Assessing the impact of SNAP-ed census-tract targeting on access to nutrition programming for low-income individuals

Introduction: One method for locating poverty geographically is census tracts where >50% of residents are below a poverty threshold. USDA has used this approach to define geographic limits for SNAP-Ed programming. Many states report inability to reach low-income audiences effectively using this criterion; USDA has indicated willingness to consider other approaches, but requires states to propose new criterion. Purpose: Document impact on reach of limiting programming to census tracts with <50% of population <185% of federal poverty level; demonstrate an alternate methodology for identifying highest-poverty areas within a state. Methods: Data on individuals in poverty and poverty rate was retrieved for all US census tracts, and analyzed to determine percentage of low-income population reachable through tracts with <50% <185%FPL. GIS-maps were developed to show geographic trends, and correlations between population size, density, and urbanicity examined. An alternate methodology was developed for Hawai‘i, clustering census tracts within locally-recognized communities, determined proportion of state's low-income population within each community, and ranking these results. Results: Census tract targeting reached 27% of low-income individuals nationally, with wide disparities between states -- 3% in Vermont to 43% in Mississippi/Texas. Reach was correlated with states' poverty level (r=.661 , p<.001) and population size (r=.424 , p=.002) In Hawai‘i, census tract methodology reached 6% of low-income residents. The alternate methodology increased reach to 50%, by targeting only 10 communities. Discussion: Targeting low-income individuals by census tract alone may result in poor reach; targeting by proportion of state's low income population located within communities can improve reach substantially.