A Study of Some Social Factors in Perception

Chapter 3

Table of Contents | Next | Previous

I. RANGE EXPERIMENTS

We have seen that the work done on the method of single stimuli shows that the subjects establish a scale even though a standard stimulus is lacking ; and every stimulus presented is judged according to its experienced place in the scale.

We presented the same single stimulus (the point of light) 100 times. Of course the subject does not know that the stimulus is always the same. How will he distribute his judgments of the distances which the light moved? An objective scale for judgment is lacking. The question is, will the subject distribute his judgments in a haphazard way because of the lack of an objective basis for comparison? Or will he himself furnish a scale, and a reference point on that scale subjectively, and distribute his judgments around that reference point? In the physical absence of an anchorage point will an anchorage be produced subjectively?

What will a group of individuals do when they collectively face such a situation, which lacks physical basis for a standard or norm? If a common scale and a common norm are produced for the group as a whole, this will be very significant as an approach to the question suggested by our review of anthropological field cases.

RANGE EXPERIMENT---PART I---(INDIVIDUAL)

There were 19 subjects in this experiment.

The observer was seated at a table on which was the key which operates the timing device. The following instructions were given:

"When the room is completely dark, I shall give you the signal READY, and then show you a point of light. After a short time the light will start to move. As soon as you see it move, press the key. A few seconds later the light will disappear. Then tell me the distance it moved. Try to make your estimates as accurate as possible. "

These instructions summarize the general procedure. Then the subject's head was placed in a Stoelting head-rest, to do away at least with the head movement. The signal "Ready" was given before each exposure by a pencil tap on the table. If the observer did not press the key within 30 seconds, the light was covered and the distance was recorded as zero. This happened seldom. After the


( 24) observer pressed the key, an exposure time of 2 seconds was used in all experiments. Generally the key was pressed by the subjects not more than 5 seconds after the exposure. The subject reported orally the distance moved. The experimenter recorded each judgment as soon as it was spoken by the subject, writing each judgment on a separate sheet of a small paper pad. 100 judgments were obtained from each observer. The subjects reported their estimated distances in inches (or fractions of inches).

At the end of the experiment the subjects were asked to fill out a sheet with these questions

1. Was it difficult to estimate the distance? If Yes, give the reasons.

2. Show with a diagram the way the light moved.

3. Did you try to find some method of your own, so that you could make your judgments more accurate?

Some of these introspections will be briefly reported later, because they further illustrate the questions raised.

Results: The actual data, the distances reported, are given in the frequency distributions in the Appendix, Table XVI, pages 55-56. The range, mode, median, Q, and P.E. (mdn.) were computed for every 100 judgments. (See Table I.)

 
Table I Range Experiment I
Subject Range Mode Median Q P.E. (Mdn.)
1 12 (0 - 12) 4 4.80 1.35 + .17
2   9 (4 - 13) 6 7.89 1.55 + .19
3 3 (½ - 3 ½ ) 1 ½ 1.72 0.51 + .06
4 6 ( 2 - 8) 6 5.45 1.04 + .13
5 6 ( 0 - 6) 3 3.45 0.97 + .12
6 2 ( 1 - 3) 2 2.16 0.53 + .07
7 13 ( 2 - 15) 8 8.39 1.70 + .21
8 11 ( 5 - 16) 9 9.62 1.32 + .16
9 2 ½ ( ½ - 3) 2 1.37 0.65 + .08
10 5/8 (0 - 5/8) 3/8 0.36 0.11 + .01
11 10 ( 4 - 14) 8 8.70 1.66 + .21
12 4 ½ ( ½ - 5) 2 2.61 1.01 + .13
13 6 ( 0 - 6 ) 2 2.96 1.05 + .13
14 9 ( 1 - 10) 4 5.21 1.28 + .16
15 2 3/4 (1/4 - 3) 1/2 0.78 0.37 + .05
16 4 ( 2 - 6) 3 4.12 0.72 + .09
17 8 ( 0 - 8) 1 2.36 1.08 + .13
18 6 ( 2 - 8) 6 6.25 0.78 + .10
19 1 1/4 (1/4 - 1 1/2) 3/4 0.80 0.21 + .03

One will note that the frequency tables show that these distributions approach normality with different degrees of skewness, in


(25) some cases positive and in some negative. The median value may deviate in various degrees from the middle of the range. These points are not of particular interest for the present problem, though such an analysis may be of special interest to the psychophysicist. For our problem the important fact in these results is that the subjects subjectively establish a range and a point within that range which is peculiar to the individual and which may differ from the range and the median point established by other individuals. Among these 19 subjects the shortest range is 1 1/4 extending from 1/4 to 1 ½ inches; the greatest range is 13, extending from 2 to 15 inches. The minimum median is .36 and the maximum is 9.62 inches. The variation of ranges and medians within these distances holds only for our specific conditions-exposure time, distance between light and observer, and diameter of the light. The facts summarized above may be readily seen in the bar diagram on p. 26.

