Measuring beliefs in an experimental lost wallet game

We measure beliefs in an experimental game. Player 1 may take x < 20 Dutch guilders, or leave it and let player 2 split 20 guilders between the players. We find that the higher is x (our treatment variable), the more likely is player 1 to take the x. Out of those who leave the x, many expect to get back less than x. There is no positive correlation between x and the amount y that 2 allocates to 1. However, there is positive correlation between y and 2's expectation of 1's expectation of y. Journal of Economic Literature Classification Numbers: C72, C92. (C) 2000 Academic Press.