[Excerpt] This writer can't see the potential in the establishment survey data that Mills sees —at least, not for collective bargaining purposes. First, the sample can never be made large enough, except at prohibitive cost, to include a sufficient cross-section of unionized firms. Granted, union and management representatives have some interest in what is happening in nonunion firms; but this writer would guess their principal interest is in what is happening in comparable unionized relationships. Second, the establishment survey is a good source of information on average hourly earnings and the like, but it is hard to believe it provides information on wage rates and scales. Union and management officials are probably more interested in negotiated wage rates than in actual earnings data. Third, as far as this writer knows, the Establishment Survey provides absolutely no information on the multitude of other matters that typically concern the parties: vacation and holiday schedules, leave provisions, seniority and job security measures, etc. These are additional reasons why this writer believes expansion of the BLS collective bargaining contract series would have a bigger payoff to the parties and to researchers than the measures Professor Mills recommends.