Planning ahead: eliciting intentions and beliefsin a public goods game

In a two-person finitely repeated public goods experiment, we use intentionsdata to interpret individual behavior. Based on a random-utilitymodel specification, we develop a relationship between a player's beliefsabout others' behavior and his contributions' plans, and use this relationshipto identify the player's most likely preference \type". Our estimationanalysis indicates that players are heterogeneous in their preferences alsoat the intentional level. Moreover, our data show that deviations fromintended actions are positively related to changes in beliefs, thereby suggestingthat people are able to plan.