Sensitivity analyses
Our first sensitivity analysis was to rerun our segregation analyses with varying degrees of freedom for the natural cubic splines and inspect differences in the resulting shapes of the smooths. We then reran our segregation analyses with an alternative segregation measure, the index of concentration at the extremes (34, 35). This is a measure of social polarization used in numerous environmental, social, and health science studies to assess the impacts of spatial social stratification. The index of concentration at the extremes was calculated as:
\(\text{ICE}_{i}=\ (D_{i}-A_{i})/\text{Total}_{i}\) (2)
where D is the number of disadvantaged people, A is the number of advantaged people, and Total is the total population of census tract i . We repeated the regression four times, using Asian, Black, and Latino alone as the disadvantaged groups, and then all except white as the disadvantaged group.