…says psychologist dressed in all black standing with a clipboard outside a punk rock concert.
Or, young people don’t really care about “conformity”, based on mostly dubious survey data.
“A Critical Test of the Assumption That Men Prefer Conformist Women and Women Prefer Nonconformist Men”
This paper involves discussion of statistics which can be confusing if you don’t work with it a lot. I’m probably not going to do a satisfactory job if I try and explain it piecemeal in an essay, so I very much recommend watching the following videos for a good primer on the subject:
So I realise laughing at a SciAm article doesn’t actually constitute a thorough critique of a published study. Fine.
“It is frequently presumed that women are attracted to non-conformist men and that men are attracted to conformist women.”
So I can’t be too mad about this, because a lot of science is just testing out ideas that pop into your head, and the scientific writing culture demands that you start your topic with how you’re exploring a yet unanswered question in the literature. But honestly, that’s not how most, if any people form ideas on what to study. They get the idea and then look to see if it’s been done before. Not saying that nobody combs the literature first, but it’s not the norm. Here they’re being upfront about “so I heard…”. Nonetheless, outside of teenagers and people with the emotional maturity of teenagers, I’ve never heard anyone say they presume this. So I kind of want to know where they’re getting this. And it’s less that I want bullshit references to literature–this is not a knock at the authors, the bullshitting is in the culture of science-writing at large–than I just want to know who they are hanging out with.
“But is this true, or is it an anachronistic myth?”
Note the phrasing. It is either true or a myth. What it cannot be in this dichotomy is something that people have never thought. I’m not trying to be haughty when I say that I genuinely haven’t heard any of my friends state this sentiment explicitly. In high schools, yeah, sure, but are the nation’s 20-somethings really concerned about this when they swipe right on their Tamogachis or whatever?
“If one were to draw exclusively on the psychological literature, one might presume the former.”
They then list a number of articles that go back as early as… 2006, i.e. 9 years before they published this. “Anachronism” seems pretty generous… Also, they use MLA formatting (or some variant), which, holy shit, that’s annoying. That’s the anachronism…. I didn’t bother reading any of those citations, which is exactly the point of throwing them out there: you are meant to think we are working with established facts and merely adding a little to what is already known, not just launching off into the wild blue yonder. Again, not an invention of these authors; that’s one of the dirty little secrets of science writing.
“Sharpening our theoretical understanding of the link between conformity traits and attraction also carries applied importance. Relationship success is a key driver of overall health and happiness, and failing to attract a partner is associated with low subjective well-being (Dush & Amato, 2005).”
Just want to draw your attention to this. Again, this is not unique to our authors, but a convention of the genre. “what’s the practical application of your research?” is what the people who write the grant money checks ask, so you put on the hat and you dance your ass off. Pull up an article on cancer research, and it’s the same thing. “Cancer is a disease that takes away jobs from hard working Americans (Someone et al.)…”. The fact that they straight-up say “implied importance” is the functional equivalent of ending your English paper with “In conclusion…”. It’s inelegant, but they have to put it in there, so I don’t fault them in the slightest.
“But if this assumed wisdom turns out not to be true, then it suggests that many people might be engaging in impression-management strategies that are ineffective, or even worse, counter-productive.”
This is what the media will read, and this is what will ensure they have grant money next year. I am not yet cynical enough to ascribe this to conscious, malicious intent, but I’m not so Pollyannaish as to dismiss the role of the unconscious mind in writing this. Or we can blame the culture. Science funding is a bitch and a half to procure. If pandering to the lowest common denominator keeps you afloat, you tend to do it. If we as a society want to stop the flow of crap into the media sewer, we have to start cutting personal checks to research.
The next six paragraphs are completely superfluous in my assessment. Again, hard to determine how much of that is padding because that’s how you get published and padding because of a conscious or unconscious desire to assert pretty dubious experimental data as hard, empirical facts, but it’s still too much.
“In the five studies that follow, we take a broad definition of (non)conformity, incorporating measures and manipulations that include standing out from the crowd, emphasizing uniqueness, and sticking to opinions in the face of pressure from others.”
https://deadhomersociety.files.wordpress.com/2014/07/the-old-man-and-the-lisa13.png
“Which do you think is more important: hard work or sticktoitiveness?” – Principal Skinner
“Are there any real questions?” – C.M. Burns
How do you measure “standing out from the crowd”, “uniqueness”, and “sticking to opinions” objectively? Also, how do we know this is what non-conformity is? I’m not trying to be whiny, but these are super nebulous terms, and I don’t think it’s unfair of me to want to know how you can come away with definitive statements about these vague personality traits to the point where it can be called “scientific”.
