K-12 Education and Life Expectancy

Kevin Drum in Mother Jones points out that Not Everyone is Living Longer

Longer life spans, it turns out, really do depend on just how privileged you are.

He's referring to a MacArthur Foundation Research Network report that Differences In Life Expectancy Due To Race And Educational Differences Are Widening, And Many May Not Catch Up :

Drop in life expectancy for least educated whites, slight increase for least educated blacks. It is important to note that the size of the least educated subgroup of the U.S. population has been shrinking in recent decades (down to about 8% for whites).

Their graph summarizes:

So the data is compatible with what Drum claims: it appears that life expectancy has dropped for the less educated. In the course of 18 years, a randomly selected white female dropout has lost 5 years of life expectancy while a randomly selected white male dropout has lost 3 years...but that's not the only interpretation. The white female dropout population of 1990 (WFD90) has different people in it than the white female dropout population of 2008 (WFD08).
Imagine that we could rank people in some combined ordering that combined intelligence with emotional factors including "grit"; think about perseverance, focus on a plan, supportive family, all kinds of factors that make it easier to finish an education and that also make it easier to do well in life, not only getting and keeping a job but seeking and following medical advice from stop-smoking and stop-eating-so-much to recognizing the symptoms of a heart attack and taking the pills you've been prescribed. In my understanding of the world, these factors very often (not always) go together.
Now, think of WFD08 as being an exact copy of WFD90 except that we've pushed education harder, so that the dropout population is a smaller part of the population (hey, it is). Some of the top dropouts in WFD90 have gone into the high-school graduate population instead, but perhaps some or all of their other life-factors are no different. In that case, part of what we're doing is simply lowering the margin below those who were the top of the bottom group: it could be that every individual of 2008 is better off than an identical individual of 1990 would have been, but we'll see that the average lifespan of a dropout is less. Finishing high school may have helped the graduates live longer, or it may have made no difference at all -- the data is compatible with either.
A lot of research has this kind of ambiguity; a set of international-comparison studies that were much discussed this week led economist Arnold Kling to write Good News: School Doesn’t Matter!

In my view, the policy implication is that we should spend a lot less on classroom education and instead spend more on better research, including randomized controlled trials, to find out what, if anything, makes a difference. For now, I see no evidence that the money we spend on education is anything other than an enormous waste.

I wouldn't go quite that far, but I do think it would be a better world if people were more careful with their inferences.
Or then again, maybe not.