We Aren't the World: Why Americans Make Bad Study Subjects

The stakes Henrich used in the game with the Machiguenga were not insubstantial?roughly equivalent to the few days? wages they sometimes earned from episodic work with logging or oil companies.

Henrich's approach to the ultimatum game seems flawed. He mentions that he offers the equivalent of a few days wages, which is probably too much. The game is usually played for significant, but smaller sums, such as the value of a free lunch. For a sufficiently large starting sum, even tiny portions are enough to be worth something. For example, if you were asked to decide on a split of $200 out of a total $2000, you would probably want to spite the splitter. But you would also probably be overruled by your desire to get a free $200. It's only when we start looking at a smaller total with similar proportions -- say, $2 out of $20 -- that we start to see small portions being worth sacrificing to spite the other guy.

Proportionality is a bad metric in this scenario, and he should probably use some thing like "hours of equivalent labor" instead. (And in that case, he better hope everyone is used to making equal amounts of money in such an hour, which is certainly not true in Western societies.) By sticking to proportionality as a metric long after it becomes meaningless, Henrich buries the signal in noise. He has made it too easy for the splitter to "buy off" the decider.

The Pacific Standard description of the game also misses the point when they say that (for Western subject) the game tends towards and average 50/50 split. The average isn't nearly as interesting as the highest refused split/lowest accepted split, which tells you exactly how much someone is willing to sacrifice to spite the other party/the minimum "fair" proportion. This figure tends to be down near 30%. (It is up for debate how the subjects are internalizing this number as fair... whether it is closer to, say, "half of an even share (25%)," or "half of what the splitter makes (33.33%)," or some other figure.)

He is correct in that it will be culturally influenced. That is a big part of the point. In fact, when the experiment was originally devised, it was considered surprising that people would refuse any split at all. It is, after all, free money split between anonymous parties in exchange for no work at all. The reason people behave in this "illogical" manner is because reputation has worth, and if you want to avoid being cheated in society, it pays to have a reputation for being spiteful and willing to take a small loss to inflict punishment on those who wrong you. No transaction happens in a vacuum. The point is that the social gaming conditioning "leaks through" into our behavior even though the experimenter has (usually) done his best to remove all social components that would reward such spiteful behavior.

Now, Henrich has spent a few years doing this sort of thing, and it's been looked over by plenty of competent people, so I'm presuming his team's understanding is really not so shallow as it is presented here. But still, it is a bit odd to look at this collection of anecdotes that seems to demonstrate "culture matters" and come away with the conclusion that Westerners, and especially Americans, are weird. This is especially true when so many experiments of the previous century were aimed at identifying cultural behaviors and disentangling them from basic human response... in essence, all experiments which prove both that humans are similar (because they respond similarly under highly controlled conditions) and that culture matters (because that what influences them to behave slightly differently under different conditions). An experimenter has to be keenly aware of the culture under test, because experiments can amplify subtle differences if it doesn't account for them.

Source: http://rss.slashdot.org/~r/Slashdot/slashdotScience/~3/O0cfBlb5CNA/story01.htm

rod blagojevich uconn ncaa march madness mario williams vcu unlv sam young