“Method: Participants: 74 white undergraduate students (46 female, 28 male), heterosexual, ages17 to 28 (average 20 y.o).”
So right off the bat, a paper that talks about “men” and “women” as a whole, is basing this on at least one study that a surveyed a small handful of college kids. But that doesn’t make it into the title, or the introduction, or (spoiler alert) the study conclusion or the paper conclusion. Now, maybe college kids are functionally identical to all adults when it comes to dating. I don’t have data to disprove that. Of course, I didn’t see any reference to data supporting that presented by the authors, and considering that my own personal experience has taught me that not all adults want the same things as 20 y.o. college kids, I’m going to have to ask for some pretty solid evidence before yielding on that. So the sample from which they’ve picked this first study in my eyes, and I think most peoples’ if they were queried, is enough to throw out this entire paper. But I’m a trooper, so I’ll keep going.
“Materials and procedure: Participants were instructed to imagine they were single and wanted to start dating,”
Two things. So one, this seems to imply that they did not identify the relationship status of their sample pool. You think there’s a difference in what a gal in a 6-month relationship and one who’s never dated think about a potential dating partner? More grounds for dismissal in my book. The next big thing is the word “imagine”. Hugely overrated and pompous song. Also, yeah, what a 20 year old guy who’s never dated imagines what he wants in a partner is probably not going to give you universally applicable dating info. I move for dismissal, dammit!
“… and were presented with 39 items designed to assess mate preferences for conformity behaviors…. After recording their own preferences, participants reported how attractive other people of their own gender would find the 13 characteristics, and how attractive people of the opposite gender would find the 13 characteristics. All items used a 7-point scale (1 = very unattractive, 7 = very attractive). The scales were calculated such that higher scores indicated preference for romantic partners who conform, whereas lower scores indicated preference for nonconformist partners.”
“Mate”, like this is a nature documentary. Hm, I wonder if David Attenborough would be willing to narrate. “What do you look for in a mate?” “Well, top-tier cheeky banter at Nandos…”. No, it’s not what they look for in a partner, but a mate. This is a trifle, sure, but the word does have a connotation in the States for breeding. If that’s the word they used, did they just prime the subconsciouses of their subjects to think more about fornication than long-term partnership? I’m mean college kids are going to be bad enough as it is…. I also find the 1-7 scale interesting. It seems to pop up a lot in psych studies. I’ll reiterate the argument I’ve made elsewhere that ends up yielding statistically significant results when a scale with a larger range would not. I’m not sure whether this is conscious, willful intent to warp data or unconscious, but I do find it a bit dubious.
So to two significant figures, “men” are dead-center apathetic about conformity in women (3.5 on this scale), and women don’t seem to care either (3.4). Or the scales also don’t accurately determine conformity. They both think that men will find conformity “slightly attractive”, and they think women find conformity “kind of slightly attractive. But that won’t any play after this hits the science media, so fuck it, turn on the old ANOVA machine and throw a bunch of weird numbers and Greek symbols at people because SCIENCE!
“In sum, both men and women showed a preference for non-conformity relative to what they thought other men or women would prefer.”
That is, there’s a differnce betwen their preferences and what they think other people prefer. By how much? Men think other men rank conformity as .66 points more attractive on a scale from 1 to 7. That is, while a handful of young college men find conformity neither attractive nor unattractive, they think that other dudes might find conformity “slightly attractive”. Wow! What a huge difference! Same goes for women’s assumptions about men by .60 points, and women’s assumptions about other women by .46 points. It may be statistically significant based on the tiny scale they picked, but no person who’s not a complete idiot would honestly argue there is functional significance here. I.e., study one proved than neither college males or females care about conformity in an imaginary date, but they think other boys might care a little bit. Fucking groundbreaking.
Study 2:
“Revealed partner preferences (what people actually choose in a partner) are often at odds with stated preferences.”
It’s good that they realise this. Thought experiments are one thing, but the real world is where it counts. So I assumed they’re going to follow their kids and see what kind of people they actually date?
“Participants were presented with descriptions of targets of either the same or opposite gender, who self-presented in a conformist or nonconformist way. Participants then rated each target according to how romantically attractive they found the target, and how romantically attractive they thought others would find the target.”
Oh. Okay, so the same thing, just now they have a picture and a blurb. Yeah, so this is not a revealed preference, this is another stated preference. This is kind of a major distinction that I feel like someone should have caught.
New survey cohort: 68 females, 47 males, mean age still 20.
