In the classic paradigm, people are sorted into two categories on a trivial basis, never meet anyone, and allocate points between one anonymous member of each category. Nobody can allocate anything to themselves.
Every option states exactly what it gives to each person. Nothing is hidden, nothing is timed, and you can allocate nothing to yourself.
Progress
| Starting position | Fair | Joint total | Own group most | Largest gap | Gap afterwards |
|---|
Five conclusions somebody might draw from a result like yours. Say whether each is supported, partly supported, or not supported by this kind of evidence.
If people simply wanted their own category to do well, they would take the option that gives it most. The reason the paradigm mattered is that a reliable minority does not: they take fewer points for their own side in order to be further ahead of the other. Nothing in self-interest explains that, because nobody gains anything - the allocator cannot allocate to themselves, the points buy nothing, and no one will ever know who decided what.
Social identity theory's account is that a category one belongs to becomes part of how one is placed in the world, and that placement is comparative: what a group is, is partly what it is rather than. Being ahead is therefore worth something that simply having more is not. That is a claim about the comparative structure of identity rather than about hostility, and the paradigm is evidence for it precisely because there is nothing else in the situation for the behaviour to be about.
In the original studies and almost certainly in your own profile, fairness is the single most-used strategy. The finding was never that people are biased rather than fair; it was that a pull towards differentiation shows up reliably even when categorisation is the only thing present. A demonstration in which everyone allocated unfairly would be a different and much less interesting result.
Choosing the largest gap when both categories start level creates an inequality. Choosing it when your category already holds four times as much maintains and widens one. The point-scoring is identical and the act is not, and no measure of allocation strategy will tell the two apart. That is the limit of the paradigm, not a criticism of it: it was designed to establish a minimum condition, and a minimum condition is not an explanation of anything that happens under maximum conditions.
Real intergroup relations involve categories with histories, material interests, unequal access to institutions, and memories of what has been done. A Kestrel has none of these. The honest summary is narrow and still worth having: mere categorisation is sufficient to produce differential allocation, and therefore no account of discrimination that requires prior hostility, competition or contact can be complete. Sufficient for a laboratory effect is a long way from adequate to a society.