It’s not often that wade into retractions in the mainstream media on this blog, but in this case, we’ll make an exception.
As Politico and Poynter — and probably others — have reported, CNN has retracted a story about a yet-to-be-published study in Psychological Science claiming to find a link between estrogen and elections (disclosure: Ivan’s wife works at CNN). Specifically, the researchers reported that the well-documented preference among single women for President Obama might be rooted in their sex hormones, while that of married women for Mitt Romney seems to reflect their own ovulatory cycle. Or something like that.
Here’s the money part of the piece, which can still be found floating around on the web:
The researchers found that during the fertile time of the month, when levels of the hormone estrogen are high, single women appeared more likely to vote for Obama and committed women appeared more likely to vote for Romney, by a margin of at least 20%, Durante said. This seems to be the driver behind the researchers’ overall observation that single women were inclined toward Obama and committed women leaned toward Romney.
Here’s how Durante explains this: When women are ovulating, they “feel sexier,” and therefore lean more toward liberal attitudes on abortion and marriage equality. Married women have the same hormones firing, but tend to take the opposite viewpoint on these issues, she says.
The study used an online survey, and we can’t tell whether that 20% gap is an absolute spread or a relative difference.
Durante is Kristina Durante, an assistant professor of marketing at the University of Texas, in San Antonio. According to her website:
My research draws on theory in evolutionary psychology to examine how both situational and biological factors nonconsciously influence social and consumer behavior. The overarching focus of my research program is the consumer behavior of women and families. Within this program I examine:
- The hormonal mechanisms that guide women’s decision-making
- The influence of status-seeking and competition on women’s consumer choice
- The role of social influence and word-of-mouth on women’s consumption
- The environmental factors that influence family spending patterns.
Her listed publications include:
“Ovulation Leads Women to Perceive Sexy Cads as Good Dads”, with Vladas Griskevicius, Jeffry A. Simpson, Stephanie M. Cantu and Norman P. Li, Journal of Personality and Social Psychology — covered here by Psychology Today.
“Do Women Feel Worse to Look Their Best? Testing the Relationship between Self-Esteem and Fertility Status across the Menstrual Cycle,” with Sarah E. Hill, Personality and Social Psychology Bulletin, Vol. 35, 2009, pp. 1592-1601
“Changes in Women’s Choice of Dress across the Ovulatory Cycle: Naturalistic and Laboratory Task-Based Evidence,” with Norman P. Li and Martie G. Haselton, Personality and Social Psychology Bulletin, Vol. 34, 2008, pp. 1451-1460.
Why estrogen would make single women feel sexy but turns married women off goes unexplained in the article — probably because the notion is, on its face, complete poppycock. After all, there’s no evidence we’re aware of to suggest that single women have better sex lives, while data show that married women or those in otherwise committed relationships have plenty of sex.
CNN acted within hours of posting the story yesterday afternoon, and now has this notice up on its site:
A post previously published in this space regarding a study about how hormones may influence voting choices has been removed.
After further review it was determined that some elements of the story did not meet the editorial standards of CNN.
We thank you for your comments and feedback.
It’s unclear, of course, what they meant by “some elements.”
The reporter, Elizabeth Landau, tweeted yesterday that she was simply reporting on a paper that had been accepted for publication:
For the record, I was reporting on a study to be published in a peer-reviewed journal & included skepticism. I did not conduct the study.
And she did offer this disclaimer up high in her piece:
Please continue reading with caution. Although the study will be published in the peer-reviewed journal Psychological Science, several political scientists who read the study have expressed skepticism about its conclusions.
But that’s beside the point. Hers would be a more plausible defense if Landau had written, say, a 300-word brief on the paper — how’d she get it, anyway? A heads-up from the researchers or even the journal? A university press release? — rather than a 1,000+plus-word story (we counted) with several interviews.
Even with the quotations from skeptics, the net result feels a bit like passing on a rumor that someone kicked a dog, then asking three of his friends to vouch for his good character. Why monger the rumor in the first place?
The whole thing strikes us as a symptom of “poor science, poorly reported.” The mere fact that a paper has been accepted for publication in a “peer-reviewed journal” does not make it A) worth the paper it’s printed on and B) worth spilling more ink and pixels.
But let’s not pile on Landau for writing about what she supposed — correctly, but for the wrong reasons, it turns out — would be an attention-grabbing story. The journal and its peer reviewers bear as much responsibility for accepting the paper in what, on the face of it, seems an attempt to have some timely election-related research.
Still, Psychological Science, we should note, has recently retracted two papers by Diederick Stapel, and, to their credit, showed a great deal of backbone in pushing for a retraction from Lawrence Sanna.
Science bloggers have weighed in. Read Scicurious’s take on the study here, and Kate Clancy’s critique here.
On-line survey … okay, there’s the first clue that this “research” is total poppycock.