“Procedure: Participants were presented with profiles of 20 people, each profile comprising a brief description of the person’s personality accompanied by a photograph. The 20 profiles were arranged in pairs, and each pair of profiles described a conformist target and a nonconformist target. Half the participants viewed profiles of opposite-gender targets, and half viewed profiles of same-gender targets.”
ACCOMPANIED BY A PHOTO!?!?!?! For fuck’s sake.
You know which one I find more attractive? #2, because CLEAVAGE. I didn’t even read the descriptions, and I picked 2, because I am a normal human. I’m not saying she’s objectively more attractive, but I really hope it’s obvious that a picture’s worth a thousand very persuasive words when it comes to ranking attractiveness. It boggles my mind how the authors are missing this massive, massive observation. This is why you have to read the studies. The problem with peer review is that if all the peers are equally oblivious, worthless studies get through.
Also, these are just terrible profile blurbs. It’s the kind of thing that sounds like it’s a phishing scam. If all I had was the blurb, I’d swipe right. But they put the picture. According to the geniuses who crafted this study, I’m attracted to non-conformity. Not boobs. Because “men like tits” won’t move papers. I mean, the conclusion won’t, the premise has done a number of people in various media quite well…. Also, the birthday. Why is that necessary? You run the risk of people into astrology inadvertently causing bias. probably not a ton, but it’s an unnecessary confounding variable.
Well, let’s look at the data anyway:
So conformity is now somewhere between “slightly attractive” and “attractive”, or at least the pictures of people that the researchers clipped a blurb calling them conformist. These are all attractiveness ratings 4.5 +/- .4 on a scale of 1-7. I wouldn’t call any of these differences functionally significant. I mean, the perception differences, maybe, but it’s super slight. Again, similar conclusion. Guys and gals rate conformity as slightly attractive, think their peers care about it a little bit more. This isn’t headline-grabbing. which is fine, it’s just that the authors and the media went overboard with their sales pitch.
Study 3:
“Study 3 extended Study 2 by examining (non)conformity preferences among men and women during an ostensibly live group interaction. This revealed preferences paradigm allowed us to observe mating preferences during real-time interpersonal interactions, rather than while viewing static dating profiles.”
Making progress. Maybe by the end they’ll look at actual goddamn dating preferences made in real life. We get more demographic info on this group. 59 females, 52 males, 2/3rds white, the rest are mostly Asian; 72 are single, 35 in a relationship, 3 married, 1 divorced. Grab yourself a beverage and strap in for the methodology (you can just skip to the picture if you’re feeling lazy).
“Materials and procedure: Participants were informed that they would be completing a study on art preferences and that they—along with other participants in different laboratories—would evaluate a series of images. They were led to believe that the experiment would be conducted via an interactive ‘chat’ program, where each participant could view other participants’ ratings and comments. Four images were presented in the style of an Internet chat-room, and participants were led to believe that they were images of coparticipants in the interaction. The images comprised people aged 18 to 29 who were prerated as moderately attractive (5.40-6.60 on a 10-point scale of attractiveness). Male participants were presented with four female group members; female participants were presented with four male group members. Of these, one group member varied in terms of the extent to which they converged with (conformity condition) or differentiated from (nonconformity condition) the other members’ ratings. This group member was the target evaluated at the end of the experiment. Participants were presented with four black-and-white patterns. After the presentation of each pattern, participants rated it (from 0 to 10) based on its visual appeal. Participants were always (apparently randomly) assigned to make their evaluations last (i.e., after the other four group members had rated). A comment box and a scorebox were included under each group member’s photograph so that participants could view the responses of others before making their own rating.
For participants in the nonconformist condition, the target rated two of the patterns as clearly more visually appealing compared with the other group members (on average three points higher), and two of the patterns as clearly less visually appealing compared with the other group members (on average three points lower). Conversely, for participants in the conformist condition, the target gave a score for every pattern that was within 0.3 of the scores of other group members. No comments were provided by the target group member until the final pattern, at which point the target stated, ‘Looks like I’m going to have to go against the crowd again!’ (nonconformist condition), or ‘Look, I’m happy to go with the crowd again!’ (conformist condition). Following the pattern evaluation task, participants evaluated the target (seemingly selected at random) on a range of attributes. The target participant’s photo was displayed on the screen and randomized so that each photo was equally likely to be associated with a conformist or nonconformist target. A summary table of all participant ratings was displayed before the questionnaire was administered, and the target’s ratings were highlighted in red. Participants rated how interesting, warm, intelligent, likeable, and friendly (all scales of 1-7) they found the target. These items were combined into a single scale of positive regard. To measure romantic attraction toward the target, participants rated the extent to which they found the target attractive (1-7), and the extent to which they agreed with the statements ‘I would like to go on a date with this participant’ and ‘I think this participant would make a desirable long-term romantic partner’ (1-7).
