The Effect of Verbal Suggestion on Output and Variability of Muscular Work
- Author:
Charles W. Manzer
New YorTc University Statement of the Problem This study investigates the effect on the output and variability of muscular work of telling the subject that his task is “easy,” “medium” or “hard,” when actually the task is of uniform objective difficulty. Subjects and Apparatus The subjects were fifty men and fifty women students between the ages of eighteen and twenty-five years. The subjects of each sex were divided at random into a control and an experimental group with twenty-five subjects in each of these groups. The muscular work studied consisted of maximal contractions on the Smedley hand dynamometer. Among the subjects of each sex both groups, control and experimental, had practically the same mean initial strength of grip. Table 1 contains the means and the probTable 1 Means and Probable Error of Means of Initial Contractions by Both Control and Experimental Groups (Kilograms)
Men Subjects Control Group Experimental Group Women Subjects Control Group Experimental Group Number of cases……… Mean of first contractions. Probable Error of Mean.. 25 36.84 0.621 25 36.62 0.677 25 25.32 0.517 25 25.04 0.525 Difference between means of control and of exper. groups Probable Error of Difference Difference/P.E.diff. 0.22 0.291 0.76 0.28 0.233 1.20
able errors of the means of the first contractions. Among the men subjects, the difference between the means of the control and of the experimental group is 0.76 times its probable error; the correspond248 ing difference among the women subjects is 1.20 times its probable error. Since both differences are insignificant statistically, the control and experimental groups may be considered as samples from the same population.
Experimental Conditions and Procedure
The work was done by each subject individually in a small room, free from noise or other distraction. While doing the work the subject was seated at a table. The dynamometer on the table was clamped vertically to an iron stand which was free to move over the surface of the table at the convenience of the subject. Both the apparatus and the record were completely hidden from the subject throughout the experiment. The experiment was described to each subject as a study of work and rest and so far as is known no subject suspected its real nature. The following instructions were read to each subject at the beginning of the experimental period: This experiment is concerned with work and rest. The work consists of squeezing a handle which compresses a spring. There are three different loads: easy, medium, and hard. There is also a standard load. We are interested in comparing the easy, medium and hard loads with the standard load. Do your best each time you squeeze the handle. In order that you may do your best, you will be told what load you are using. The work is done in pairs: standard, easy; standard, medium; standard, hard; standard, standard. Do your best each time. After each contraction, the reading on the dial of the dynamometer was recorded by the experimenter and the pointer set back to zero. Each subject was given a practice period long enough to familiarize him with the apparatus and procedure but not so long as to produce fatigue.
In a series of preliminary trials it was found that, irrespective of the load reported to the subjects, there was a tendency to do more work in the second member of a pair of contractions than was done in the first contraction. In order to control this tendency to accentuate the second contraction of a pair, the suggestions used were given half the time just before the first contraction and half the time just before the second contraction. That is to say, half the time the loads were reported to the subjects as standard, easy ? standard, medium; standard, hard; while for the other half of the time, they were given as easy, standard; medium, standard; hard standard.
The work which was done in each contraction following the experimental suggestions, easy, medium, and hard, was expressed as a per cent of the work done in each contraction following the control suggestion, standard. For example, in the pair, standard, easy, the work done in the first contraction was used as the base in finding the percentage, while in the pair, easy, standard, the work done in the second contraction was used as the base. That this procedure did in fact correct for the tendency to accent the second contraction of a pair is shown by the percentages of work done in the control series (Tables 2 and 3). As would be expected theoretically, the percentages of work done in these control series, in which standard, standard were the suggestions used throughout, are all very near one hundred.
Each subject was given two tests; each test consisted of twenty pairs of contractions. In the experimental series, the verbal suggestions were given in the following order: standard, standard; standard, easy; standard, medium; standard, hard; standard, standard; standard, standard; standard, hard; standard, medium; standard, easy; standard, standard. When the test was repeated, the order of giving the suggestions was reversed: hard, standard; medium, standard; easy, standard, and so on. There was a onesecond interval between the members of a pair of contractions and a fifteen-second interval between successive pairs. The rest between tests was ten minutes. This interspersing of rest throughout the experiment was designed to keep fatigue at a minimum. Following each verbal suggestion, spoken by the experimenter, the subject made a maximal contraction on the hand dynamometer. The task, as has been said, was of uniform objective difficulty throughout the experiment.
Results
The results of this study will be presented under two heads: Work and Variability. The data of the men and of the women subjects will be treated separately.
