"Humans"? This was one study of students playing a particular game in a modern American university. The jump to "(all) humans" seems unjustified. But suppose the researcher might have wanted to show the effect of cockiness.
Heh. Well, I agree that it's premature to generalize definitively to humans in general, and this is a problem with psychology experiments in general: any particular experiment is necessarily very restricted in scope, and generally N is very small, so it's hard to have confidence in anything more than simple linear models.
Occam's razor suggests that it's more likely, given this one data point, that it's true of all people than that it's only true of the people that happened to be in the sample. But that clearly doesn't justify holding that view very strongly; it's based on a small amount of evidence.
