If traditional institutions reduce suicide, as Douthat suggests, then suicide rates should be lower in the South, where religiosity is highest, or in the inland West, where marriage is most common. Instead, suicide rates are lowest in the northeastern corridor, with Washington, D.C., and New York rounding out the bottom of the list. Suicide is most common in the sparsely populated states of the interior West. If the 30 percent increase is worthy of a headline, then western states deserve a spot in the headlines, too: The suicide rate in states like Alaska or Montana is more than 300 percent higher than in D.C. or New York.
So let’s look at Douthat’s column:
In the 1990s, the suicide rate dipped with the crime rate. But since 2000, it has risen, and jumped particularly sharply among the middle-aged. The suicide rate for Americans 35 to 54 increased nearly 30 percent between 1999 and 2010; for men in their 50s, it rose nearly 50 percent. More Americans now die of suicide than in car accidents, and gun suicides are almost twice as common as gun homicides.
This trend is striking without necessarily being surprising. As the University of Virginia sociologist Brad Wilcox pointed out recently, there’s a strong link between suicide and weakened social ties: people — and especially men — become more likely to kill themselves “when they get disconnected from society’s core institutions (e.g., marriage, religion) or when their economic prospects take a dive (e.g., unemployment).” That’s exactly what we’ve seen happen lately among the middle-aged male population, whose suicide rates have climbed the fastest: a retreat from family obligations, from civic and religious participation, and from full-time paying work.
The hard question facing 21st-century America is whether this retreat from community can reverse itself, or whether an aging society dealing with structural unemployment and declining birth and marriage rates is simply destined to leave more people disconnected, anxious and alone.
Nate Cohn argues that this is all wrong because “If anything correlates with suicide rates, it’s a states’ [sic] population density: In populous areas, suicide rates are low; in the sparsely populated hinterlands, suicide rates are high. Perhaps depression and loneliness is particularly harsh in desolate areas, and maybe it’s easier to cope in a major city like D.C. or New York.” Which is — well, actually, it’s not that different than what Douthat wrote, is it? — given that it’s far easier to become “disconnected from society’s core institutions” when you’re living in “sparsely populated hinterlands.”
Cohn seems so desperate to disagree with Douthat that he fails to notice now similar their explanations are. Perhaps Cohn could benefit from investigating the studies that Brad Wilcox cites here, in a post that outlines the data that Douthat is drawing on. Cohn doesn’t seem aware of those studies. He repeatedly insists that “the data” doesn’t support Douthat’s claims, but doesn’t actually cite any data that supports his view: he refers only to two very general datasets and one New York Times story that summarizes some recent research in this way:
Preliminary research at Rutgers suggests that the risk for suicide is unlikely to abate for future generations. Changes in marriage, social isolation and family roles mean many of the pressures faced by baby boomers will continue in the next generation, Dr. Phillips said.
Or, precisely the explanation that Douthat suggests. Nate Cohn seems to have gotten into a panic at the thought that “society’s core institutions” could have anything to commend them and let that panic chase him far ahead of any actual information about American suicides.
Or maybe he has data that contradicts the explanations offered by the very researchers he links to. I can’t tell from his post. And his only references to (again, uncited) data are at the state level, which is far too coarse a grind for this kind of question, given how different urban and rural cultures can be in the same state. If you’re going to insist that you have data that proves that someone else’s explanations for a social phenomenon are definitely wrong, you should link to that data, shouldn’t you?