Editorial Comment Upon the Effect of an Audience
Floyd Henry Allport
There are a few points of interest in a comparison of the results of the foregoing study with other investigations. With the kind permission of Dr. Gates these suggestions are added herewith.
It has been pointed out above that the originally superior subjects are on the whole those most unfavorably affected by the audience. If we compare the results of the highest and lowest subjects in the "audience" group, not with equivalent classes in the control group, but with each other, the differences dependent upon original ability stand out vividly. The following table presents this comparison from the data given.
| INFLUENCE OF THE AUDIENCE WITH REFERENCE TO INDIVIDUAL DIFFERENCE OF ABILITY | ||||
| Test | Average Improvement in Terms of Test Scoring Unit | Percent of Subjects who Improved | ||
| Highest 8 subjects | Lowest 8 subjects | Highest 8 subjects | Lowest 8 subjects | |
| Coordination | 1.5 | 5.5 | 63 | 88 |
| Color-Naming | 1.9 | 2.9 | 75 | 88 |
| Analogies | -5.4 | 12.5 | 25 | 100 |
| Word-Naming | -2.6 | 1.4 | 37 | 63 |
(343)
It will be seen that in every case, both in average improvement and per cent who improved, the lowest eight subjects gained more than the highest through the influence of the audience. The difference is striking in the per cent of subjects showing improvement in the tests of analogies and word-naming. In two cases the average of the highest group was actually lowered in the period under observation by the audience. It is true that the lowest subjects of the control group tend to show the same superiority in improvement after the rest period. This tendency, however, is not nearly so marked as in the other case. In "average improvement" it occurs in only three of the four tests, while in "per cent who improved" it occurs in only two tests. No doubt these results in the control group may be explained by the fact that the poorer subjects were those who were at first unfavorably affected by working in a group; but who after a period of habituation were spurred to greater effort by the sights and sounds of those working about them.
But notwithstanding this situation in the control group, the quantitative difference between the improvement of the lowest as compared with the highest subjects in the control group and the corresponding improvement in the audience group seems to indicate that in the latter the audience itself affects the inferior individuals more favorably than it does the superior.
Other investigators, such as Mayer, Moede, and the present writer have found a frequent inverse correlation between original ability and tendency to improve under the influence, not of an audience, but of the "co-working" group itself. That is, working together tends to benefit the work score of the slower reactors; more than it does that of the quicker reactors, as measured in comparison with work done alone. This is particularly true when the setting involves competition, a condition no doubt true to some extent of all co-active performance. I have elsewhere interpreted this phenomenon as due largely to differences of rapidity of stimulations from the work of the quicker and slower workers, respectively, combined with the rivalry incentive. We have from Dr. Gates' study the interesting suggestion that the presence of inactive onlookers also may spur the slowest workers, while it ha little influence (and sometimes a deleterious one) upon the speed of the more rapid. It may be that the audience renders the slower workers more conscious of their slowness, and increases their incentive to rival, and perhaps excel, their fellow workers.
It should be borne in mind that working in a group has been repeatedly found to stimulate an increase in quantity of indi-
(344) -vidual output. No doubt such co-working in the fairly large groups of Dr. Gates' experiment had already raised the level of effort to a point much nearer the physiological maximum than would have been the case if the individuals had been tested alone. Under such circumstances the audience may have had the effect in some cases of over-stimulating and impairing muscular coördination and mental work (cf. Triplett's findings upon "over-rivalry" in children). In other words we have here the resultant of two social influences, fellow-workers and an observing audience, rather than the isolated effect of the latter condition. It might be worth while to repeat Dr. Gates' method using solitary individuals, in one set of trials working wholly unobserved, and in another set working before an audience.
For references bearing upon the above suggestions consult the following:
Mayer, A., "Ueber Einzel und Gesamtleistung des Schulkindes," Archiv für die Gesamte Psychologie, 1903, I, 276-416. Moede, W., "Der Wetteifer, Seine Struktur and sein Ausmass," Zeitschrift für Pädagogische Psychologie, 1914, XV, 353-68. Triplett, N., "The Dynamogenic Factors in Pace-making and Competition," American Journal of Psychology, 1897, IX, 507-32. Allport, F. H., "The Influence of the Group upon Association and Thought," Journal of Experimental Psychology, 1920, III,159-182; also, Social Psychology, ch.. 11.
F.H.A.
