Recent empirical studies on outcomes in employment discrimination litigation all reach the same conclusion: Plaintifs have little chance of success. But these studies rely on summary judgment decisions and trial verdicts, gleaned from reported opinions, electronic docket entries, and data collected by the Administrative Office of the Courts, and they acknowledge that this is just "the tip of the iceberg." Until now, settlement outcomes, which account for 70% of case resolutions, have been rendered invisible because of confidential settlement agreements. Along with the "vanishing trial" syndrome, secret settlements have created an information vacuum, skewing the public policy discourse about employment discrimination litigation and leaving the judiciary, litigants, and attorneys without benchmarks by which to negotiate resolutions. This Article reports on an anonymously coded dataset of 1,170 cases settled by federal magistrate judges in Chicago over a six-year period ending in 2005, and provides the first detailed analysis of settlement outcomes. These data reveals that employment discrimination claimants, who account for over 40% of the dataset, obtain a mean recovery of $54,651. Among the factors analyzed as they relate to settlement amount are: type of discrimination, type of adverse employment action, stage of litigation, and amount of lost wages. These data indicates that employment discrimination litigation is neither jeopardizing American business nor resulting in undeserved windfalls for disgruntled employees. Plaintiffs are achieving a reasonable degree of success through settlements, measured against their lost wages. The Chicago project should be replicated and refined to provide an even more complete picture of settlement outcomes.