Work. Tables 2 and 3 show the mean work done, the probable error and the probable error of the mean of the work series following each of the verbal suggestions of both control and experimental groups. The percentages for the control groups in Tables 2 and 3 are based upon the control contractions which correspond in location in the total work series to the contractions in the experimental series where the experimental suggestions, easy, medium,
Table 2 Work Done by Control and by Experimental Groups Men Subjects Control Group Suggestion Used Standard. . Standard. . Standard. . Mean (per cent) 100.76 100.68 100.21 Probable Error P.E., 5.06 4.92 5.49 0.72 0.70 0.78 Experimental Group Suggestion Used Easy Medium. Hard… Mean (per cent) 93.52 95.86 104.87 Probable Error 11.94 8.90 5.29 P.E.-? 1.69 1.26 0.75 Table 3 Work Done by Control and by Experimental Groups Women Subjects Control Group Suggestion Used Standard. Standard. Standard. Mean (per cent) 100.66 100.92 100.46 Probable Error P.E. 5.90 5.81 6.42 Mean 0.83 0.82 0.91 Experimental Group Suggestion Used Easy…. Medium. Hard…. Mean (per cent) 99.42 99.65 107.49 Probable Error 8.67 8.64 9.08 P.E., 1.23 1.22 1.28
and hard, were used. It should be emphasized that the only verbal suggestion used throughout the entire control series was the word “standard.” In view of the fact that the mean per cent of the control groups is 100 (or very near 100), it appears that the control suggestion, standard, produced a uniform base or level of work from which to measure the effect of the experimental suggestions. In the light of this uniformity of work following the suggestion, standard, we shall conclude that any significant difference between the work following the control suggestion and that following the experimental suggestions is indicative of the effect of these suggestions on work. That output of muscular work is actually affected by verbal suggestion can be seen upon examining the means, the probable errors and the probable errors of the means in the experimental groups (Tables 2 and 3). In the work of these experimental groups, there is no such uniformity as that which appeared in the work of the control groups. The means of the experimental group of men subjects (Table 2) vary from 93.52 per cent to 104.87 per cent; the probable errors range from 5.29 per cent to 11.94 per cent. Among the women subjects, the means of the experimental group (Table 3) vary from 99.42 per cent to 107.49 per cent; the probable errors range from 8.64 per cent to 9.08 per cent. The Men Subjects. The means in Table 2 indicate that the suggestion, easy, resulted in a reduction of output of approximately seven per cent; medium led to a reduction of about five per cent, while hard was followed by an increase of five per cent. The statistical significance of these differences is shown in Table 4.
Table 4 Comparison of the Effects on Work of the Verbal Suggestions, Standard, Easy, Medium and Hard: Reliability of the Differences Men Subjects Standard and EasyStandard and Medium Standard and Hard Easy and Medium Medium and Hard Easy and Hard Difference between means (per cent) P.E. of Diff Diff./P.E.diff -7.24 1.84 3.93 -4.82 1.44 3.35 +4.66 1.08 4.31 +2.34 2.11 1.11 +9.01 1.47 6.13 + 11.35 1.85 6.14 Women Subjects Standard and Easy Standard and Medium Standard and Hard Easy and Medium Medium and Hard Easy and Hard Difference between means (per cent) P.E. of Diff Diff./P.E.diff. -1.24 1.48 0.84 -1.27 1.46 0.87 +7.03 1.57 4.48 +0.23 1.73 0.13 +7.84 1.77 4.43 +8.07 1.77 4.56
This table contains the differences between the control suggestion and each of the experimental suggestions and also the differences between each pair of experimental suggestions, the probable errors of these differences and the critical ratios. Among the men subjects, the differences between standard and easy and between standard and medium are 3.93 and 3.35 times their probable errors, respectively. The difference between standard and hard is 4.31 times its probable error. Accepting the convention which regards a difference that is four or more times its probable error as significant, the statement is justified that the suggestions, easy and medium, almost certainly lead to a reduction in output among the men subjects (99.6 chances in 100 and 98.7 chances in 100, reEFFECT OF VERBAL SUGGESTION 253 spectively), while the suggestion, hard, is followed by an increase in output that is fully significant, since this difference is 4.31 times its probable error.
When the experimental suggestions are compared one with another, it is found that the suggestions, easy and medium, have essentially the same effect on the output of the men subjects?both these suggestions cause a reduction in work. Easy causes a somewhat greater reduction than does medium but the difference between the effects of these two suggestions is not significant since it is only 1.11 times its probable error (77 chances in 100). The suggestion, hard, causes an increase of nine per cent above the work done following medium. Each of these differences is more than six times its probable error.
The Women Subjects. Among the women subjects (Table 3), the control suggestion, standard, produces a highly uniform output. Unlike the men subjects, however, the women do as much work following the suggestions, easy and medium, as they do after the control suggestion, standard. Such slight reductions in output as these suggestions cause (about one per cent) are only 0.84 and 0.87 times their respective probable errors. The suggestion, hard, on the other hand, causes an increase in the output of the women subjects of seven per cent; this difference is 4.48 times its probable error (Table 4) and is therefore statistically significant. Comparisons among the experimental suggestions show no difference between the effect of easy and that of medium. Hard causes almost eight per cent more work than does either medium or easy; the differences between hard and easy and between hard and medium are more than four times their respective probable errors. The difference between the effects of verbal suggestion upon the work of the men and that of the women found in the present study is similar to the difference found by Hurlock1 between the effects of praise and reproof upon boys and girls. Hurlock found the boys the more responsive and because of the fact that the girls were more conscientious and interested from the start, the girls could not raise their rate of achievement as much as the boys when incentives were used. The “conscientiousness” of the women subjects in the present study may have prevented their output from falling below “standard,” but they did respond to the suggestion, hard, with an increase in work.
i E. B. Hurlock, An evaluation of certain incentives used in school work. J. Educ. Psychol., 1925, 16, 145-159.
