The results of the ultimatum game do vary a great deal between cultures.

Yup, there’s some variance among people of different cultures.

Just to clarify, my point is that experimental results show people acting in a manner inconsistent with what a simple understanding of self-interest would predict. That fact remains true across cultures – albeit to differing degrees.

I'm skeptical of evolutionary psychology as a scientific endeavor, because the basic theory can explain anything in retrospect--but I've never seen a researcher make a prediction based on EP and then verify it via testing.

Funny you would say that....

Where cultures differ environments differ: a hunter-gatherer tribe may be so closely interdependent that their optimal strategies in the ultimatum game would differ than those of people in more individualistic-independent environments. In a tribal setting, direct competition for resources or sexual partners might be more intense: in such a setting, if you're directly competing with the person proposing the split, it makes no sense to accept less than 50/50....

I like it! That seems like a sound basis for predicting that people in small, close-knit communities would demand greater equity than people in larger, more atomistic societies.

But the last time I looked at experimental results, I saw the exact opposite outcome. That is, people from cosmopolitan, market-based economies offered and demanded GREATER equity; people from more traditional, isolated societies offered and demanded less.

The authors hypothesized that people in market economies were accustomed to interacting with strangers. They identified with BEING a stranger, and therefore identified with the need to accord equity to strangers. And they recognized that building a reputation of rejecting inequitable treatment might prove expensive in the short term but pay off during future negotiations – either with this specific individual, or with others who had learned of your reputation.

In contrast, people in isolated societies were accustomed to frequent, repeated dealing with kin, punctuated by rare interactions with strangers. In such societies there is much less value placed on equity. When dealing with kinsmen, the terms of any one transaction are not so important given the high degree of sharing – both in an economic and gene-pool sense. But the calculus for dealing with strangers is completely different. Because they did not expect the interaction would be repeated, the benefits of establishing a reputation for demanding equity were unlikely to compensate for the short-run costs. And acts of “altruism” would be less likely to result in benefits to your gene pool than acts of selfishness.

Having now set this contrast sharply into focus, I must sheepishly acknowledge that I haven't found a citation to the studies I have in mind. Sorry!