The theory of democracy requires that the outcome chosen by majority rule from among the set of available alternatives should reflect the preferences of the voters in some meaningful way. When an outcome is said to be a "majority rule winner", no majority coalition should be able to imagine a different outcome that all coalition members would prefer. Procedural rules should simply be neutral conveniences enabling a voting body to determine the majority rule winner efficiently and easily. Unfortunately, none of these desirable propositions is true in general for majority rule voting. When voters have strong differences of opinion, then there may be no single alternative that can be chosen that reflects the "will of the people" in any meaningful way. For any possible selection, there will always be some majority coalition that would prefer a different choice. When this is true, different procedural rules may lead to quite different outcomes simply by protecting favored outcomes from a majority coalition that would prefer a different choice. It is troubling to realize that any possible democratic procedure could determine the outcome in a way that is in some sense arbitrary and "independent" of the preferences of the voters. But if so, it is important to understand the effects of alternative procedural rules on democratic decisions. The proposed research conducts controlled laboratory experiments in which subjects are paid amounts that vary with the alternative selected from a two-dimensional grid. Subjects have quite different preferences over the alternatives to be selected, but the preferences are held constant from one voting body to another. But different voting bodies are allowed to vote over the alternatives using quite different procedural rules. Different rules determine, for example, the procedure by which alternatives are introduced for consideration and balloting. If rules are simply neutral means by which voting bodies conveniently arrive at invariant outcomes, then the outcomes chosen should not differ from one experiment to another. However, early experiments indicate clearly that different rules lead systematically to quite different outcomes, even though the preferences of the voters are held constant. Different theories make quite different predicitions about the outcomes to be selected under different rules; with sufficient experimentation, it should be possible greatly to increase our understanding of the effects of the rules of the game on democratic choice.
