[Excerpt] A major development in systems for the enforcement of individual employment rights is the use of alternative dispute resolution (ADR) procedures to resolve claims by employees. At their best, ADR procedures may hold the potential for greater accessibility by employees to enforcement of substantive employment rights, while avoiding burdens of excessive costs for the public and employers in processing claims. On the other hand, ADR procedures, particularly mandatory employment arbitration procedures, have also been criticized for producing the privatization of justice and denial of effective enforcement of employee rights. In this paper, we present the results of a new empirical study of employment arbitration. Despite the growing importance of employment arbitration in the workplace, empirical research on this phenomenon remains in its infancy and views on arbitration are often characterized by assumptions and anecdotal impressions. In the analysis presented here we attempt to systematically examine some of the common assumptions about the decision-making of employment arbitrators. In particular, we examine three propositions that are often injected into discussions of arbitral decision-making: 1) Arbitrators will tend to favor compromise decisions, proverbially “splitting the baby” between the two parties. 2) Arbitrators will be less inclined to award very large damage claims of the type more sometimes seen in jury decisions. 3) Arbitrators will prefer to award at least some small, token amount of damages to a party bringing a case rather than deny any recovery. We analyze these propositions using a unique dataset developed from analysis of employment arbitration case files of the American Arbitration Association (AAA), arguably the leading provider of employment arbitration services in the country.