Page:Journal of the Optical Society of America, volume 30, number 12.pdf/56

 All subjects were similar in that they had some interest in color and passed the Ishihara test for color blindness, but they differed widely in respect to the nature of their color interest and the character of their special training. In general they may be classified as follows: professional colorists in science or industry, 10; amateur or apprentice colorists, 8; professional physicists specialized or interested in color, 5; professional psychologists specially interested in color, 5; psychology students, 5; business executives with a broad or commercial interest in color, 3; specialists in dyes and pigments, 2; specialists in physiological optics, 2; graphic arts research, 1. There is much overlapping and this classification, like any other, is rather arbitrary.

The task of making visual estimates of colors and color differences is obviously a psychological one and the fact that one of our psychological observers, F. L. Dimmick, was opposed to the procedure as too demanding is duly noted. It should be noted, further, that the psychologists as a group were inevitably aware of numerous sources of error, technical imperfections, and general observational difficulty in the procedure. The various specific difficulties which seemed of any essential importance to the writer have been revealed in the preceding section on “complications.” Most of them appeared to bother the nonpsychological observers relatively little or not at all. But one of these, D. L.MacAdam, also objected on psychological grounds. The subcommittee is grateful to both of these subjects for carrying through their series of observations in spite of their objections.

The advantages of having such a diverse group of subjects are; first, that participation by a wide representation of the color-interested population yields correspondingly representative results and, second, tends toward wide application of results. A disadvantage of the fact that the data represent a heterogeneous sample of color-interested individuals is that they may not prove wholly satisfactory to any particular individual in the group. Lack of psychological training in scientific observation will tend to make the group results incorrect for the specialized sense-psychologist whose judgments in general might be assigned a higher validity than those of anyone else. Other observers would be more susceptible to the various recognized sources of error indicated in the section on “complications. Some of the sources of error are not even recognized by most of the untrained observers. These errors undoubtedly affect the results in ways not intended by the observers, but, on the other hand, they show various systematic tendencies which should correspond to the way the Munsell samples look to them. An example of the latter is the tendency to see chromatic colors as lighter than does the more experienced observer. As already indicated, there is something to be said for the view that what is really wanted here is a system of color which will seem right to the bulk of the population which will use it. Probably this population would never contain more than a minority who were specially trained in technical color observation. A desirable later step, perhaps, would be to produce a surface-color solid which would be as correct as the most rigorous technical training could make it. In the meantime, however, we shall hope to make some comparisons between results secured from sense-psychologists and conscientious colorists; possibly the differences in their results will not prove to be intolerably large after all.

Averaging the estimates

Representative visual determinations were computed for the hue, value, and chroma of each of the standard Munsell colors as perceived against each of the three experimental grounds. Each of these determinations is an average of the individual estimates of the participating observers. The averaging procedure included two principal steps: (a) summarization of all of the available data, and (b) computation from the summary of the arithmetical means of hue, value, and chroma. A detailed description of these two principal steps is given below.

(a) The original vectorial records entered by the observers in the data sheets, of the form already illustrated in, were summarized on sheets of the form illustrated in. In this particular form, which was used in conjunction with the constant-value charts, the hori