The objective was to examine the relationship between injury rates and socioeconomic factors for children in Hamilton County, Ohio, using small-area analysis. The subjects were county residents less than 15 years old who were hospitalized or died of injuries between January 1, 1993, and December 31, 1995; they were identified through a population-based trauma registry. The census tract was the unit of analysis; the rate of injury per 100,000 population was the dependent variable. Risk factors included median income, level of education, percentage below the poverty level, percentage unemployment, percentage non-Caucasian, and percentage families headed by females. There were 2,437 children meeting the case definition; injuries per census tract ranged from 0 to 2,020.2 per 100,000 per year. Census tracts with higher injury rates had lower median incomes, more people with less than a high school education, more unemployment, more families headed by females, more people living below the poverty level, and more non-Cancasians han those with lower rates. In a regression model, percentage of people living below poverty level, percentage of those who did not graduate from high school, and percentage unemployment were significant risk factors for injuries, P<.001. Since small-area analysis examines associations on an ecological level rather than an individual level, these studies should always be interpreted with caution because an association found at the level of the census tract may not apply at the individual level. Inverventions to reduce injuries should target socioeconomically disadvantaged children living below the poverty level and those in areas with fewer high school graduates and more unemployment.