Sex sells. We know that. I’ve seen the effect in the last few days at F1000: we published a press release about an evaluated paper. It’s about premature ejaculation, and is receiving two to three times more views than our other releases (stop it, Brooks).
So it’s perhaps not surprising that Nature gets all excitable about a study in Psychol Sci that’s talking about speed dating and role reversals. Ooh, hot buttons. And it sounds like a pretty stunning result:
simply reversing which sex rotated demolished [a] well-established sex difference
What the researchers found (and I can’t look at the source paper, much to my chagrin, because it’s still only in press) is that when you have blokes sitting at tables in a speed-dating event the well-documented ‘pickiness’ of women disappears. So they say.
tinkering with the speed-dating format alters human behaviour, dramatically changing the outcome.
But hang on a minute. Let’s forget about the sort of (sad, friendless) people who might be tempted by the offer of speed-dating opportunities on university campuses in return for the right to analyse their dating behaviour, and how that might slew the data from the outset, and look at the numbers.
When women sit, and men ‘rotate’ (as is, apparently, usual at these sorts of things), men said ‘yes’ 50% of the time and women said ‘yes’ 43% of the time. Well, to this biologist’s eyes that’s quite a small difference. It might be statistically significant, but the true, biological—or even psychological—significance seems small. You have to remember that one of the last experiments I did in the lab was looking at changes in exon expression, where anything under a two-fold change was considered noise and ignored (at least in the first round of analysis).
When they made the blokes sit down and the women do the walking, there was an ‘astonishing’ effect. Men said ‘yes’ only 43% of the time and women 45%. So, the selectivity—or ‘pickiness’—of women didn’t change, and, let’s be honest here, neither did the men’s. Not really. 50% down to 43%. That’s ‘astonishing’?
And what’s all this guff about ’embodiment effects’ that should be explored further? Isn’t that what you might expect Father Jack to say in his more lucid moments?
Maybe I’m missing something, but frankly, it would be nice if psychologists would work on something that might make a difference. The only conclusion I can draw from this ‘research’ is that speed-dating agencies might be better off by making sure men sit down at these cattle shows: because that way they get 14% more repeat custom. Or something.
<span class="Z3988" title="ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.jtitle=Psychol.+Sci&rft_id=info%3Adoi%2F&rfr_id=info%3Asid%2Fresearchblogging.org&rft.atitle=Role+reversal+undermines+speed-dating+theories&rft.issn=&rft.date=2009&rft.volume=&rft.issue=&rft.spage=&rft.epage=&rft.artnum=&rft.au=Finkel%2C+E.J.%26+Eastwick%2C+P.+W&rfe_dat=bpr3.included=1;bpr3.tags=Psychology”>Finkel, E. J. & Eastwick, P. W (2009). Role reversal undermines speed-dating theories Psychol. Sci
Changes in cell activity or something in a laboratory are one thing. How whole actual people operate in the real world is something quite different.
A 7% difference in behaviour in a real-world situation is actually quite large.
Imagine an additional 7% decrease in smoking prevalence in women caused by some simple change in proceedure like who sits first or who speaks first?
Speed dating is a trivial example but using tightly controlled biological effects in laboratories as a yard stick for the importance of real-world effects is a little silly.
Small effects over millions of people can have very large public health effects. Conversely, large tighly controlled laboratory findings often turn out to have no real world applicability.
Then again look at their sampling of undergraduates and marvel at the lack of real-world applicability…
The numbers itself are completely irrelevant without errorbars.
Well, there is that: but I can’t see the original paper. And I suspect that with only 15 events and 350 participants error bars are pretty meaningless anyway.
the sort of (sad, friendless) people who might be tempted by the offer of speed-dating opportunities
a bit harsh, no?! I’ve heard it’s really fun. Apparently the trick is to take a friend of the opposite sex with you, to spy on your potential match-ups before and after the actual “dates” and pass on the inside info. You do the same for them. Although one friend took this too far when she went to an event with a big group of women from work… and about 60% of the men present at the event were their male colleagues!
But yeah, it’s hardly earth-shattering research, is it?
Isn’t “sad, friendless” a case of potential libel?
I’d say it was fair comment.
I thought they were just all reporters for community- and student papers.
which sounds plenty skewed to me.
Come on. A 7% difference in a totally artificial situation meaning anything in ‘real life’? Worse, deriving principles from this (which is what they’re doing).
The other thing I didn’t bring out was confounding factors. Yeesh. Let’s just not, eh?
I’m not saying this is worthwhile as a study. I’m just saying that a 7% difference cannot automatically be dismissed as being unimportant- like my smoking example.
