in laws, practices, and regulation affect consequences of workplace injuries and illnesses. To accomplish these goals, we will use both parametric and nonparametric statistical methods designed to provide unbiased measures of the impacts of interventions (which, in this case are the workplace injuries and illnesses). This study will compile and analyze individual longitudinal data on more than 500,000 injured workers in five states. The primary data on injured workers in each state will consist of workers' compensation administrative records on all lost-time injuries and unemployment compensation longitudinal quarterly earnings and employment data covering at least 7 consecutive years. Comparison groups will include workers with very short-term injuries, workers with workers' compensation claims involving only medical payments, and uninjured workers matched on preinjury employer and preinjury earnings. We will test the sensitivity of our results to the choice of statistical method and comparison group.
