This study demonstrates that readily available socioeconomic data routinely collected by the US Census can be used to estimate the incidence of burn injuries within the 66 counties of the populous six-state New England region of the USA. The burn data were collected during the National Burn Demonstration Project and included New England residents admitted for hospital care of burns sustained between 1 July 1978 and 30 June 1979. Linear regression analysis revealed strong associations between calculated burn rates and a number of socioeconomic variables. Associations with five such variables are described, including per capita income, percentage of persons below poverty level, percentage of residences built prior to 1940, percentage of adults with 16 years or more of education, and percentage of persons moving since 1975 with previous residence in the same county. Estimates of burn incidence for counties, together with a previously reported study at the level of census tracts for a major Standard Metropolitan Statistical Area, can be used to reduce the time and cost of burn injury case reporting by health care providers or case-finding efforts for large population groups, and can be used to predict the effectiveness of social and economic programmes and policies that improve the overall well-being of county populations.