The introspections give qualitative support to the conclusions summarized above. The answers filled in to the question : Was it difficult to estimate the distance? show that they feel the lack of reference points. Let us quote some of the representative ones:

1. "Darkness left no guide for distance."

2. "No set position from which to judge how far."

3. "Didn't know direction it would move."

4. "Lack of visible neighboring objects."

5. "No fixed point from which to judge distance."

The answers given to the third question give a qualitative statement to support the conclusion reached on the basis of quantitative results, to the effect that they established a subjective basis of comparison. Some examples of the point are the following

1. "Compared with previous distance."

2. "Judgments are all relative."

3. "Compared successive judgments."

4. "Approximated distance of spot from me, and used that."

5. "First estimate as standard."

6. "Thought of using radium dial of watch for judging distance. "

We find this verification over and over again in the introspection obtained from the observers in the group experiments.

RANGE EXPERIMENT---PART II (INDIVIDUAL)

The special task of this experiment was to find whether, after once a range and a norm within that range are established, they


( 26) sherif_1935a_3.gif


( 27) persist on subsequent occasions. In this part of the range experiment, 4 subjects were used. Instead of 100, 300 judgments were obtained from each subject, in three different sessions on different (lays within a week. The range, mode, median, Q, and P.E. (mdn.) for each 100 judgments are given in Table II, below. These results show that once a range and a point within that range are established, there is a tendency to preserve them in the subsequent sittings.

Table II Range Experiment II
Subject Range Mode Median Q P.E. (Mdn.)
One          
I 4 (2-6) 3 3.62 0.72 0.09
II 5 (1-6) 3 2.95 0.66 0.08
III 5 (1-6) 3 3.37 0.56 0.07
Two          
I 2 (1-3) 2 1.66 0.53 0.07
II 3 (½ - 3 ½ ) 2 1.9 0.52 0.07
III 2 ½ (½ - 3) 2 1.61 0.47 0.06
Three          
I 5/8 (0 - 5/8) 3/8 0.29 0.10 0.01
II 6/8 (1/8 - 7/8) 3/8 0.46 0.15 0.02
III 6/8 (1/8 - 7/8) 2/8 0.33 0.11 0.01
Four          
I 1 1/4 (1/4 - 1 ½ ) 3/4 0.67 0.2 0.03
II 1 1/8 (1/8 - 1 1/4) 4/8 0.55 0.15 0.02
III 1 1/8 (1/8 - 1 1/4) 5/8 0.58 0.13 0.02

II. THE GROUP AS A FACTOR IN PERCEPTION

The facts in the above experiments led to the conclusion that (1) every observer establishes a range of his own ; (2) the judgments within that range are fairly normally distributed around a median value (norm). We started with the individual, to find out the individual reactions first. With the group experiments we extend our method to an important field of social psychology. The question becomes: What will a group do when confronting such an unstable situation? Will different individuals establish their own ranges and the norms within those ranges, or will the group establish a range (scale) of its own, and produce a norm (a median value) peculiar to itself? This involves one of the most debated questions in social psychology. We are concerned with the production of a new standard in a group (reminding us of "group mind"


( 28) theories), and with the idea of the arousal of a new norm in a group situation so well developed in Durkheim's "Formes Ιlιmentaires de la Vie Religieuse."

A further question is this : How much convergence of ranges and medians (norms) will there be (a) when the individual in one session faces the situation alone and then is brought into the group situation; (b) when he faces the situation in the group first and then alone?

There were 8 groups of 2 subjects and 8 groups of 3 subjects. Four groups started with the individual situation (one session for each individual), and then functioned as groups. Four groups started as groups (3 sessions-all subjects of the group present in all 3), and were then broken up and studied in the individual situation. These arrangements are shown below. As before, 100 judgments were taken from each subject in each session.

Starting with the Individual Situation:
Session I II III IV
 

 1

     
  Individual  2 Group Group Group
  3      
Starting with the Group Situation
Session I II III IV
       

1

  Group Group Group Individual   2
        3

The general plan above holds true for the groups of 2 and 3 subjects.

The experimental setting in general is the same as in previous experiments. The exposure time (after the key is pressed) is the same. The head-rest was, however, not used, as the previous experiment showed us that it does not make much difference. As the subjects were new to the experimenter, he could not tell from the voice who was giving a judgment. Each subject pressed a pushbutton at the same time as he gave his judgment aloud. This pushbutton operated a signal light (yellow, red, or green as the case might be), which could be seen only by the experimenter, as it was separated by a partition from the vision of the subjects. It must be repeated that the colored light was very dim; it did not have intensity enough to make anything in the room visible. (See Chapter Two.)

