Krieger et al. paper is so timely for me as my current research project is actually related to discerning the role of census-tract-level factors vs. individual level factors in the observed urban-rural disparities in cervical cancer outcomes. Moreover, I will be looking at interactions of various factors in recognition of the importance of contrasting intersections of risk factors/experiences, e.g. such as race/ethnicity-income intersection mentioned in the paper by Krieger et al. So, I'm happy that my research will address both these issues -- using census-tract-level data rather than cities/counties and studying combined experiences rather than each separately.
It also got me thinking that it's not just the steeper gradient that we might observe in the estimates when using measures that combine indicators of inequalities, such as race/ethnicity and income, but we could also have some interesting and maybe even unexpected findings. Like in a recent paper on the association between neighborhood-level redlining and lending bias with breast cancer mortality, the researchers observed that NHW residing in areas of redlining fared worse in terms of breast cancer mortality vs. NHB living in areas of redlining. One of the explanations that researchers suggested was that NHB could have become more resilient and might have benefited from the protective community effect given that they (NHB) comprise the majority living in redlined neighborhoods vs. NHW. Anyway, studying joint segregation/inequality effects can lead to interesting results!
A related comment to the Jones et al. paper. I wonder what we would see if we contrasted individual and neighborhood level factors. For example, what's the difference in individual air-pollution exposure between Asians living in an Asian-overrepresented neighborhood vs. Asians living in a White-overrepresented neighborhood. I'm not sure if authors looked at such combinations. Again, to the point how interesting it is to contrast joint effects.