In other news I’ll be scouting the brewing delights of Seattle next week.
Before knocking on speed-dating as a method for examining how people meet, interact, and make decisions about attractiveness and the future viability of a partner I would suggest reading the methodological papers published on this topic. They are not in press. They are published. They also suggest that speed dating populations do not deviate from the general population, except they are single (which is the population the researchers interested in studying) and (if I remember correctly) they are more extroverted. Hardly the individual difference confounds you suggest. Obviously no situation is perfect, but if you have methods to improve the internal validity of an interpersonal relationship study that can include randomization of experimental conditions, then please be my guest.
Second, I am unsure of your field of expertise, but effect sizes in the social sciences are generally small. People come with lots of random error, thus masking potentially large effects. Many respected statisticians have spoke/written on this (I believe the most influential was Cohen, 1988, but I might be mistaken. I haven’t looked at this paper since introductory stats).
Finally, for some time now people assume that women are more selective then men for evolutionary reasons. This kind of assumption is based on what scientists theorize about evolution as well as observed sex differences (the 50 – 43 difference in the study). By making that difference go away it suggests that social norms and socialization are also important factors and that men and women are more similar than different. Besides, reversing or making an well known/established effect disappear by a very subtle manipulation (so subtle you think it means nothing) is always a big deal. Especially since the manipulation represents a social norm embedded in Western (most? all?) culture. That is, the male should approach the female and not vice versa. This is probably not the perfect manipulation of this process, but we are talking about human subjects and there is only so much a researcher can do ethically and legally.
Of course there could still be plenty of things wrong with the study (I haven’t read it either, but did hear a talk that mentioned it briefly), but the points you pick from your read of the media don’t seem to reflect concerns that have not been addressed previously.
Brilliant — you see, I’m not just interested in how scientists communicate with the public, but in how scientists from different disciplines communicate with each other. To me, this study seems like so much hand-waving and behind-the-hand sniggering.
Telling us how I’m mistaken, educating me, is a good thing.
@Nathaniel In other news I’ll be scouting the brewing delights of Seattle next week.
I would highly recommend, if it’s still there, the Red Door Brewery in the Fremont district.
When I was an undergrad, I took a semester course called The Psychology of Close Relationships. At the time I was doing a biology degree. I have to say that I was impressed by much of the methodology of these sorts of experiments, given the restrictions on human experimentation. I believe even experiments on undergraduates are likely to have merit if robust and reproducible — as in many of the findings I discussed about male/female bias, earlier.
The source paper should be available if it was written about. The journal concerned will have said “no embargo”, so it should provide a copy to anyone who wants it. I think this is becoming increasingly common for journals that have a large print backlog. (Anecdotal obs, n vague, no error bars.)
Hrm. It was ‘in press’ but I couldn’t find it on the advance section of the Journal website.
Someone has blundered?
bq. 50% down to 43%. That’s ‘astonishing’?
The odd ratio is 1.33. For epidemiologists this is a fairly large effect. Remember that there are a pile of other factors that influence the choices, so something so inconsequential has to be compared to (for example) bad breath and interest in the lbw laws to get an idea about how odd it is.
Yep. If a journal opens up a paper so the press can report on it, it should be open to any of the journal’s readers once the press reports are “out there”. Unless its an embargo break, which I doubt in this case, given the journal doing the reporting.
As the editor that commissioned this piece, I’d like to clarify a couple of points. Matt obtained a copy of the paper (accepted but unplublished) direct from the researchers at a conference – not from the journal. He checked with the researchers to see if it was OK to report on the paper and they said it was fine – the journal had not forbidden them from speaking to reporters while their paper was in press.
We should also be clear about what an embargo policy involves. Some journals (including Nature) make their papers available to journalists a week or so before publication under the condition that they don’t publish any stories on the work before the embargo lifts. Psychol. Sci. doesn’t use this system as far as I’m aware. Matt heard about the findings thanks to good luck and good contacts – not an embargoed press release.
After our story went live, the journal press released the paper (without embargo) and it was subsequently picked up by some other publications. The paper will appear in the next issue of the journal. But anyone who wishes to see the paper before then can download it from one of the author’s websites, here.
On the story itself, earlier speed dating studies had claimed men were less picky than women – thus strengthening various conclusions about the evolution of behavioural differences between the sexes (I seem to recall that these findings have appeared in a number of books about evolutionary psychology, as well as newspaper articles). I felt a paper that cast some doubt on these conclusions would be of interest and deserved some play.
Cool! Thanks for contributing here and giving us that perspective.