A substantial number of attempts to test the traditional structureperformance hypothesis of classical industrial organization economics have used weighted average concentration ratios as one of several independent variables.1 In a recent issue of this journal, Stanley Boyle (1973) claims to demonstrate that "the statistical weighting methods which have been employed by those developing these average concentration measures cannot and have not yielded the type of concentration estimates which were sought." The purpose of this comment is to show that (1) Boyle's main hypothesis is left unexamined since he does not correctly define and confront the "type of concentration estimates which were sought; (2) the alleged superiority of an alternative "three-digit" concentration measure based directly on statistics-of-income data is not supported by a properly specified empirical test; and (3) aggregation does have important implications for conclusions based on profit-concentration regressions which are ignored by Boyle. Weighted average concentration ratios were introduced as a partial index of industry structure in several research efforts of the 1960s.2 The main reason for developing and using weighted average concentration ratios is to facilitate analyses for which other relevant variables (e.g., income-statement data) are available only for broader industry classifications than the Standard Industrial Classification (SIC) four-digit classifications-the most common source of concentration measures. However, the goal of the weighting procedure is not, as Boyle suggests, to create a concentration measure for the broader industry class that