A case for standard theory?

Using data from an experiment by Forsythe, Myerson, Rietz, and Weber (1993), designed for a different purpose, we test the "standard theory" that players have preferences only over their own mentary payoffs and that play will be in (evolutionary stable) equilibrium. In the experiment each subject is recurrently (24 times) randomly matched with ever changing opponents to play a 14 player game. We find that assuming risk-neutrality for all players leads to a predicted evolutionary stable equilibrium that, while it can be rejected at the 5% level of significance, is nevertheless remarkably close to "explaining" the data. Moreover, when we assume that players are risk-averse and we calibrate their risk-aversion in one treatment with a simple game, this theory cannot be rejected at the 5% level of significance for another treatment with a more complicated game, despite the fact that we have close to 400 data points.