Accepted for publication by a peer-reviewed journal … okay, I can see how the reporter is not totally to blame here. She is not a science expert. Her specialty is getting some comments and writing up the stuff that is made available to her. She doesn’t know that journals run junk along with good science. But I remember refusing to write a few stories that seemed just too ridiculous to be true, and being glad when my employer didn’t have to run apologies for telling readers something that was completely false. There should be some gatekeeping in the serious, professional media. The wide-open, unmoderated channel is personal posts on the internet where anything goes and the buyer knows s/he should beware. And then there is the National Enquirer, where you can bet that if they run it, it’s not true, even if they have photos, which are undoubtedly doctored or misrepresented.
The responsible course for a serious journalism organization to take, if there is concern about not having the eye-catching-but-probably-false story that the competitor is running, is to do some serious reporting on why the eye-catching story is false. Big journalism prizes have been won for doing this. It can be called public service reporting or reporting in the public interest rather than reporting on what the public is interested in. Unfortunately, this requires more investment than many news organizations are willing to provide.
Shame, double shame, and triple shame on the researcher for creating this garbage and on the journal’s reviewers for accepting it for publication. One has to wonder whether the new information age is actually improving our access to information or is just giving us a larger helping of the same old wives’ tales that have always been the staple of the average person’s “knowledge”.
JudyH, agreed about shame on reviewers and editor who accepted what sounds like extremely bad research.
Your suggestion about online data collection indicating low quality research is ill-informed. There have been a lot of side by side comparisons of internet and noninternet data sources and the results generally look good. See eg
http://homepage.psy.utexas.edu/homepage/students/buhrmester/BuhrmesterKwangGosling_PoPS_inpress.PDF
The more standard approach of testing college undergrads has many disadvantages particularly when research topics relate to political behavior.
“Women from the U.S. (42 states) participated for a small payment via an Internet hosting site (MTurk). ”
The website they used (Amazon Mechanical Turk) is designed for things like this. What’s the problem with this method?
Well, it is hard to assess the ovulation or estrogen level based on the internet data, which is clue of this study…
How about Landau’s editors? Surely, they bear some responsibility as well.
Indeed. News organizations that have quotas for column inches or air minutes or words posted or number of clicks that the output must receive are inviting their reporters to merely re-write (or some cases that I know of not even that) whatever is sent to them by individuals and outfits that want publicity for something. The easier the re-write job and the more outrageous the topic, the more likely the publicity release will be presented to the consumer.
Every supervisor in the chain participates in the dilution of quality, either by jawboning about maintaining profits without specifying how or by setting standards that are difficult to meet via legitimate news reporting methods. A reporter who relies on “including skepticism” as evidence that the requirement of “balance” has been met is letting a just-following-the-rules approach substitute for intelligent evaluation of whether the item should be used at all.
So the women who responded to the on-line survey are the real “swing voters”. The unattached women vote for Obama when they are ovulating and for Romney the rest of the month, while the attached women do the opposite? Or maybe all women prefer “none of the above” unless they are ovulating and then they diverge toward the conservative or liberal side? Those hysterical women, always thinking with their reproductive organs instead of their brains. No wonder women are incapable of serious analysis and need men to tell them for whom they should vote. We should repeal the nineteenth amendment.
Is that sensational or what? Romnesia linked to high estrogen levels! Is that true also if you have to share a husband? And how the hell could a man then get it? I bet Dr. Durante is up to something big here. National Science Foundation take note.
There is a huge market, including breakfast television shows and magazines sold at checkouts in supermarkets, for this kind of, dare I say, science. Durante is one of the many suppliers of this product for consumption by the masses. This parody of science, by its very nature, uses a very simplified version of the scientific method. It begins with a hypothesis. The fancier the better as it has to mesmerize the common man with the attention span of a fruit fly. And it ends there as the hypothesis is also the conclusion. And you know that your shot at science really sucks when even political scientists doubt your claims.
I find it troubling that this post and the comments are happily calling the study “poppycock”, and yet none of us have actually read the paper yet. Currently the best we have to go on is a reporter’s take on the methods and results, and a clip from the author’s website about her overall research program. Dismissing a research study based on an anecdotal report – and a retracted one at that – strikes me as bad science. Why not put away the pitchfork until the article appears, then judge it by its merits or lack thereof.
I have to agree with Marc J and Frank. The tone here is degenerating into name-calling and bomb-throwing, rather than reasoned discussion. It would be a shame to see Retraction Watch become the PerezHilton.com of the science world.
I’ve always worried that a venue for discussion of retractions could easily become a haven for people who are suspicious of science or hostile to scientists.
I would not agree that it is only people hostile to scientists. There are plenty of scientists in the field who think Psych Science is a bogus journal, and so the priors for something published there being BS are pretty high! That’s the nature of the journal. Psychological Science only publishes “surprising” and “novel” results. By definition, these results are likely to be false positive outliers: 20 people do the study, 19 get the boring, expected, and true, result; 1 person gets the “surprising” false positive (due to the usual alpha=0.05 used to define statistical significance). This one person publishes paper in Psychological Science. Of course, the results does not replicate because it is bogus in the first place. This is the norm, I’m afraid.
