Andrew Sullivan links to this Scientific American article, entitled “Men and Women Can’t Be ‘Just Friends’.” Adrian Ward, the author of the article, probably didn’t write that headline, but it’s a fair summary of his article:
The results suggest large gender differences in how men and women experience opposite-sex friendships. Men were much more attracted to their female friends than vice versa. Men were also more likely than women to think that their opposite-sex friends were attracted to them—a clearly misguided belief. In fact, men’s estimates of how attractive they were to their female friends had virtually nothing to do with how these women actually felt, and almost everything to do with how the men themselves felt—basically, males assumed that any romantic attraction they experienced was mutual, and were blind to the actual level of romantic interest felt by their female friends. Women, too, were blind to the mindset of their opposite-sex friends; because females generally were not attracted to their male friends, they assumed that this lack of attraction was mutual. As a result, men consistently overestimated the level of attraction felt by their female friends and women consistently underestimated the level of attraction felt by their male friends.
So the study must have found huge differences between men and women, right?
Well, let’s take a look at what Ward didn’t report: the numbers (pdf link). First, Ward reports that “men were much more attracted to their female friends than vice versa.” The study authors asked their subjects1 to rate their attraction on a scale from 1 (“not at all attracted”) to 9 (“extremely attracted,”) with 5 defined as “moderately attracted.” Here’s what they found:
The average report for women was 4, versus 5 for men. Ward could have more accurately reported that men’s and women’s average responses were “almost alike.”
Now let’s check out the differences between reported and perceived attractions. Using the same nine-point scale:
Again, there is a difference here, but contrary to Ward’s reporting that difference is modest. Men perceived their female friends as slightly less than moderately attracted to them (4.5), when those female friends were in fact… slightly less than moderately attracted to them (4). Women perceived their male friends as slightly less than moderately attracted (4.5), when those male friends were in fact moderately attracted (5). On average, women slightly underestimated attraction and men slightly overestimated attraction, but on average both women and men were only a half-point off the mark, and the differences between estimates and reality fell within the margin of error.
There is one graph in the study showing a large difference between the sexes, which is on the question of if people see sexual attraction to an opposite-sex friend as a cost or a benefit of the friendship.
All study participants were asked to list up to ten ways that their “cross-sex friendships enhanced their lives or were beneficial to them,” and up to ten ways “cross-sex friendships complicated their lives or were costly to them.” Although both men and women were more likely to list attraction as a disadvantage, there was a notable difference in the percentages — women were much more likely to list attraction as a disadvantage.
There was also a notable similarity – the majority of respondents, male or female, didn’t list attraction as either a cost or a benefit.
Ward opened his article by asking “Can heterosexual men and women ever be just friends?” This study did address that question. Of the college-aged respondents, 98% of men and 97% of women said that they had opposite-sex friends. Among the respondents age 27-52, 86% of men and 88% of women reported having opposite-sex friends.
In other words, the overwhelming majority of both men and women reported that they have opposite-sex friends. Strangely enough, Ward didn’t find that result worth reporting.
- The sample for this part of the study consisted of an unrepresentative sample of 88 heterosexual students at the University of Wisconsin, all fulfilling a “course research participation requirement,” race not reported but I’d bet a comic book they were nearly all white. For a later part of the study, involving adults age 27-52, subjects were recruited by asking the students from the earlier part of the study for names and addresses of older friends and relatives. 80% of the participants in the second half of the study were from just a couple of states in the Midwest, nearly all were married, and once again race was unreported but… [↩]
That’s… really irritating.
Of course I don’t have any opposite sex friends, personally, because boys are icky. Especially Amp.
Using undergrads as a sample and then generalizing to the population as a whole is problematic for a whole lot of reasons. I seriously doubt that most people would say that the sexual dynamics of the undergrad years — away from home for the first time, easy access to alcohol for usually the first time, many people living in co-ed dorms, spending pretty much all day with mostly-single people of about the same age, and for many of them, still lots of puberty hormones going wild — are really anything at all like the sexual dynamics of the rest of their lives.
For the love of…something, please, somebody, make science journalists take a basic data analysis class…
Even ignoring the biased samples, based on the reported confidence intervals there’s about a 20% chance that, in general, heterosexual men and women are correct about how attracted their friends are to them, and this study just happened to include a group of people who weren’t. Effects that seemed more certain than that disappear all the time–there are lots of published and unpublished studies conducted each year, after all. The self-reported attraction level to the friend is only about 0.5% likely to be equal between men and women, however.
A further source of possible bias: these were students in a class asked to bring a friend along, not a survey of all the opposite-sex friends these students had. So this is the attraction level between the student in the course and a friend they could convince to come along with them–a subset of friends that, I imagine, might have a very different distribution of attraction levels than their overall population of opposite-sex friends.
I also note the annoying elision of “heterosexual” from the headline, though of course it’s included in the first sentence. But everybody knows what they mean, right? Bah.
