Is life expectancy for the worst-educating shrinking?

Now here’s a scary headline and lead from the NYT:

Life Spans Shrink for Least-Educated Whites in the U.S.

For generations of Americans, it was a given that children would live longer than their parents. But there is now mounting evidence that this enduring trend has reversed itself for the country’s least-educated whites, an increasingly troubled group whose life expectancy has fallen by four years since 1990.

The rest of the story is full of lots more scary language, including references to the “years of life” supposedly being “lost” by folks at the bottom, some not-very-convincing speculation about causes, and a comparison to the shrinkage in Russian longevity after the Soviet system came apart.

The good news is that the bad news isn’t true, or at least that the study in question – done by a heavyweight team and published in Health Affairs – doesn’t show that it is true.

The shoe drops in the sixteenth paragraph:

Researchers said they were baffled by the magnitude of the drop. Some cautioned that the results could be overstated because Americans without a high school diploma — about 12 percent of the population, down from about 22 percent in 1990, according to the Census Bureau — were a shrinking group that was now more likely to be disadvantaged in ways besides education, compared with past generations.

Professor Olshansky agreed that the group was now smaller, but said the magnitude of the drop in life expectancy was still a measure of deterioration. “The good news is that there are fewer people in this group,” he said. “The bad news is that those who are in it are dying more quickly.”

In other words, women in roughly the bottom decile of the educational-attainment distribution in 2008 had a lower life expectancy than women in roughly the bottom quintile of that distribution in 1990. No sh*t, Sherlock. It’s clear from the paper that the authors had the data to do an apples-to-apples comparison, matching the bottom 12% in 2008 with the bottom 12% in 1990, but chose not to.

Now, the finding that whites in the bottom decile have lives about four years shorter than those in the bottom quintile — that is, the group consisting of the very-worst-off tenth and those just above them — is shocking enough. It’s at least suggestive evidence of the human costs of extreme inequality. It might even lead you to say that the policy of making the poor poorer acts as a kind of silent “death panel.” But what it doesn’t suggest is that things are somehow getting worse. (The fact that the U.S. is slipping in the world life expectancy league tables is much more bothersome, but the NYT reports that fact without analyzing it.)

Moreover, it would be hasty to conclude – as the journal paper and the news account both implicitly assume – that the difference came entirely from the social disadvantage created by not graduating from high school, as opposed to the social disadvantage and personal characteristics reflected in not graduating from high school.

Schoolwork is mostly boring, especially if you’re not very good at it, and dropping out means more time, right now, for either paid work or leisure, at some expense in future opportunities. You’d expect, other things equal, that people who “give no thought to the morrow” would be more prone to drop out, and also more prone to do other things that provide current benefit at future cost: smoking cigarettes, for example. You’d also suspect that they had, on average, less attentive parents.

So what we have here is not evidence of a health crisis, or even of growing inequality, but merely an excellent bad example for those who need to teach about sample selection bias.

Fortunately, there’s lots of actual bad news to fret about, so you needn’t miss your MDR of worry.

Comments

  1. politicalfootball says

    Is life expectancy for the worst-educating shrinking?

    I don’t think you have to worry about this personally, professor. I’m sure you’re a fine educator.

  2. EB says

    Thanks for parsing this out. Saw the story and realized right away that the study means nothing unless the proportion of the population designated “least-educated” remained the same (and that it was vesry unlikely to have remained the same).

  3. says

    I think the created/reflected distinction is one that’s not really appropriate, and potentially dangerous. It’s way too easy to go from “reflects existing deficits” to implicitly or explicitly marking those deficits as intrinsic to the person.

  4. Barbara says

    I read your post three times and I still can’t figure out what “affirmative argument” you are making, other than a technical exercise in criticizing the work of others (which you do so cryptically as to be ineffective in your own right). Hence, I classify you as being either unconvincing or a poor communicator. Close to 50% of the population cohort studied lack health insurance. Yes, of course, their “lack” of health insurance is a reflection of their “lack” of education, or maybe it’s a “lack” of something else, but so what? The study shows that, for whatever reason, declining economic fate is correlated with reduced life expectancy. So your conclusion is — what — why worry because it’s not “worse than it used to be.” Or maybe it is but it wasn’t actually studied. Are you saying this study is not a further reason to support equalizing access to health care because these people are “destined” to die young anyway? Have you ever actually tried to help a person in this cohort get access to medical care? The burdens to access are overwhelming.