“Nonconformists (M= 5.02) were regarded more positively than conformists (M= 4.65). Participant gender and ethnicity had no significant effects, either alone or as an interaction with conformity.”
If you skipped over the methodology text, the “rating” is actually an aggregate of several different factors of how interesting, warm, intelligent, likeable, and friendly they were rated. That this is equated with attractiveness is a bit of a bait and switch. Somebody saying some mechanical statement designed by oblivious psych researchers about how the look like they have to “go against the crowd again!” might just come off as more weird and annoying than the “nonconformist”, which isn’t a measure of attractiveness, but social comfort.
“On romantic attraction, only an interaction between participant sex and conformity emerged. For men, nonconformist women (M= 4.17) were seen to be marginally more romantically attractive than conformist women (M= 3.50). In contrast, there was no significant difference between how romantically attractive women found nonconformist (M= 3.27) and conformist men (M= 3.63). Ethnicity had no effect, either alone or as part of an interaction.”
Again, on the scale of 1-7, the only one of those rankings that was above pure indifference was the men rating “nonconformist women” as “slightly attractive”. Playing with the very marginal differences is so worthless when the functional differences are nothing to write home about.
Study 4:
“It is an open question whether these ‘clean,’ decontextualized reports of preference would hold up in the messy cut-and-thrust of real-world…”
By Jove, I think he’s got it!
“… mating.”
>:( There’s that mating again. Maybe they’re trying to goad National Geographic into picking up their paper. How about dating, or long-term relationships? Hell, they’ve already found some married couples.
“Indeed, a critical indicator of whether people are attracted to conformist or nonconformist people is whether conformists and nonconformists are successful in their romantic endeavors. Therefore, Study 4 was designed to switch perspective and test whether individual difference variables conceptually associated with conformity and nonconformity predict real-world dating success.”
All right, so they’re actually going to look at some actual relationships. That’s promising. If they do a decent job with the setup, this might be worth a glance.
“Study 4 sampled from both Western cultures (United States and United Kingdom) and from a collectivist culture (India), where one might expect that the preference for nonconformity found in Studies 1 to 3 could be reversed. If it is true that men have a preference for conformist women, it might be expected that this would be particularly evident in a country like India, which is still governed by relatively traditional gender roles.”
So this is weird to me. It makes sense that they want to look at multiple cultures, but this raises a red flag for me. The samples from the first 3 studies were Australian University students. I get that the Aussies may not speak for everyone, but the consistency was good from a design standpoint. Using multiple cultures adds a lot more variables. And while I applaud the ambition, this is now apples to oranges. We’re no longer looking a specific, more homogeneous group of people but we’re lumping the results into the same paper. That seems kind of sketchy.
In this sample group, we have 821 people, 457 of whom are male and 364 female, 515 from the “US/UK”–because, you know, those are basically the same country– and 306 from India. Average age is 28. There are two scales used to detect conformity, and two for non-conformity that were adapted from other researchers. For conformity, they query for “Independent self-construal” on a scale from 1-7–“Items from this scale include, ‘My personal identity, independent of others, is very important to me’ and ‘I act the same way no matter who I am with'”–and “idiocentrism”–“Items include, ‘I often do my own thing’ and ‘One should live one’s life independently of others'”. For non-conformity, they query for “Interdependent self-construal”– Items from this scale include ‘It is important for me to maintain harmony within my group,’ and ‘It is important for me to respect decisions made by the group'”– and “allocentrism”– ”’I usually sacrifice my self-interest for the benefit of my group’ and ‘I hate to disagree with others in my group'”. I realise that’s a little tortuous to read, but notice how these are not mutually exclusive. Notice the “my group” in the items that ostensibly test for “conformity”. In my experience, conformity is doing what everybody is doing. My group is not everybody. You can still think yourself non-conformist by bucking the crowd but also think it permissible to be loyal to your friends. One could argue that’s hypocritical of people, but nonetheless, they may still feel that way.
The folks were also asked to rate a few ways of measuring how successful they felt their various kinds of relationships were on scales of 1-5. So if they were in a casual dating relationship, basically, do they feel it’s going well. They bolded all the results that yielded statistical significance. The β is the regression coefficient (similar to an r^2). β tells us how much one variable influences the other, and is measured in units of standard deviation. So a β of 1.5 indicates that a change of one standard deviation in one variable will result in a change of 1.5 standard deviations in the other. So a larger β means there’s a larger effect from the variable.
