Neighborhood racial/ethnic concentration, social disadvantage, and homicide risk: an ecological analysis of 10 U.S. cities

Abstract

Homicide is one of the leading causes of death among African-American and Hispanic men. We investigated how neighborhood characteristics associated with social disadvantage explain racial/ethnic homicide gaps in 10 U.S. cities. The test hypotheses were that (1) higher concentrations of African-Americans and Hispanics would be associated with higher homicide rates and (2) the relationship between racial/ethnic concentration and homicide would be attenuated after adjusting for neighborhood characteristics (e.g., unemployment, median household income, low educational attainment, and female headship). The test hypotheses were examined using separate Poisson regression models, which adjusted for spatial autocorrelation. Homicide rates were greater in neighborhoods with higher concentrations of African-Americans and Hispanics than in other groups, and the association of neighborhood racial/ethnic concentration with homicide was reduced after adjusting for neighborhood social disadvantage variables, especially percent female head of household and percent persons with less than a high school education. We also found that the relationship between neighborhood racial/ethnic concentration and homicide was explained more by social disadvantage variables in some cities than in others. Based on our findings, policy makers may wish to consider implementation of policies that (1) expand early childhood education programs and higher education opportunities and (2) encourage economic and community development initiatives in socially disadvantaged neighborhoods.

FIGURE 1.

Homicide rate by percent African-American and percent female headship.

FIGURE 2.

Effects of having 10% higher percentage of African-Americans in a particular census block group on homicide mortality rate within each of the 10 cities and overall. The “unadjusted” results only take into account background variables, year, mean homicide of spatially adjacent neighborhoods, percent males aged 14 to 24 years of age, and area in square miles. Adjusted results also included covariates: percent persons with less than a high school education, percent female head of household, median household income, and unemployment rate as well as background variables. Boxes and lines represent log relative risk and 95% confidence intervals, respectively.

FIGURE 3.

Effects of having 10% higher percentage of Hispanics in a particular census block groups on homicide mortality rate within each of the 10 cities and overall. The “unadjusted” results only take into account background variables, year, mean homicide of spatially adjacent neighborhoods, percent males aged 14 to 24 years of age, and area in square miles. Adjusted results also included covariates: percent persons with less than high school education, percent female head of household, median household income, and unemployment rate as well as background variables. Boxes and lines represent log relative risk and 95% confidence intervals, respectively.