Social and Economic Well-Being by Race and Hispanic Origin

Fellows, Jeffrey jmf9 at cdc.gov
Mon Oct 5 11:19:00 PDT 1998


A group of us in my division have been examining this issue (as it related specifically to violence) for the past few months, and the consensus coming out of it is that low SES trumps race/ethnicity. It also seems to be a consensus in overall health as well. Most studies I've seen that find race significant have usually operationalized SES poorly, e.g., they used education or housing density only. So there is a good deal of error that may partially be captured by a race variable. The housing density one is important if residential segregation exists.

Most of the debate is centered on which SES measure matters most (income inequality, concentrated poverty, low occupational status->segmented labor market theory, etc.), and how low SES translates into interpersonal violence. I think the study you mention (I don't know for sure) may not have controlled for income inequality between racial groups. In other studies, the inclusion of a racial income inequality measure is meaningful and an independent race variable drops out (becomes nonsignificant). The ways in which low SES translates into interpersonal violence is important, and reflects the complexities of the problem and the richness of the literature. What the JAMA article may have presented is the community effects on ill-health associated with low SES. It is common to see risk factors associated with individuals or racial groups that are really intermediaries between SES and adverse health (e.g., high blood pressure, diabetes, etc.) There is at least one condition (sickle cell) I think affects African Americans and not European Americans, but I don't know if this is accurate given possible ambiguities of racial categories; SES may matter here as well?

I don't have time to elaborate more now. I will keep folks updated on any significant results. We are also looking at other social factors related to violence perpetration and victimization, including social capital, collective efficacy, social networks, social norms, and the ways these concepts overlap and/or interact.

I hope I answered this sufficiently. I'll look for the article in my files. If there is something in it that I didn't address I will send an addendum.

Jeff

Doug wrote:

Fellows, Jeffrey wrote:


>>Doug, thanks for the post. From what I have seen of the research on
>>socioeconomic status and health (including violence-related morbidity and
>>mortality), the race/ethnicity dimension of health disappears as a
>>statistically significant relationship once SES is controlled for. So a
>>report on the social and economic well-being by social class would have
been
>>much more accurate. Of course, I am preaching to the congregation here.


>There isn't much data on this sort of thing broken down by SES rather than
>"race," is there? Though I do recall an article in the New England Journal
>of Medicine in the early 1990s that showed that even after controlling for
>income and education, blacks on average had worse health outcomes than
>whites. Is there any kind of consensus on this in the literature?


>Doug



More information about the lbo-talk mailing list