Even though they got a number of hits with statistical significance, these regression coefficients are all terrible. The highest were .21, just a hair over fifth of a standard deviation. That’s glorified noise. These effects, while detectable, are minute. The authors conclude:
“Thus, the relationship between success and nonconformity mirrored the preferences observed in Studies 1 to 3.” I agree, they mirror the complete insignificance. At best, they show a mild relationship between people who rank themselves as “independent” and having success in “casual relationships” and a mild relationship between sex (as in gender) and “cultural attributes of collectivism” in relationships for casual sex (for both sexes, but only for Indians) and also in committed relationships (for males only). Additionally, “where interdependence was high, the relationship between independence and success was positive and significant (one-time sexual encounters: β = .35, casual sex: β = .25, casual dating: β = .26)”. Again, these variable relationships are so small that without emphasis on qualifiers like “slight”, the results are misleading. What I absolutely did not get out of this study is anybody really caring about conformity or non-conformity in any meaningful way in long-term relationships, which is what everybody thinks of when you talk about “what men/women want in relationships”. The fact that casual hookups are thrown into the statistics orgy is dubious to me.
Study 5:
“In Study 5, we sought to complete our research by taking into account the dyadic nature of romantic interactions.”
Dyadic interactions are interactions between two people (the prefix dy-/di- meaning 2). E.g. Itchty and Scratchy compose a dramaturgical dyad.
“In Study 5, participants judged the level of conformity in both their current and ex-partners. They also rated their level of attraction and attachment to these targets.”
We asked people to rate their exes. These results will be totally objective and fair in every way.
“Dyadic data are costly and time-intensive to collect.”
Translation: “Ugh, that sounds like something approaching actual work. I didn’t get into sociology/psychology to talk to people! Can’t we just have some undergrads fill out a survey and call it a day?” This is not in the authors’ defense, but from a regulatory standpoint, interviewing people is still considered research on human subjects, and doing it by the book does involve a pretty terrible amount of paperwork, and by extension a fair amount of expense.
The surveyed folks were found through a work-online service from Amazon, where you can basically fill out surveys for minimum wage. Nope, no sampling bias to be found there. “To minimize the extent to which the ratings of the ex-partner were contaminated by bitterness about the break-up, we measured and controlled for ‘who broke up with who'”. I literally laughed out loud when I read that. For whatever reason, the authors broke with the earlier studies and allowed both heterosexual and bisexual people answer. I don’t have anything against bisexuals–some of my favorite porn actresses are bisexual!–but in the event that they might have different personality preferences than heterosexual people, then we’re mudding the waters and less (appropriately) able to make statements with definitiveness. Nonetheless, they got 185 males and 125 females with an average age of 32. 294 had an ex and 243 had a current partner. So pretty different from our college kids back in studies 1-3. “Nonconformity” and attraction were rated on 1-7 scales.
Results: No relation between conformity and partner attractiveness, “however, participants reported feeling more attracted to their ex-partner the more nonconformist their ex-partner was rated to be, β= .18”. So again, piddling amount.
We then get a rather hefty 12 paragraph conclusion discussing the results of the study as a whole. Weirdly, none of them mention how most of their data show that people care very little about conformity (or rather, their metrics therefore). Because this was not about forming conclusions that best fit the data, this was about cherry-picking data that fit preconceived conclusions. In the process of writing this, I was talking to a friend of mine who’s an analysts with a formal education in statistics. I’ve been complaining about statistical values in these papers (and others) that claim to show significant relationships when that’s still very much up for debate. My buddy gave me the depressing news that in social sciences, regression values of .2-.4 are considered completely acceptable for demonstrating a relationship, which just boggles my mind. Different disciplines have different standards for what they deem to be meaningful relationships. A math teacher’s website provides the following handy illustration. The higher up the chain in sciences you go, the larger you need those values to be. I’m in medical biology, and while someone might entertain an r^2 of .4 as showing a possible connection, my boss, should I come to him saying I found a “correlation” with that number, would tell me to get the hell out of his office. You want something around the .8 range. My buddy in mechanical engineering prefers it to be at least a .9. So perhaps it’s elitist of me to thumb my nose at a .2; at the end of the day, it’s all up to subjective interpretation anyway. But justifying that interpretation, even if ultimately impossible, is still a necessary part of communication in science. Scientists, and in my experience the social scientists seem the worst at this, do a very sub-par job of effectively relaying the uncertainty to the world. And we don’t stop them. Because uncertainty isn’t satisfying; it’s terrifying, even. But that’s what we have. And people deserve to know the truth to the best of our abilities to determine it. That’s what science is all about.