I had the honor to work with Rich Gonzalez and with Robert Savit . We worked on an experiment called The Minority Game. Our research team studied strategy formation and social cognition.
I ran studies of up to 17 participants and participated in lab meetings. While there we submitted a paper for publication, however, it was not accepted.
Here is more information about The Minority Game as explained in our paper -
The Minority Game: General Structure, Computer Simulations, and Analytic Results
Consider a repeated game of N (odd) players. At each time step of the game each player chooses to join group 1 or group 2. Each player in the smaller group at that time step gets one point, while each player in the larger group gets nothing. Thus players are rewarded for making choices that are different than those made by most other players. The game is repeated for T time steps. Players are not privy to the choices of other players and do not communicate directly with them. The only information available to players is the list of which groups were minority groups for all previous time steps of the game.
The Minority Game: The Human Experiments
The results of the computer simulations are quite intriguing. But it is unclear how relevant they are to groups of humans. To address this question, we performed a set of experiments with humans playing the Minority Game.
Description of the Games:
An odd number of participants played the minority game for 400 time steps. At each time step participants had five seconds to join either group 1 or group 2. Participants in the minority group were monetarily rewarded, while participants in the majority group were not. Participants were given only the history of minority groups, updated after each time step, and a running total of their winnings. They were not allowed to communicate with each other.
After the game, each participant was asked to complete a questionnaire. These questionnaires indicated a high degree of heterogeneity in the players self-reported heuristics and strategies. This heterogeneity of the agents is central to understanding the rich and counter-intuitive behavior of these games.
Most of the games discussed here were as described above and were endogenous, in the sense that the history of minority groups that the players used to make their decisions were the ones actually produced by the choices of the players in that game. We also performed exogenous experiments. In these games participants were told that they were to play the minority game as described above, but rather than updating the history of minority groups with ones actually produced by the participants choices, the participants were provided with a list of minority groups produced by a previous game (with the same number of participants). Thus, their decisions had no effect on the next minority group to appear in the list. Participants were rewarded if their choice of group matched the one that appeared next in the list.
The results reported in this paper are based on a total of 29 endogenous games whose participants, N, number from 5 to 23. These games consisted of fourteen 5-person, two 7-person, five 11-person, three 15-person, two 17-person, one 19-person, and one 23-person games. In addition, we performed 10 5-person exogenous games.
Summary: When humans play a repeated endogenous game in which they are rewarded for being in the smaller of two groups, those players that adopt the simplest strategies generally reap the greatest rewards.