    • politicalfootball says

      I try to maintain my cynicism about the “Reality-Based,” who I reckon are often suckers, but Barbara, I feel compelled to defend Prof. Kleiman here. He is merely saying that the study doesn’t offer convincing evidence in favor of the proposition it purports to prove. That is, it doesn’t support the notion that “Life Spans Shrink for Least-Educated Whites in the U.S.”

      As I say, I try not to get too excited about getting the facts straight, but if you’re not going to sweat the facts, you need to pay attention to who your friends are. Prof. Kleiman is on the side of the angels on this one.

      • Barbara says

        Then why can’t he be more forthright in saying that? You write for a general audience, you need to understand how what you write comes across. Basically, this came across as mostly undercutting the study (no problem with that) but the rest of his propositions were posed as “maybe” points — the whole came across as frankly a bunch of mush. It still does IMO, and if I didn’t already KNOW where Mr. Kleiman stands I would not have “assumed” anything about whether he sides with the angels.

        • J. Michael Neal says

          Whether Mark Kleiman sides with the angels is entirely incidental to this post. Even if it meant that he didn’t take that side, it doesn’t change the fact that the study mentioned is complete crap as it is written. Worse, they had all of the data necessary to make it valuable, but they chose not to use it. As for not being forthright in his position, the following is exactly that:

          “So what we have here is not evidence of a health crisis, or even of growing inequality, but merely an excellent bad example for those who need to teach about sample selection bias.”

          It’s hard to be more succinct and clear than that.

          Your objection seems to be that he didn’t make any policy points that you agree with. If that’s all you’re looking for, you’re at the wrong blog. Sometimes the methodological questions are both interesting and important, and the proprietor posts about them fairly frequently. My advice is to either skip the posts where he (or someone else) does that. The rest of us often find them very interesting.

          • Barbara says

            “Schoolwork is mostly boring, especially if you’re not very good at it, and dropping out means more time, right now, for either paid work or leisure, at some expense in future opportunities. You’d expect, other things equal, that people who “give no thought to the morrow” would be more prone to drop out, and also more prone to do other things that provide current benefit at future cost: smoking cigarettes, for example. You’d also suspect that they had, on average, less attentive parents.”

            If that isn’t a judgment about the worth of the population being studied I don’t know what is. He said that — I am not criticizing him for failing to say something else.

          • NCG says

            I couldn’t find the reply button for Barbara, but that’s what this is meant to be.

            I do think there are occasional bits of intellectual snobbery here in the RBC. Here and there, a suggestion that the less intelligent are lesser.

            But, I have never seen anyone wanting them to die young. And bad parenting can happen to any baby.

            And I think Mark just really doesn’t like it when people do sloppy work. They should just re-crunch their data, or whatever, and all will be right again.

          • J. Michael Neal says

            If that isn’t a judgment about the worth of the population being studied I don’t know what is. He said that — I am not criticizing him for failing to say something else.

            Except that it’s a lot more than a criticism of the population being studied. It’s a statement of a fact that has been empirically demonstrated many, many times: the poor and poorly educated have a much shorter time horizon for their preferences than do the well-off or the better educated. You may not like it, but there isn’t any disputing it. It’s an inconvenient, even demoralizing, fact.

            If it becomes impossible to state a fact that is as obvious to anyone who studies such things as this one is because one is afraid of being seen as criticizing a class of people, then serious work becomes impossible. Knowing this fact, about the only value judgment to be read into what Mark said is that this fact is a bad thing. Which it is. We need to work to overcome it. Pretending that it doesn’t exist is wrongheaded.

  5. MikeM says

    Moreover, this study apparently compares people born in the 20s with those born in the 40s (or even earlier). It seems to be a cross-sectional study of some sort, and cross-sectional studies can be very misleading. I haven’t read the paper, only the abstract, but I would expect that a longitudinal study, starting with cohorts from the 1920s and 1940s, would be more informative. This is over and above Marks observation about the shrinkage. And if I’m right about the subjects, many are still alive, so studying death records doesn’t tell the whole story.

