(Return to index)

A clear discussion of this test and a list of the investigations carried on by means of it will be found in Whipple's "Manual of Mental and Physical Tests." For the present purpose a list of fifty words was used, as follows:
 

attractive	loud	dangerous	drunk	aristocratic
helpless	idle	rash	sacred	broad
prompt	savage	graceful	coarse	noisy
foreign	defective	vague	ancient	innocent
cowardly	public	masculine	hostile	gentle
peculiar	wise	ripe	fickle	gay
harmonious	shallow	slovenly	timid	soothing
beautiful	cloudy	talkative	sickly	stormy
fertile	tragic	doubtful	lazy	sharp
brief	expensive	stale	victorious	past

These words were typewritten on a card in two vertical columns. This card was laid face downward before the subject at each trial, and at a given signal from the operator the subject turned it over and went through the list as rapidly as possible, naming the opposite to each word in order. The opposites only were uttered, the words on the card not being pronounced. The time was taken by the operator with a stop watch to the fifth of a second. As in the color naming test, no errors were allowed. The words were arranged in 20 chance orders, and thus the order was never the same for successive trials. F2 was already well down on her practice curve at the beginning of this test, having previously gone over this list of opposites a great many times as operator.

This is a test of controlled association processes. Hollingworth [1] says of it, "The opposites test is a much used one in experimental and educational psychology, and has been found [p. 54] to correlate to a fairly high degree with other tests designed to measure mental ability. . . . This test indicates the ability of the individual to select the appropriate response from the host of ideas which follow in the wake of a stimulus word. It is an index of speed, accuracy, linguistic feeling, and of ability to repress useless or irrelevant ideas. . . . It is a test of association processes, but of association processes of a considerably more complex kind than those involved in the color naming test."

Woodworth and Wells [2] take the association test to be "A measure of mental alertness," and "of efficient mental control." It may also be added here that these investigators found the opposites test to be the best representative of controlled association tests.

Stern [3] also gives figures to show the high degree of correlation between the opposites test and other tests for efficiency. Since it correlates closely with other mental tests, the inference is fair that it should yield a representative measure of general mental efficiency at a given period.

The complete daily records of the seven subjects are given in Table XIV. The records converted into averages as under the tapping test (Table II) are given in Table XV.

Examining the contents of Table XV, we find that the records marked (*) form a consistent part of an ordinary practice curve. They show no uniform tendency to rise above the records preceding, and thus to indicate characteristic lapses from efficiency. It is to be noted, particularly after the first great drop in practice, that here, as in the other tests, every average does not improve over the average preceding it. This is just as true for control records as for records marked (*), and is indeed a well recognized feature of practice curves [4] without regard to sex.

To summarize the results of Table XV briefly and specifically, out of the fifteen critical averages herein recorded, eight are slightly better than, four are slightly worse than, and three are the same as the record preceding. This is about what would [p. 56] be found in the case of ally twelve averages taken at random from the table.

The mean variability of records made during critical periods reveals no uniform tendency. The M. V.'s for the opposites test are greater than those for the tapping test; yet they are low enough to demonstrate that the figures are quite reliable.

Table XVI was compiled in the same way as Table III, described under the tapping test. The purpose of Table XVI, as of Table III in the tapping test, is to compare the average performance of the critical period with a standard of efficiency (which is the average of the averages of periods immediately before and after); to examine into the average performance of four critical days, excluding the first; and to determine whether the first day shows a remarkable degree of impairment.

F1 is first considered in this table. On the first month, while the records were more variable than later, owing to the great amount of practice, the record of F1 for the critical period is 38.9* as compared with 37.0, the standard of efficiency for that particular period. The difference is, however, within the M. V. On the remaining two months the records of the critical periods equal the standards (30.1*-29.8, and 27.2*-27.0, respectively).

In the case of F2 the first two periods do not measure up to their standards (31.2*-29.5, and 27.7*-26.3, respectively). The difference between the first of these, however, is within the M. V. On the remaining two months the records of the critical periods equal the standards (24.5*-24.3, and 23.9*-23.7, respectively).

In the case of F3 the critical period is twice the same as the standard (32.7*-32.8, and 29.8*-29.8, respectively), and once slightly worse than the standard (39.9*-39.3), the difference being within the M. V.

The record of F4 shows the critical period to be twice slightly worse than the standard (33.7*-32·9, and 25.2*-24.5, respectively), the difference being both times within the M. V.

The record of F6 shows the critical period to be slightly better than the standard in one of the two cases where the record for the whole month was completed (54.8*-56.0), and in the other instance to be slightly worse than the standard (39.5*-38·6), these differences being well within the M. V.[p. 57]

It must be concluded that this experiment discovers no real influence, either adverse or favorable, on the average of performance for critical periods. An examination of the control records will show the same result that the critical records yield.

Turning now to the first days, it is seen in the case of F1 that all records for first days are very slightly worse than the averages in which they are included, as well as slightly worse than the averages that precede them. The first two months for F2 show first days considerably worse than the averages in which they are included, and worse than the averages preceding them. On the last two critical periods for F2 the first days equal the averages in which they are included. The record of F3 shows one first day worse than the record in which it is included, one the same as the record in which it is included, and one slightly better than the record in which it is included. F4 shows one first day worse and one first day better than the average in which it is included. F6 shows one first day better and two first days worse than the averages in which they are respectively included.

This general result would seem to indicate a degree of impairment on first days in the opposites test, nine first days out of the fifteen herein recorded being worse than the averages in which they are respectively included, as well as worse than the average which precedes. But a scrutiny of control records shows the same "worseness" on pseudo-first days, as witness for example the record of MI (control for F2) on the fourth month, or the record of F2 (control for F1) on the first month. This fact, combined with the smallness of the differences, renders the conclusion very doubtful that there is genuine impairment on first days. The only degree of "worseness" which is noticeable is in the case of the first two first days for F2, and it must be remembered that on these days F2 suffered pain for a short time. It seems not at all improbable that the opposites test was adversely affected by this suffering, especially since F2 shows no impairment whatever on the two remaining first days, when no pain was experienced.

When curves are platted from Tables XIV and XV, the menstrual periods are not characterized by inefficiency. The curves follow the downward direction of the ordinary practice curve, as they did in the case of the color naming. No period [p. 58] of maximum efficiency recurring regularly in each month and no "cycle" are revealed, even after the curves have attained a fair practice level. On the whole, the performances of F1, F2, F3, F4 and F6 follow the same course as those of M1 and M2, and are distinguished from the latter by no peculiarities.

As in the case of the color naming, it may be objected that this test, after many repetitions, became automatic, and that the processes of association, inhibition, etc., were no longer truly measured, although the varying order of the words did preclude the possibility of pure memory work. Therefore five trials were made with each subject when the experiment was entirely finished, as follows. The opposites to the words on the card were typewritten in exactly the same way as the words themselves, in two vertical columns. The subject then read the opposites, turning the card over, and performing exactly as he or she had done previously. The records for these trials are given below. They show how far the process was from being automatic, even after more than a hundred and twenty-five trials. The subjects could articulate the opposites when placed before them, about twice as rapidly as they could call them up by association and articulate them.[p. 59]

[1] Hollingworth, op. cit., p. 18.

[2] Woodworth and Wells. op. cit., p. 5, 59.

[3] Stern, Die Differentielle Psychologie, 1911, p. 488.

[4] E. L. Thorndike, op. cit., pp. 166-169.