Classes can’t fix stupid. Many journalists are simply dumb. To present this material as it was presented requires either an absolutely fiendish commitment to an explicitly dishonest narrative, or garden variety stupid. I prefer to believe the latter; then the problem is a mismatch between abilities and employment, rather than someone being in the employ of His Satanic Chocolate PopTart-Eating Majesty.
I’ve never, ever, understood the reason for these stories. I mean, it’s one of those topics where you have to ignore the blindingly obvious to even ask the question. I’m aware of literally dozens of friendships between heterosexuals of opposite sex, and my sample size isn’t exactly huge.
But not only is this a question, it’s one that is persistently “answered” in the negative. Why? Can anyone explain to me what is the motivation behind this? Normally if there are persistent studies that go counter to easily attested life experiences then there’s some sort of ideology behind them. But, in the Western world, at least, I can’t think of an ideology that can benefit from arguing that male/female friendships are impossible that’s anywhere near mainstream. Am I hopelessly naive?
Well, it kind of goes along with the “guys are always thinking about sex” idea, and thus ties into a lot of rape culture — “You have to always be careful about men — even the guys that you think are your friends actually just want to have sex with you.”
Oh, look, they’re trying to do statistics. With t-tests. Adorable, it’s almost like they’re real scientists.
Oh, and they used least-squares regression rather than orthogonal regression for non-causal associations. With no correction for multiple tests at all.
Give me 6 months, some custom stats software with extra-big buttons, and a bag of M&Ms and I can have a rhesus monkey correctly avoiding their blunders. 6 more months and I can realize my lifelong dream to have a monkey that savagely attacks anyone who doesn’t use a mixed-model repeated-measures ANOVA when they should (not that these authors should have, clearly; an infant could see a nested approach would be the correct one here).
It’s like watching someone who’s only built tiny model rockets try to figure out how to design a mixing nozzle for a Saturn V.
Mokele: Your 6 months and a bag of M&Ms comment made me laugh out loud.
yrs–
–Ben
Can you explain to an ignorant person like me what this means? Feel free to talk to me like I’m a small child. :-p
Mokele@7: That is the standard of the field in most social science fields–something that should certainly be changed, but it’s not something that makes this extra-laughable compared to other studies of this type.
In related news, I’m currently enjoying the bizarre phrase “cross-sex friendships are a historically recent phenomenon” in the abstract of the paper. (I poked a historian friend about this and she said, essentially, that friendships of any kind as we conceive of them today are a historically recent phenomenon.)
@Eytan Zweig, unfortunately there are many people for whom performative gender differentiation is not merely a fetish, but something so dear to their hearts that they are uncomfortable unless they believe it is everyone’s fetish.
Amp @9 – Statistical tests are not interchangable; each test you run has been designed based on particular assumptions about the data. Furthermore, they are not foolproof, and each test introduces a chance of false positives (i.e., an outcome where a random result appears to be a significant difference). If you run a lot of tests, the likelihood of false positives compounds and the more tests you run the more likely at least one is to be wrong. Statisticians who do they job properly are aware of this and introduce corrective mechanisms to their tests that take into account how many tests are run.
Mokele is pointing out that the authors of this paper used statistical tests that make inappropriate assumptions about the type of data they are dealing with, and they did not introduce any corrections for the amount of statistical tests they ran. Which means that they were far more likely to generate false positives.
As Harlequin pointed out, while this is lamentable, it’s something that’s very common in social science publications, and not a specific issue with this paper. I’ve been guilty of it myself in my grad student days (or at least, I’ve been guilty of using statistical tests by simply aping other publications without understanding how they worked).
Mythago – Yeah, I know that. But I still don’t get the conclusion. This isn’t “Men are from Mars and women are from Venus and thus friendships between the two include a lot of miscommunication” or whatever.
Back when I lived in Israel, there were some people who liked to write articles about how Kosher laws were justified by all sorts of health concerns. They invoked all sorts of pseudoscience to explain why every time you mix milk and dairy you shorten your lifespan, and how secular Jews are foolishly falling into traps laid for them by their envy of gentiles.
That was silly, and transparently agenda-driven, but it was understandable why people who wish to push the Kosher agenda would both believe it and/or use it as leverage for their views. They still didn’t take the extra step of “and therefore, cheeseburgers don’t exist”. That’s what I don’t get,
But perhaps I’m overthinking this.
@Eytan Zweig I think it might actually have a lot to do with the “men and women are from different planets” idea, in that there’s an assumption that due to the ‘fundamentally different goals’ men and women have, the participants aren’t acting in good faith. That is, the men really just want to have sex with the women and would prefer if there was no companionship of any other flavour, while the women just want a boyfriend surrogate – someone to tell all their feelings to – and either they ‘can’t get’ the real deal, or they have a bf but then this external relationship is practically considered cheating. So, both participants in the friendship are using each other for mutually exclusive ends, and thus said friendship is invalid. All of which is wrong in a bazillion ways, of course : (
As a personal aside, I especially resent the whole ‘if you are generally attracted to people of x gender, then it’s impossible to have a meaningful/fulfilling relationship with anyone of that gender, especially if they are attracted to your gender’ pile of nonsense; the implication that simply because I’m bisexual all my friendships are fraudulent and worthless (or at least worth less) is both preposterous and hurtful.