  6. Ken Doran says

    Life expectancy is important, very likely the best single indicator of social welfare. Some of Mark’s concerns did occur to me. I think what we need here is an independent review from a someone who knows the subject inside out and can write for a lay audience. If anyone can point to a good example, or recruit one from inside the academy, I for one would appreciate it.

  7. Barbara says

    J Michael Neal, it is not a statement of fact. It’s stating a hypothesis about the population that is drawn almost entirely from negative stereotypes. I almost always agree with Mark, but having relatives in real time in the category makes me personally affronted when I read paragraphs like the one I excerpted. Seeing the obstacles they face in trying to improve their lot — obstacles that are very often enhanced by judgments such as those displayed in that paragraph — makes me much less willing to jump to the conclusion that they are the author of their own fates. Of course, we all are to some extent, but it’s also the case that what we reward is not entirely tied to merit, and frequently enough has a lot to do with self-reinforcing trends that make a bad situation even worse than it needs to be.

    That is, for instance, doctors (highly educated and highly compensated) have office hours that almost never extend to evenings and weekends and they don’t even take phone calls during lunch. Hourly wage workers find it extremely difficult to justify taking time off to visit a doctor even when they are sick, much less for preventive services, and sometimes can’t even get through to make an appointment because they aren’t “allowed” to conduct personal business on company time (and so, when they are free at lunch, find the doctor’s staff unavailable). So somebody finally makes it to an ER for something that should have been treated in an office and meets medical staff who are positively contemptuous that the person doesn’t find it “important” to maintain their health the way the doctor and the doctor’s friends would.

    That’s the world these people live in and until it is ameliorated in their favor I think it appropriate to avoid jumping to “factual” conclusions about their self-defeating habits.

    • J. Michael Neal says

      Yes, actually, it is a fact. It isn’t the only fact, but it is one. Like all statements about categories of people, it applies in varying degrees, including often not at all, to particular members of the category. Nevertheless, it is true on average.

      I think you are also reading value judgments into Mark’s statement that are not necessarily there. I don’t doubt that you hear them regularly, since I do, too. But without any statement as to how the things he relates come about, there isn’t really any value content.

      If we’re going to argue anecdotes, I see a lot of those problems on a regular basis in volunteer work that I do. It gets very frustrating trying to tutor adult students who have zero interest in math and very little ability to work through even very simple problems. I know that most of that comes about because they went to schools that use teaching methods seemingly designed to convince students that math is tedious and that they aren’t any good at it. (I know this because I also tutor elementary school students and see their homework.) However, knowing why they ended up with a complete indifference to the level of math necessary to do basic bookkeeping doesn’t change the fact that a lot of them have that indifference.

      • Barbara says

        “Like all statements about categories of people, it applies in varying degrees, including often not at all, to particular members of the category.”

        Your idea of a fact is different from mine.

        • J. Michael Neal says

          If you choose to deny that it is impossible to make factual statements about group tendencies, then our ideas of what constitute facts are indeed very different.

  8. retr2327 says

    One apparent finding of the study that hasn’t received any particular discussion (perhaps because I’m just misreading the charts?) is a striking disparity between the trends for Hispanics, both male and female. It seemed, to my eyes at least, that those two groups showed a dramatic improvement in longetivity, while other groups declined. The improvement is a good thing, obviously, but what’s causing that? Any thoughts?

    One possible hypothesis I’d throw out there might be the recent decline in immigration rates: you’d expect immigrants who have been here longer to start climbing up the ladder and doing better on a wide range of measures, so a drop in immigration rates could result in a sampled population with a longer average time in the U.S. But I have no idea if there’s any merit to that.

  9. says

    Ofshanky et al‘s error is all the more surprising in that the previous big study of the same issue in 2008 by Meara, Richards and Cutler did run an appropriate check:

    Because rising educational attainment over time could affect our results, we recomputed estimates after equalizing the shares in the high- and low-education groups. We did this by randomly reassigning some people with twelve years of education to the high-education group in the early period to match education shares in the latter period. The patterns we documented are robust to the changing composition within education groups.

    Mike M: the Meara study included longitudinal data.