The study compares compulsory arbitration with final-offer arbitration. Research utilized separate analyses of the quality of arbitral settlements, the bias of negotiated settlements, and the probability of a negotiated settlement. The results suggest strongly that final-offer arbitration statutes intend one thing, while actual arbitration brings about something quite different. Ultimately, this kind of arbitration leads to what the arbitrator wants, completely independent of the parties' preferences except as they affect the arbitrator's preferences. Compulsory arbitration does not yield to the same incentives to go along with the arbitrator's wishes. An examination of modeling and theoretical considerations concludes that to design better arbitration schemes, impasses must be avoided as often as possible. One note and 12 references are included.