In a national sample of 302 police interest arbitration awards, arbitrators issued 1,5 7 4 salary decisions and 2,684 nonsalary decisions which researchers were able to evaluate. Using a definition of "winning" that is deliberately designed to record as many union wins as possible, researchers found that 47 percent of the decisions were for unions and 40 percent were for employers (the remainder could not be classified with the information available).
Arbitrators were more likely to favor unions in their salary decisions and employers in nonsalary issues. Most arbitrators provided explanations of their decisions, citing comparability and ability-topay evidence most frequently. Arbitrator decisions and criteria citations were consistent in that most references to salary-related evidence were expressed in language favoring unions and most references to nonsalary-related evidence favored employers.