There are 8 kinds of people in the U.S.

Those who divide the U.S. into 8 kinds of people, and those who don't, further subdivided by urban and rural residence, and high and low income . . .

You may well have read news accounts of the study by Christopher Murray et al, published yesterday in PLoS Medicine, about disparities in life expectancy among various segments of the U.S. population. The newspaper account I read did not explain it very well, however. It is called Eight Americas. As Tony Snow would put it, "There's nothing new here," and it is in fact an extended exercise in the reification of constructs. But it still manages to be instructive, at least as a way of dramatizing what we already know.

Because they needed historical data going back a few decades, they used the 1977 "OMB statistical policy directive 15" standard for classifying the population by race, which means people get to be White, Asian, Native American, or Black. You can't be Latino, or anything else, such as Brazilian or Haitian or Somali or Assyrian. In fact, based on more recent data, the authors classified Asians living in places where 40% or more of the Asian population by the old standard is in fact Pacific Islander (Hawaii and Alaska, basically) as part of an amorphous Middle America, which is mostly white.

Anyway, starting with this 19th Century pseudo-scientific racist system, they went on to look at broader socio-demographic characteristics of counties and, based on what felt good at the moment, divided us all into 8 categories: Asians who don't live near a lot of Pacific Islanders; White people in low income rural counties of the northern plains; White people in Appalachia and the Mississipi valley who live in low-income counties; all other White people; Black people living in low-income counties of the South; Black people living in urban counties with high homicide rates; Native Americans in the Mountain and Plains states, mostly living on or near reservations; and everybody else, who is collectively called Middle America.

All this may strike you as arbitrary and odd, even vaguely offensive. Whatever. They then calculated life expectancy and mortality rates at various ages for these categories. As you know from your assiduous, dedicated reading of this blog, life expectancy is also an elaborate construct that does not necessarily represent any actual individual person's chances of living to any actual particular age. Nevertheless it does provide a snapshot of the condition of sub-groups of the population at a given moment.

The payoff is that disparities in life expectancy among the various "Americas" (and sorry if you don't happen to be in any of them, if you don't exist I guess your life expectancy must be zero) are as great as some of the most dramatic international comparisons. For example, the Asian females of America I have a live expectancy 12.8 years longer than the rural Black women of America 7 -- similar to the gap between women in Japan and womenin Nicaragua. The Asian males of America 1 can expect to outlive the urban Blacks of America 8 by 15.4 years -- similar to the gap between Icelandic men and men in Uzbekistan.

Okay. As I said before, nothing new here, just an elaborate way of telling us what we already know -- which is not nearly enough since we have such limited data on ethnicity, social and economic status, health care access and utilization, environmental exposure, and behavioral risk factors for the U.S. population, as the authors say. Still, it does seem more than disgraceful, doesn't it?

Oh look -- terrorists! They might kill somebody!