After Thomas Frank wrote What’s the Matter with Kansas?, he was subject to withering criticism from Princeton political scientist Larry Bartels, who argued that Frank’s portrait of working class voters who vote Republican on culture war grounds was highly misleading. Frank had an effective rejoinder, but it settled nothing. Henry Blodget summarized the exchange at Crooked Timber.
I blogged Larry Bartels’ take-no-prisoners critique of Thomas Frank last week; now Frank has come back with an equally frank rejoinder. It seems to me that Frank’s riposte to Bartels on the issue of how to define the white working class strikes home, although some of his other jabs miss the target. Since the definitional question is key to Bartels’ critique, it looks to me as if Franks comes out ahead (unless Bartels comes back with a more convincing justification). Still, I find some of Frank’s apologia unconvincing. He’s absolutely right to say that working class conservatism is still important, even if it isn’t a majority phenomenon – but there is a tendency (which Frank is by no means immune to) to generalize from the particular and to draw big conclusions on the basis of limited and somewhat impressionistic information. National survey data is imperfect – but it can serve as a very useful corrective to these tendencies, and it’s unfortunate that more pundits don’t use it (or at least acknowledge more directly the limits of the kinds of information that they do refer to).
Drawing big conclusions on the basis of limited and somewhat impressionistic information is, alas, what literary intellectuals do, and it is important. It’s not always right. It is not always the basis for sound policymaking. But it offers a useful lens for approaching and understanding a dense, complex social world that can’t be captured by the extremely, extremely crude tools of social science. Susan Sontag’s “On Camp” offers a story that is hard to tell through rigorous data analysis. Joan Didion’s reflections on postwar California weren’t the work of a professional sociologist or political scientist. But they resonated with a large audience. While I’m strongly inclined to disagree with Frank’s characterization of working-class conservatives — Ross and I have outlined our objections to the Frank thesis — arguments from statistical analysis often miss the point.
Which is why I was struck by Jessica Grose’s thoughtful, intelligent, but ultimately puzzling take on Lori Gottlieb’s Marry Him in Slate. Gottlieb’s book is calculated to generate maximum controversy: it is aimed at a small, affluent, highly influential group of women, and it is essentially an effort to scold many of them. It is hardly surprising that it has become a subject of endless, roiling debate among a certain kind of educated middle or upper middle class person.
Grose is deploying fairly crude statistics to make the case against a highly impressionistic account. We could criticize Gottlieb for not producing a work of social science. But the truth is that we don’t have very good statistical tools for capturing the very small and narrow tribe of women that Gottlieb is talking about: for example, not every woman who graduates Duke or UC San Diego occupies the same psychographic niche, though that is masked by statistics.
So what about those screenwriters and female execs? For them, Gottlieb is making a provocative argument: They will lose their last chance at happiness if they don’t lower the bar. Built into her argument is the assumption that all women over 40 would be happier as part of a couple, even with a nice bore of a husband. In the Atlantic article upon which her book is based, Gottlieb writes, “Madame Bovary might not see it that way, but if she’d remained single I bet she would have been even more depressed than she was while living with her tedious but caring husband.”
Here again, the statistics don’t necessarily bear her out. Married women—and here we are talking about upper-class married women—are getting less, not more happy in their marriages. The percentage of college-educated white women who describe their marriage as “very happy” dropped from 74 percent in the 1970s to 68 percent in the 2000s, according to Brad Wilcox, director of the National Marriage Project. Now, this doesn’t mean that they would have been happier unmarried. But it does mean that marriage has become a whole lot more complicated for the upper-class woman than it was in the ’70s, fraught as it is with time pressures and conflicted roles. Gottlieb talks about women who divorce and then remarry—and found that “most of them reported no increase in satisfaction or that they were less happy than they were in their first marriages.” But she doesn’t talk about the women who remain single in their 40s and beyond. [Emphasis added.]
Of course the statistics “don’t necessarily bear her out.” I can’t imagine a scenario in which the statistics would necessarily bear her out, as this is confusing, ambiguous, highly uncertain terrain in which we’re dealing with lots and lots of very contradictory evidence. And if we’re talking about a small but influential group of screenwriters and female execs, it is ethnography, and not statistical analysis, that is our best bet for deriving some rudimentary understanding of what the heck is going on. Ethnography is notoriously difficult, and it invites the possibility of selection and confirmation bias. In very small groups, however, it is pretty much the only game in town.
Is it at least possible that married women are more happy than they would be if there were not in said marriages? A time-slice comparison doesn’t make much sense in this context. It could be that married women circa 1975 were happier than married women circa 2005. It’s not clear that this has any bearing on Gottlieb’s argument.
In fact, levels of well-being have gone down since the ’70s for both married and never-married women, according to the happiness study by Wharton economists Betsey Stevenson and Justin Wolfers that got so much press last year. And as Elizabeth Gilbert points out in Committed, married women are more likely to suffer from depression than singletons. Marcus Buckingham, writing about women’s happiness in the Huffington Post, notes, “Women’s happiness with their marriage sinks below men’s at age 39.” Rather than settling in unhappily with her snooze of a spouse, this research indicates that modern Mme. Bovary might be better off setting up a bachelorette pad in Paris.
The Stevenson-Wolfers study is interesting, but it doesn’t have any bearing on Gottlieb. Rather, it reflects many of the broader cultural trends that Gottlieb describes. As for Gilbert’s observation, one has to assume that there is selection bias involved. I haven’t read Gilbert’s book, but I imagine that married women are more likely to have comprehensive health insurance coverage that would allow them to get diagnosed for depression. I’m certainly not suggesting that single women are in reality more depressed. Rather, I think we should throw cold water on these statistics.
As for the Gottlieb thesis itself, I’ll say only this: the article sparked many conversations, the book will likely sell many copies. I see Gottlieb’s argument as part of a continuing conversation about how social roles are evolving and how they should evolve. For whatever reason, her portrait rings true with many people and it frustrates many others. As someone who likes to see people reading and arguing, this doesn’t strike me as a bad thing.