The growth-value and small-large scales are popular, but they suffer from two deficiencies: They are not grounded in theory, and they clear definition. Shefrin and Statman offered a remedy in a theory in which the quality scale-the scale that separates companies from bad-plays a central role. Here, the ratings in the Fortune survey of company quality are used as direct measures of perceptions of company quality and a proxy for the Fortune quality ratings is constructed from the BARRA, Inc., list of company characteristics. This quality scale has applications in style selection and style rotation. Categorizing companies along the growthvalue and small-large scales is now a common practice. Both scales are imperfect measures of quality, the scale that separates companies perceived as good from those perceived as bad. Good companies have such characteristics as management and products and services. The growth-value and small-large scales have gained widespread use because empirical evidence, such as that presented by Fama and French, shows significant differentials in stock returns along these scales.1 The growth-value and small-large scales have two deficiencies, however, one related to of theoretical foundation and one related to ambiguity in measurement and application. Black noted that no theory underlies the return differentials along the growth-value and small-large scales.2 Discussing the Fama and French findings, Black wrote lack of theory is a tipoff: watch out for data mining (p. 9). As to the measurement issue, no general agreement exists about the characteristics that distinguish from value. Some analysts use book-price ratios, others use earnings-price ratios, and still others use dividend yield. Moreover, the terms growth and often are robbed entirely of meaning. Bogle reported that, as of the end of 1992, Philip Morris Companies was the largest holding of both and value mutual funds.3 As to size, Chan and Chen point out that the distinction that matters is not the distinction between small and large companies but the distinction between financially distressed and financially healthy companies.4 Moreover, although size is most often measured as market value of equity, measures such as book value of assets are also used. Shefrin and Statman offered a remedy for the theoretical deficiency of the growth-value and small-large scales within a behavioral capital asset pricing model.5 Asset prices in the behavioral capital asset pricing theory are the outcome of an interaction between two kinds of traders-information traders and noise traders. Information traders know the relationship between characteristics of companies and the return distributions of the stock of these companies. In contrast, noise traders make systematic errors as they assess the relationship between characteristics of companies and the return distributions of the stocks of these companies. The error that is relevant here is representativeness, a cognitive heuristic described by Tversky and Kahneman.6 The error manifests itself in the belief that stocks are stocks of companies. Statman and Shefrin provide direct evidence that, on average, investors indeed believe that stocks are stocks of compa7 nies.