I like Jon’s reply. Maybe Dr. Durante did twenty on-line surveys on various consumer decisions that women make, always comparing the results with ovulation. Tide or Gain detergent? Eggs laid by caged or free-range chickens? The red dress or the black dress for the anniversary dinner? A career as a waitress or as a university professor doing on-line surveys? Nineteen of these surveys showed no significant statistical correlation of the decision with the date of ovulation — not even the choice of eggs! — but the twentieth survey hit paydirt.
Dr. Durante studies “the hormonal mechanisms that guide women’s decision-making”. We already know that’s how women make decisions. The real scientific question is how the mechanisms operate. Apparently this is more complex than we have heretofore realized, since the very same hormones have opposite effects depending on the woman’s relationship status. But a very significant finding is that women in committed relationships favor Romney. Lesbian couples are a two-fer for the Republicans this November. 😉
How can you declare the work “poor science” or “complete poppycock” without having read the paper?
Ah, Marc J has said the same thing. I agree, wholeheartedly.
Perhaps poppycock is the wrong word lets go with something more accurate we are evolutionary directed to vote for the sexier person as they obviously have better genes and thus when we vote we are inherently desiring to sleep with the person we vote for…
Should we let people vote to match their menstrual cycles either when estrogen is at a maximum or minimum depending on what they want according to the alleged results?
It this kind of research that anyone with a background in statistics can produce high quality graphs to show a true causation.
I love your sense of humor, Sengupta. Thank you.
Yes, Hilary Clinton was the Democratic candidate for president in 2010. Most of the men voted for her because they would rather sleep with her than with Obama. … Wait, men vote for rational reasons, like the candidate’s policies on economics, civil rights, and foreign affairs. It’s women whose higher brain functions are overridden by their hormones. Men are not impelled toward foolish and unachievable goals by their hormones. Nosirree!
I took a look around, and the whole paper is available on Durante’s faculty page here:
http://business.utsa.edu/faculty/kdurante/files/Durante_PresidentialElection_Hormones.pdf
To note, she didn’t track any shifts in individual women, just compared two groups who were at different points in their cycle. There was also no control group of post menopausal women or women on birth control to establish a baseline, which seems like a pretty big oversight.
Thank you a million times.
Follow the links (at the bottom of this post) to Scicurious and Kate Clancy, science bloggers who have seen the paper and explain in much better detail what the researchers did and why it’s problematic.
Links, please.
I agree with Jon’s remark about Psychological Science. The journal seems to focus on publishing articles that are “newsworthy” (i.e., they will attract attention in the mainstream media) even though they may represent mediocre science that does little to advance our understanding of psychology. As someone who has reviewed manuscripts for the journal, I find that its review process is less rigorous than that of other journals. I rarely read the journal anymore, but once in a while they publish articles on a topic about which I have some expertise. Unfortunately, those articles tend to be lousy and represent examples of bad science, reinforcing my negative opinion of the journal.
To be clear, my comment is not about the specific article mentioned above, but about Psychological Science in general, based on my experience with the journal.
In his blog, Dr. Kurzban clearly notes why this retraction is not warranted. He also suggests that Retraction Watch do a retraction of it’s own piece.
http://www.epjournal.net/blog/2012/10/halloween-horrors-sandy-wreaks-havoc-on-east-coast-hormones-influence-political-views/
I usually love Retraction Watch, but this is irresponsible. Making rash assumptions based on limited data (“shame on reviewers and editors”) is the exemplar of bad science. The condescension toward psychology is something I’d expect from a different blog. This blogger (not me) does a nice job summing up the situation — the retraction at CNN is worlds away from the Sanna/Stapel-type retractions, and to lump this person in with that is absurd.
The article in question might be crap — it’s not my exact field and I haven’t read it — and I don’t make any claim about its quality. I am not objecting to any of the specific criticisms made about the paper. But I do treating a retraction from CNN as the same level as a journal retracting a paper. And the commentators dismissing anything using an online survey as poppycock…..I guess I expected more.
http://www.epjournal.net/blog/2012/10/halloween-horrors-sandy-wreaks-havoc-on-east-coast-hormones-influence-political-views/
The condescension is not about experimental psychology, “Longtime Journal Editor”, it’s about the well-known practices of a specific journal, by people who review for that journal and have been following it for a long time.
Well, have a big dose of schadenfreude over this. So so sick of psychological theorising that goes straight to reductionist explanations of complex human behaviour and cries ‘evolution!’ at every turn.
It is psychology once again demonstrating its inferiority complex, trying to get itself associated with ‘real science’ to overcome its ‘soft science’ stigma.
And maybe if psychologists were required to broaden rather than narrow their professional preparation they could get away from the sort of naive nonsense that infests psychological research, conducted apparently totally unaware of what is happening in other disciplines.
Bah, humbug.