If you read enough behavioral-science research, you’ll notice that the subjects are often undergraduates. This pattern makes sense for researchers because (a) there’re usually lots of students nearby and (b) they work cheap. But does it make sense to make generalizations about behavior using only the sample of backpack-bearing humanity that can be found roaming free on our nation’s campuses?
Probably not, argue the authors of a new paper. They say researchers should be looking not only outside of the campus but outside of the country—at least, if they want to make universal claims about human behavior. That’s because people in Western, educated, industrialized, rich, and democratic societies (or WEIRD, to use their clever acronym) are atypical and their responses to even seemingly basic tests don’t match up with the rest of the world.
For example, you know the visual illusion that makes two equal lines look as if they’re of different lengths? Most of us in the industrialized world fall for that one every time. However, the San foragers of the Kalahari are not fooled. They know those two lines are equal.
Or take the so-called dictator game. That’s an economic test in which a participant is given some money and told that the sum either can be shared with another person (known as “the responder”) or can be kept. It’s entirely up to the first person; that’s why he or she is the dictator.
People in the United States tend to split the money roughly 50-50 with the other person. Foragers in Tanzania are much less generous, giving the other person about one-quarter of the sum.
The point is that most of the subjects in such experiments come from a single, thin slice of the world’s population. You might even call them outliers. That is no problem if you’re interested only in that statistical sliver, but you haven’t really discovered something about “people,” per se.
So what to do? The authors of the paper suggest steering clear of universal proclamations and, whenever possible, broadening the pool of subjects.
From the paper:
There are literally untapped billions of people around the world who would be willing to participate in research projects, as both paid subjects and research assistants. The amounts of money necessary to pay people who might normally make less than $12 per day are trivial vis-à-vis the average research grant. Development economists, anthropologists, and public-health researchers already do extensive research among diverse populations, and therefore already possess the contacts and collaborations. Experimentalists merely need to work on building the networks.
Sounds good. Although I have to wonder whether the foragers of Tanzania really want to waste their time with a bunch of psychology grad students.
(The paper, published in Behavioral and Brain Sciences, is available here. The authors are Joseph Henrich, Steven J. Heine, and Ara Norenzayan.)
11 Responses to Why We’re All WEIRD
11245928 - July 2, 2010 at 5:06 pm
Hmmm. Have these guys ever had a basic anthropology class?
stevegaa - July 2, 2010 at 6:57 pm
That is what the “Limitations of this Study” section is for.
princeton67 - July 2, 2010 at 8:33 pm
Who isn’t from a “thin, statistical slice”? Certainly not the “foragers of Tazania”? What group of people could be said to be representative? Ultimately, everyone is unique, a member of an ultrathin slice of one. Even identitcal twins differ by birth order.
mxb22 - July 2, 2010 at 9:30 pm
It gets complicated. Consult the Journal of Cross-Cultural Psychology.
richardtaborgreene - July 2, 2010 at 11:23 pm
MY GOD!!!!! Americans discover the minds and hearts of the rest of the world DIFFER!!!! Wow—amazing—Americans finally finally finally getting a clue!!!! Amazing, after billions of journal articles filled with crap—NSF funded crap. The blind funding the blind. Now that this article has been published, we can ALL go back to cheap, near, easy and WRONG–our basic behavioral paradigm of research for tenured university jobs and status enhancing grants plus the occasional NYTimes article on Diener or Nizbet (the only two scholars the NYTimes has phone numbers for apparently). About ten years ago a paperback was published on some female professor at Harvard whose lifework was celebrated by a conference whose results were the book—she had spent her entire academic life showing how “means” and correlation among means hid the fact that in distributions two or three or more entirely different causal mechanisms got blended into what the mean were doing. When you looked at the lumps in the distributions for patterns within the bottom, middle, and top of the distribution, its extremes, you found entirely different causal mechanisms, missed by all those means correlating with means. Behavioral research is, unfortunate. Take Harvard Creativity research applied at P&G and other firms, enabling them, after tweaking a couple of dozen variables, to copy popular Japanese products 8 years later in the US—creativity “improved” several Harvard Business Review articles said, rather loudly. So its (behavioral research’s) intellectual sins have practical power, they make application of its results, well, silly. This does not even include the bigger problem of “effects” being published by everyone–journals and NYTimes—without boundary conditions or magnitudes. So we human-monkeys tend to in certain (unspecified) conditions exhibit, among tens of thousands of effects, ONE here-specified effect (entirely drowned by a few hundred dozen other much BIGGER and longer LASTING other effects not included in OUR study). Crap in crap out. It all starts with professors in behavioral sciences who have—entire generations of them—ABSOLUTELY no understanding of the statistical tools they commonly use. Generations of stat dummies at top ten colleges have “taught” terribly badly and wrongly statistics to generations of other dummies (now tenured full professors). The game works here just as the regulate-oil-off-coast game works till BP just as the bankers-in-government regular bankers-not-in-government game works on Wall Street—phony people in phony people and results out, this is how civilizations die, not with a whimper but with a bang.
richardtaborgreene - July 2, 2010 at 11:34 pm
Actually, statistics is a status marker (very very low status) so senior people never teach it (thank goodness—they have no clue as to what it is and does as any decent lunch conversation uncovers in its first 2 minutes). We instead find the least “with it” most impecunius, bad-english-speaking part-time substitute grad student assistant secretary, ad hoc, ad junct, ad lib to teach a stat course, freeing the big people from the grime of actually doing science. Pardon my french.
rodriguezfeo - July 3, 2010 at 10:13 am
Considering that there is ongoing non consensual experimentation here and abroad, the concept of the “untapped” billions is frightening to say the least.The problem with some of the research is that it’s lacking in empathy. Lack of empathy is what led generations of scientists to fool themselves into thinking that they understood humanity or that “studying” people can be broken down into components – biology, environmental controls, behavioral models etc. This “empathy” has been eliminated because the researchers often confused with something called bias. But the example in the article demonstrates that all researchers have biases.If those that took so long to discover differences in humans want to come up with overarching theories about humanity based on behavioral models, the real question has to be: for what purpose?
rodriguezfeo - July 3, 2010 at 10:16 am
*often confused it with
sgreerpitt - July 3, 2010 at 10:30 am
sociologists and anthropologists have always recognized this basic flaw in psychology.
agpbloom - July 3, 2010 at 1:51 pm
Henrich, Heine and Norenzayan say the following:”The authors of the paper suggest steering clear of universal proclamations and, whenever possible, broadening the pool of subjects.”But wait…is this not also a “universal proclamation”?Or is it just an across-the-board precription?Are the two one in the same?The interaction between the “universal” and the specifics of cultural context seems to be the interesting thing. Undue emphasis on either one seems to render the human sciences impotent…even if it does make for interesting partisan debates.
agpbloom - July 3, 2010 at 1:53 pm
across-the-board PRESCRIPTIONSorry about the misspelling.