If a strict equilibrium is suggested as the solution of a strategic game in normal form and if some but not all players are committed to their solution strategy, a reduced game results with only the still non-committed as active players. The reduced game property (or consistency axiom) demands that the solution of the reduced game is given by the original solution strategies of its active players. However, postulating the reduced game property is asking for too much: consistent equilibrium selection in general is not possible if certain other requirements (existence and optimality) are granted (Norde et al., 1996). Does the reduced game property have at least some behavioral appeal? We test this experimentally by confronting players with a solution proposal before letting them decide both for the original game and for its reduced games.
