An internet resource developed by
Christopher D. Green
York University, Toronto, Ontario
(Return to index)
[p. 42] VI
SPEED AND ACCURACY OF PERCEPTION
THE COLOR NAMING TEST
The color naming test was conducted as follows: The card designed by Woodworth and Wells  was placed face downward at each trial, before the subject: At a signal from the operator the subject turned the card over, and began naming the colors as rapidly as possible from left to right. No errors were allowed. If the subject made a mistake the operator, who held an identical card, quickly said "No," and the subject corrected the error before proceeding. Thus mistakes and confusions were accounted for in the time consumed by them. The time of performance was taken by the operator with a stop watch to the fifth of a second. The same order of the colors was used throughout after the first month, in the case of M1, F1 and F2, since it was thought that some orders always yielded poorer records. In the case of all the other subjects the four orders of the colors were used in rotation, from trial to trial, throughout the experiment. The illumination was the same for all trials.
This test has been used by various investigators. Hollingworth  found that it was sensitive to so small an influence as 1 to 2 grains of caffein alkaloid. To quote from this author, the test "is designed to measure the speed with which the name or idea can be brought to consciousness upon the sight of the object, which is in this case a color. . . . As the experiment progresses it affords a measure of the individual's ability to improve by practice, his degree of interference as shown by the tendency of a preceding idea to inhibit or interfere with the correct perception and expression of the next stimulus."
Table XI gives the complete daily records of all seven subjects [p. 43] and requires no comment. Table XII gives the result of converting these records into averages of five-day periods each, as described for Table II under the tapping test.
In Table XII it is seen that the averages marked (*) form a consistent part of an ordinary practice curve. They show no uniform tendency to rise above records preceding them, and thus to indicate inefficient performance. It will be noted, particularly after the first great drop in practice, that every average does not improve over the average preceding it. This is just as true for control records as for records marked (*), and is a familiar feature of practice curves without regard to sex.
Table XII shows that out of the fifteen critical averages herein noted, nine are slightly better than, four are slightly worse than, and two are the same as the average preceding. This is about what would be found in the case of any twelve averages taken at random in the table.
The mean variability of records made during critical periods is neither uniformly greater nor uniformly less than that of other records. The M. V.'s in the color naming test are greater than in the tapping test, but they are small enough to demonstrate the reliability of the figures.
Table XIII was compiled in the same way as Table III in the tapping test. The purpose of Table XIII was the same as that of Table III, i. e., to compare the average of the critical period with a standard of efficiency (which is the average of the averages of periods immediately before and after it); to examine the average performance of four critical days, excluding the first; and to determine whether the first day shows a remarkable impairment.
F1 is first considered in this table. The critical period is once worse than the standard (33.8*-31.6), and twice slightly better than the standard (36.7*-38.2, and 27.5*-28.1, respectively).
In the case of F2 the critical average is twice slightly worse than the standard of efficiency (41.9*-40.7, and 28.6*-28.1, respectively), the difference being both times within the M. V., and twice slightly better than the standard (44.4*-44.9, and 34.3*-31.7, respectively), this difference also being well within the M. V.
In the case of F3 the critical average is once the same as,[p. 44] once slightly worse than, and once slightly better than the standard of efficiency (40.3*-40.4, 37.1*-36.6, and 33.2*-34.1, respectively), these differences all being within the M. V.
F4 shows both critical averages included in her record slightly worse than the standard, the difference being well within the M. V. (43.7*-43.2, and 40.8*-40.2, respectively); and F6 shows practically no deviation in either direction from the standard of efficiency at critical periods.
It must be concluded that this experiment shows no real influence, either adverse or favorable, on the average of performance for critical periods. An examination of the control records will give the same result that is yielded by the critical records.
Table XIII demonstrates that first days of critical periods bear about the same relation to the averages in which they are included and to the averages preceding as is to be found in the case of the pseudo-first days of the control.
When curves are platted from Tables XI and XII, the menstrual periods do not rise in " blocks " of inefficient performance. In only one case out of twelve is the critical " block " high between the "blocks" preceding and following, i. e., the second period in the case of F1. This one instance is typical of what we should expect to find in all cases, or at least in a great majority of cases, if there were marked impairment during menstrual periods.
On the whole, the performances of F1, F2, F3, F4 and F6 follow the uninterrupted course of the familiar practice curve. They are distinguished from the performances of M1 and M2 by no peculiarities.
It may be objected that this test after scores of repetitions became automatic, and that the processes of perception, inhibition, association, etc., were no longer truly measured; or even, since the same order of the colors was used in some cases, that these subjects must have memorized the colors in order. Therefore five trials were made with each subject after the experiment was entirely finished, as follows. The names of the colors were typewritten on a piece of white paper from left to right, so that reading them would involve the same number of eye movements as the naming had previously involved. The subject was then required to read the colors in the same way as he or she had [p. 45] previously named them. This performance, though by no means purely automatic, was undoubtedly much more nearly so than the naming of the colors. The time in seconds of each of the six individuals for this test is given below. The records show how far the color naming was from being a memory feat or a purely automatic performance, even after a hundred and twenty-seven trials and more.
 Woodworth and Wells, Association Tests, Psych. Mon., No. 57.
 H. L. Hollingworth, op. cit., p. 16.
 E. L. Thorndike, op. cit., p. 168-169.