糖心原创

Game-theoretical and monetary payoffs in laboratory Experiments

Abstract:

Experimental research in economics relies on performance-dependent monetary incentives to implement game-theoretical games in the laboratory. While the set of players and the set of strategies are unambiguously defined and controlled by the experimenter, game-theoretical payoffs are typically either assumed to coincide with monetary payoffs, or are estimated ex-post based on observed actions and some sort of hypothesized rational behavior. We follow a different path and discuss results from an experiment on simple 2-person games in which participants were repeatedly asked to report their expectations on the opponent’s behavior and their own level of satisfaction for each possible outcome of the game. This approach allows us to reflect on experimental methodology by directly comparing monetary incentives with (perceived and declared) game-theoretical payoffs. We find that although repetition and experience help, they are unable to completely align game-theoretical payoffs with monetary incentives. While most participants seek to maximize money earnings in the experimental laboratory and they tend to perceive the games in the intended way, a small - yet non-negligible - fraction of the subject pool consistently interprets the game and, therefore, acts in an unexpected way.

Keywords: experimental method, induced-value method, prisoners’ dilemma, rationality