Separate black and white infant mortality models: differences in the importance of structural variables.

In the United States, the disparity in black and white infant mortality persists despite reductions in overall infant mortality. In 1988, 23 of the 50 states had a black infant mortality rate that was more than twice as large as its white infant mortality rate. This study assesses whether state-level structural variables relate differentially to states' black and white infant mortality rates. With the state as the unit of analysis, separate black (N = 34) and white (N = 50) multivariate models of infant mortality were produced and compared. The structural variables accounted for 57.7% of the variance in states' black infant mortality rates and 35.2% of the variance in states' white infant mortality rates. Proportion black, percent with bachelor's degree or higher, percent below poverty, and the index of dissimilarity each made a unique contribution to the black infant mortality model. Percent with bachelor's degree or higher was the only measure that made a significant unique contribution to the white infant mortality model. Thus, although both black and white infant mortality rates were higher in states with smaller percentages of the population having a bachelor's degree or higher, black infant mortality rates were also higher in states where proportionately more black persons lived, where there were higher levels of residential segregation in the urban areas, and, contrary to what was expected, where smaller percentages of the population lived below the poverty level. This study supports the need for race-specific models of infant mortality.