SCIENCE AUGUST 11, 2013
In June, Psychological Science, an academic monthly that calls itself the "the highest ranked empirical journal in psychology," published a peer-reviewed article titled "The Fluctuating Female Vote: Politics, Religion and the Ovulatory Cycle." Lead-authored by Kristina Durante, an assistant professor of marketing at the University of Texas at San Antonio who has previously written on ovulation and behavior, the paper upset and woefully-entertained many readers when an early version first became viewable online last year. The final article concludes:
In two studies with relatively large and diverse samples of women, we found that...ovulation led single women to become more socially liberal, less religious, and more likely to vote for Barack Obama. Conversely, ovulation led women in relationships to become more socially conservative, more religious, and more likely to vote for Mitt Romney...These findings suggest that the ovulatory cycle might play an important role in women's politics.
There is no shortage of psychological studies about female fertility these days. In July, the same journal published a paper titled, "Women Are More Likely to Wear Red or Pink at Peak Fertility." And last year, another serious journal article attempted to answer the eternal question, "Do Women Prefer More Complex Music Around Ovulation." (The short answer: No.)
Durante and her coauthors, according to critics, added another such footnote to the long and inglorious history of sexist pseudo-science, going all the way back to the ancient Egyptians' diagnosis of hysteria around 1900 B.C. (The cause of this female-only disorder way back then? Spontaneous uterus movement within the female body, obviously.) In a credulous CNN article about "The Fluctuating Female Vote," which was removed by the network (but remains available here) after an online furor over its existence, a Rutgers professor of political science and gender studies says, way down near the end of the piece, that the paper was sexist: "There is absolutely no reason to expect that women's hormones affect how they vote ... [and there is] a long and troubling history of using women's hormones as an excuse to exclude them from politics and other societal opportunities."
Many academics, for a variety of reasons, are concerned about lowering journal standards. To Andrew Gelman, director of the applied statistics center at Columbia University, Durante's paper is representative of a "growing crisis" in social science journals: published and subsequently publicized research findings that are "fragile, unreliable, cannot be replicated, and do not generalize outside the lab to real-word settings." Peer review, that age-old method of academic self-policing, now appears to be wholly insufficient.
Indeed, perhaps even more troubling than the “Fluctuating Female Vote”’s casually sexist conclusion is the shoddy math and science that produced it. A lay reader might not suspect it: The paper, after all, says it relied on two “relatively large and diverse samples of women,” and it included plenty of graphs and P-values supporting its results.
But, to a trained eye like Gelman’s, it was “sloppy work,” as he pointed out in a blog post after its publication. For one thing, the authors found subjects for their two studies through a get-paid-to-take-an-online-survey system on Amazon.com, which may not have resulted in a diverse sample group or accurate determinations of fertility. The subjects were also only asked questions (such as: Who would you vote for in the 2012 presidential election?) at a single moment in time, not accounting for changing attitudes. Additionally, certain women (those who were at the beginning or end of their ovulatory cycles, and thus deemed unreliable due to menstrual or premenstrual symptoms) were excluded from the study. As, notes Gelman, was some potentially revelatory data analysis (comparing respondents in different parts of their cycle using variables such as birth year, party identification, marital status: "Just a whole damn table," he begs, "showing these differences for every possible variable") that could have presented alternative stories to the one humdinger they ultimately proposed. Gelman's reprimanding blog post concludes:
It’s not innovative data collection, it’s not great theory, it’s not great data analysis, it’s not a definitive data source, it’s nothing. What it is, is headline-bait that’s not obviously wrong.
It's not clear why Psychological Science, which has a citation ranking that regularly places it in or near the top ten psychology journals worldwide, would have fallen for such a "sloppy" paper. According to its editor, Eric Eich, the journal now receives well over 3,000 submissions a year from top researchers in clinical psychology, evolutionary psychology, industrial psychology and the like, all of which are read in their entirety by a specialist in the given subject, as well as a generalist. About two thirds are rejected after an initial review by the editorial team, while the other third receives "extended review" by two or three outside experts. Most of those, too, are "triaged" for not being "sufficiently groundbreaking." In the end, only about eleven percent of submitted papers are chosen, making publication in PS slightly harder than gaining admission to the University of Pennsylvania as an undergrad. Eich's first guiding principle throughout this process is to "do not harm." But ultimately, he seeks to fill his pages with "the most exciting, innovative research that is apt to interest a broad cross-section of readers." Kristina Durante's paper on ovulation and voting made the cut.
Durante herself has been mostly mum since "The Fluctuating Female Vote" came out. But she recently opened up to me a bit by email about her thinking. On why she has written so frequently about the link between behavior and ovulation:
We usually don’t see our hormones, we can’t feel them, and we’re not aware that they are doing much of anything. Yet thousands of studies have found that hormones like testosterone and cortisol have profound effects on many important behaviors. Scientists have historically paid little attention to ovulatory hormones because, like with other hormones, people used to think that ovulatory hormones have no effect on behavior. It wasn’t until scientists started to test questions more rigorously that they found that ovulatory hormones have profound influences on women’s everyday behavior. It has always been fascinating to me that something so invisible as biology has such a very visible effect on our daily choices.
A better understanding of ovulatory hormones has the potential to advance not just the study of women’s behavior, but also of women’s mental and physical health. This work also has another important practical implication. Understanding the causes of behavior can help change behavior. For example, if we know that the ovulatory cycle subconsciously biases women’s behavior in a particular way, this is important to know so that women can de-bias their behavior. After all, knowing the causes of our behavior empowers us to have more control over our choices.
How exactly one can precisely "de-bias" one's hormonal influences on behavior is, of course, fertile ground for another set of controversial studies and articles.
In any case, it is true, as Durante told me, that men must also endure studies connecting the silent, internal activities of their bodies with their outward politics: Another recent Psychological Science paper, "The Ancestral Logic of Politics: Upper-Body Strength Regulates Men's Assertion of Self-Interest Over Economic Redistribution," concluded that "men with greater upper body strength more strongly endorsed the self-beneficial position." So far, no public outcry against these findings has been reported.
Charles Bethea is a journalist based in Atlanta. He has written for The New Yorker, The New York Times Magazine, and GQ, among others.