But statistical thinking also leads to bad intellectual habits. One is what I call "dismissing the minority." Consider the uproar spawned among chattering feminists over Sheryl Sandberg’s Lean In, another manifesto urging women to climb the corporate ladder. Statistical duels have broken out all over about how many women are housewives or full-time mothers, how many are working full time and so on. Jessica Grose at Slate is annoyed by a New York Magazine story on highly educated women who have chosen to be stay-at-home moms, and she writes,
Women like Kelly Makino—the social worker featured most heavily in the piece—are still pretty rare. Of women with graduate or professional degrees, 75 percent of them who had a child in the past year work, and 60 percent of those women work full time. When you look at highly educated women who have older children, about 86 percent of them are in the work force. So we’re not talking about hordes of women who are “too busy mining their grandmothers’ old-fashioned lives for values they can appropriate like heirlooms, then wear proudly as their own,” as New York claims. Such women, if they exist, are a minuscule sliver of the whole pie.But is 14 percent a "miniscule sliver"? We're talking about tens of thousands of women. About 14 percent of Americans are black, but nobody thinks it's odd for the media to cover black people. Yes, it can be annoying when lazy reporters invent trends based on what two of their old college friends are doing. But the fact is that American mothers are, when it comes to paid employment, highly diverse: some are corporate climbers like Sheryl Sandberg, some work full time but consider motherhood at least as important, some work part time, some stay home. Follow them over the course of their lives and you see all sorts of different strategies, moving in and out of the work force and making different kinds of compromises of time and money and child care. There is no single, dominant narrative, other than the need to make choices.
I most often object to arguments that dismiss the minority when it comes to human evolution. I regularly encounter arguments that go something like, "70 percent of documented hunter-gatherer societies practice X, therefore this is the human evolutionary norm." Garbage. There are very few "norms" across human societies, but there is a vast amount of diversity. If you find that, say, 60 percent of hunter-gatherers revere the spirits of their ancestors, that doesn't make ancestor worship some sort of human norm that can then be used to explain the evolution of religion. What it shows is that some people revere their ancestors and some don't, and your evolutionary model had better be able to account for both.
It is interesting to know what the majority of any sample is, and what it is doing, but it is not the whole story. If your theory can't explain the minority as well, your theory is bunk. And if you want to understand, say, American mothers' relationship to the corporate world, you need to think about all mothers, not just whatever portion of the total is doing what you are.