Also, John Scalzi wrote a piece on the subject (not crappy science reporting/doing, but male-female friendships) a while back, which might be of interest: http://whatever.scalzi.com/2004/04/19/reader-request-week-2004-1-boys-and-girls/
Yeah, I guess I was really being naive. And resentful, as well. I guess that a large part of my reaction is that I simply don’t like being told that most of my friendships are based upon a lie. Or that I am too much a slave to my sexual desires to be able to ever consider (emotional) intimacy and companionship more important than sexual attraction.
Hell, I guess I’m supposed to be threatened by my wife’s male friends now, aren’t I? I was so busy thinking about my own friendships with women, it never occurred to me that every man who my wife might have coffee with is secretly plotting to take her and claim her as their own. I guess I was such a fool to assume that she may actually have the wherewithal to realize what was going on in her life.
@Eytan Zweig: because it was not threatening to the pseudo-science people if Gentiles ate cheeseburgers. They were not threatened by cheeseburgers existing or that people other than Jews ate them. Whereas for gender- dichotomy fetishists, there seems to be a profound need to deny that their preference is just that; they’ve built their interest on a belief that the world IS a particular way.
Mythago:
Yes. The pseudo-science people do not see themselves as complementary to cheeseburglars, as yin to their yang, or vice versa.
Dichotomists, on the other hand, see their own selves as part of a complementary pair. If you force a change in their perception of the sex which they are not, then you necessarily force a change in their perception of the sex which they are, of themselves.
In other words, any change of world-view which impinges on gender, which has its tentacles in freakin’ everywhere, threatens a change in self.
Scary stuff. Go down that road and people have a lot of power over you. Better to ditch the whole construct and work, as much as possible, with reality.
Grace
EZ –
Yes, this is my answer to the question “why be invested in this framework.” It validates kyriarchical norms about control and policing of women, which is the very mentality that underlies rape culture, family annihilators, and a host of other entrenched, gendered social ills.
Thanks again everyone who responded to me – this has been a very helpful and educational discussion. Not just for the topic itself, but also because it’s yet another lesson on how easy it is for me to fail to connect the dots when it comes to the treatment of gender in society.
All of the criticisms leveled at this study seem on-target: from the underlying assumptions to the methodology to the sample and reliance on self-reporting (always, always biased, even unconsciously) and the inappropriate extrapolation to the overall populace.
Until very recently (the advent of neuropsychology), it was arguably fair to call ‘social science’ a contradiction in terms precisely because of flaws and limitations like these.
As a researcher whose work falls variously under both these fields (social science and neuropsychology), I would dispute both parts of this claim – there was a lot of good research as well as bad for decades, and neurological methods introduced lots of new and impressive sounding ways people could misinterpret and misrepresent their terms.
Yeah … Good social science has means of correcting these errors. Bad neurosociology ignores them.
Also, I think it is important that this study comes from an EvoPsych! background, a scientific endeavor which is riddled with under-researched and poorly performed studies of exactly this type.
The pattern of failures and assumptions doesn’t lie on Sociology, in this particular case.
yrs–
–Ben
“Arguably fair”? What does that even mean? Either it was or it wasn’t. Why the fig leaf?
(That said, it amuses me what things get tagged as ‘social science’ and what don’t; economics is no more scientific than psychology, but it shouts louder about using math.)
Sorry, I should’ve been clearer. I meant that as a comment on social science, it was an exaggeration… but not a very big one.
The development of neuropsychology marked the beginning of the Science of psychology. Before that, we had a field that could be better labeled the guesswork of psychology.
How so? What change did neuropsychology introduce?
Copyleft: So when did biology “become a science” then?
Math does not make a science. Methodological observation makes a science.
That’s so wrong that I can’t tell if you’re trying to be funny or deliberately insulting. (Marriage and Family Therapist here, and faculty in various MFT, counseling, and psychology programs.)
Copyleft, I also have to object to: ” and reliance on self-reporting (always, always biased, even unconsciously)”
I can’t see a non-self-reporting way to measure how attracted people are to their friends. I can see how you might better measure people’s attraction to their friends (getting them to using clickers to mark every time they thought sexually or romantically about a specific friend), but it would still be self-reporting, and would probably suffer from similar problems of bias and self censorship. Also, you’d probably get a high drop out and refusal rate that would seriously distort your results. It would also probably be difficult to get paired friends to do a clicker experiment and one of the virtues of this study’s methodology is that it used paired friends (although it then failed to do anything terribly interesting with that design feature: why aren’t there any charts showing the distribution of differences within friendships, rather than purely aggregate statistics?). Additionally, I think there is actually value in specifically looking at how willing peo0ple are to self report on their relationships in specific ways. More from an actual sociological or psychological perspective than from the EvoPsych perspective of these researchers, but learning how people self-describe their relationships is valuable.
Eytan Zweig: the “neuro” makes it sound more science-y, therefore it is.
Pingback: Can Men and Women Be Friends? — The Good Men Project