From Security Analysis, 1940 edition:
To sum up this discussion of qualitative and quantitative factors, we may express the dictum that the analyst’s conclusions must always rest upon the figures and upon established tests and standards. These figures alone are not sufficient; they may be completely vitiated by qualitative considerations of an opposite import. A security may make a satisfactory statistical showing, but doubt as to the future or distrust of the management may properly impel its rejection. Again, the analyst is likely to attach prime importance to the qualitative element of stability, because its presence means that conclusions based on past results are not so likely to be upset by unexpected developments. It is also true that he will be far more confident in his selection of an issue if he can buttress an adequate quantitative exhibit with unusually favorable qualitative factors.
But whenever the commitment depends to a substantial degree upon these qualitative factors—whenever, that is, the price is considerably higher than the figures alone would justify—then the analytical basis of approval is lacking. In the mathematical phrase, a satisfactory statistical exhibit is a necessary though by no means a sufficient condition for a favorable decision by the analyst.