In this connection it should be noted that difference in sex, which is generally regarded as important in the effectiveness of suggestions and incentives, has not been controlled in this study. Any generalization regarding the effects of these verbal suggestions on muscular work would have to await the findings of a study similar to the present one carried out by a woman experimenter. Variability. The effect of the verbal suggestions upon variability can be found by comparing the work done in the control member of the pairs of contractions with that done in the experimental member; that is to say, by finding the correlation between the work done in the pairs of contractions. Theoretically, it should be expected that the correlation between pairs of contractions made under the same objective experimental conditions, separated by only a one-second interval and given as a response to the same verbal suggestion (standard, standard), would be very near unity. Such coefficients would provide a base with which to compare the effect of the experimental verbal suggestions, easy, medium, and hard, upon variability of output. In other words, if the experimental verbal suggestions, given just before one of the members of the experimental pairs of contractions, should cause an increase in variability of output, such increased variability would be shown in a reduction of the correlation between the work done in these paired contractions. It will be recalled that one member of the paired contractions was preceded by the control suggestion, the other member was preceded by one of the experimental suggestions. Table 5 contains the coefficients of correlation found between pairs of contractions in both the control and the experimental series of the men and of the women subjects. The correlations in the control groups show that the data approach very closely to the theoretical expectations stated above. Among the men subjects the correlations in the control group range from r = .9434 to r = .9439. These coefficients show that the verbal suggestion, standard, given throughout the control series, leads to output of great uniformity. A similar effect appears in the control group of the women subjects; the coefficients are of the order of r = .9400. In the experimental series (Table 5) it will be observed that the suggestion, easy, reduced the correlation to r = .6930 among the men and to r = .7652 among the women. The suggestion, hard, lowers the correlation among the men to r = .8906 and among the women to r=.8850. In the case of the suggestion, medium, the
Table 5 Variability op Work Done by Control and by Experimental Groups: Correlations Between Pairs op Contractions Men Subjects Control Group Suggestion Used Standard. Standard. Standard. .9405 .9248 .9334 P.E.r .0110 .0157 .0122 Experimental Group Suggestion Used Easy Medium. Hard .6930 .8210 .8906 P.E.r .0496 .0311 .0197 Women Subjects Control Group Suggestion Used Standard. Standard. Standard. .9435 .9439 .9434 P.E.r .0105 .0104 .0105 Experimental Group Suggestion Used Easy…. Medium. Hard…. .7652 .7867 .8850 P.E.r .0395 .0364 .0207
coefficients are intermediate between that for easy and that for hard in both groups of subjects. The conclusion seems justified that all the experimental verbal suggestions increase the variability of output, though they differ regarding the extent of increase in variability produced. Output is most uniform when subjects are working under “standard” conditions.
This study justifies drawing the conclusion that even when the task is of constant objective difficulty the description of the task to the worker as easy or hard affects the amount and the variability of the work. When he is making muscular contractions, the subject is reacting not merely to the load but rather to a total situation. Changes in such an apparently minor element in the total situation as the subject’s attitude toward the difficulty of his task result in significant changes in output and in variability.
Summary and Conclusions
The problem studied was the effect on the output and variability of muscular work of telling the subject that his task was “easy,” “medium” or “hard,” when actually the task was of uniform objective difficulty. The work consisted of pairs of maximal contractions on the Smedley hand dynamometer.
The subjects were fifty men and fifty women college students. Subjects of eacli sex were divided into two equal groups, control and experimental. Among the subjects of each sex, the control and experimental groups had the same initial mean strength of grip.
The word standard was the control suggestion. The words easy, medium and hard were the experimental suggestions. The following conclusions are drawn:
1. Among the men subjects, the suggestions, easy and medium, cause a reduction in output of 7.24 per cent and 4.82 per cent, respectively, below the work done following the control suggestion, standard. The suggestion, hard, causes an increase in work of 4.66 per cent above that done following the control suggestion. The critical ratios of these differences are 3.93, 3.35 and 4.31, respectively. The work done following the suggestion, hard, is 11.35 per cent greater than that done following the suggestion, easy; the critical ratio is 6.14.
2. Among the women subjects, the suggestions, easy and medium, cause no reduction in output that is statistically significant. Hard causes an increase above the work done following the control suggestion of 7 per cent; this difference is 4.43 times its probable error. The work done following the suggestion, hard, is 8 per cent greater than that following the suggestion, easy; the critical ratio is 4.56.
3. Among both men and women subjects, variability of output is increased by all three experimental verbal suggestions. The suggestion, easy, increases variability most, while hard causes the smallest increase in variability.
Disclaimer
The historical material in this project falls into one of three categories for clearances and permissions:
Material currently under copyright, made available with a Creative Commons license chosen by the publisher.
Material that is in the public domain
Material identified by the Welcome Trust as an Orphan Work, made available with a Creative Commons Attribution-NonCommercial 4.0 International License.
While we are in the process of adding metadata to the articles, please check the article at its original source for specific copyrights.