The instruction sheet ran as follows:

"When the room is completely dark, I shall give you the signal READY, and then show you a point of light. After a short time


( 29) the light will start to move. As soon as you see it move, press the key. (Press it the moment you see the light move. Don't wait for the other persons.) A few seconds later the light will disappear. Then tell me the distance it moved. When you give your estimate, press the push-button. Try to make your estimates as accurate as possible."

This also describes the general procedure. After the subjects read the instruction sheets they were told that they could give their judgments in any order and they could change the order from time to time. In accordance with this, the subjects changed the order in which they gave their judgments during the course of the experiment. Each of the 100 judgments obtained from each subject in each session was written by the experimenter on a different sheet of a small pad and then torn off. As the subjects in the group were unknown to the experimenter and the experimenter could not recognize their voices, each subject's judgments were written down on a pad of a different color, corresponding to the color of the glow produced by the pressing of the push-button by each subject.

As will be noticed in the instructions, the subjects were left free as to the order in which they would give their judgments. This was done on purpose. The task set in the present study is to find what a group, consisting of people who have not established a relationship to each other, affective or otherwise, will do when they face such a novel field of stimulation. They were told at the start to give their judgments in random order, and to change the order in which they gave their judgments once in a while. Whether the judgments of the person who utters his judgments first have more influence than the others becomes a study in leadership which is a further interesting study.

In order to find out whether the subjects became conscious of the range and norm (median value) they established subjectively, the following questions were added:

Between what maximum and minimum did the distances vary? What was the most frequent distance the light moved? The similarity of these introspectively reported ranges and norms to the ranges and norms revealed by the computation of 100 judgments, would indicate how conscious the subjects became of the range and median established in the group. Also at the end of all the experiments the following question was added to the introspection blank to find out whether they were conscious of the influence of the group on their judgments: Do you think you were influenced in


( 30) your estimates by the judgments of the other persons in the experiment? The question: How did the light move? was eliminated from the introspection sheets, because if the subjects' attention were to be concentrated on the direction they would soon find out the discrepancy of the directions they experienced and thus conic to the conclusion that it was an "illusion" after all. This teas exactly the way that Schweizer found in 1858 that the movement of the stars on a dark night was an "illusion."

Results: The data were tabulated in frequency tables, each table representing 100 judgments for each individual subject in group and individual sessions. These results give an idea of the convergence and divergence of individuals (1) in individual sessions; (2) in group sessions, when first started with the group situation and then worked on individually, and (3) vice versa. As before, the range, mode, median, Q, and P.E. (mdn.) are worked out for each subject for each experimental session. The results obtained from the groups of three subjects are given in the Appendix (pp. 56-60). Results for the groups of two subjects show essentially the same trend. Since space is limited, these are not presented here.

The crucial point for our problem is to see from the results the comparison of ranges and norms (median values) when the individuals face the stimulus field under these conditions

(a) First a subject's individual range and median are found for a whole session, and then he is put into the group to face the same situation (for three successive sessions) as a member of the group, so that we may note how much he converges toward a common range and median for the group.

(b) When the subject first faces the situation in the group (for three successive sessions on different days) and then faces the smile situation alone in Session IV on a different day, we note hour closely he sticks to the common norm established in group sessions. To give a concise picture, the median values (norms) established by each subject in each successive session are presented in the graphs on pp. 32-33.

Certain facts stand out in these tables and graphs. When subjects start with individual sessions the median values which they establish individually differ from each other considerably. When on successive sessions they work together their medians tend to converge-a "funnel-shaped" relationship, the opening of the funnel representing individual sessions. On the other hand, when subjects start in-group situation there is convergence at once, which is main-


(31) -tained in successive sessions, including the last individual session. Groups of 2, starting with the group situation, tend to keep the same general level of median values in successive sessions. In groups of 3, starting with the group situations, there may be a rise or fall, as well as a keeping to the same general level in the median values, as seen in cases of the second and fourth groups. But when there is a rise the subjects rise as a whole, and they fall as a whole when they fall. The closeness of the medians of individuals in the group, which suggests the establishment of a common norm for the group in the cases of groups starting with the group situation, is a very important fact that we wish to stress. Especially important is the fact that the divergence of the median values established by the subjects in the individual session after the group sessions is small. Compared with this, when the subject starts with an individual session and then is brought into a group the convergence of the medians (see the funnel-shaped relationship) is not so close-suggesting that if individual faces a stimulus situation and patterns it in his own way first, the group influence is not so dominating as when he faces the situation in the group first.

This point is subjected to a special analysis. The differences in the medians of each possible pair in each session and the reliabilities ole.gif are computed. In addition to this, for groups starting with the group situation, the differences between the medians of the last (3rd session) group session and the individual session (4th session) were computed. This is one of the crucial tests for our problem. These results for all groups are presented in the tables on pp. 34 11.These results show that the differences of the medians of the pairs of subjects are very small in cases of groups starting with a group session, and are statistically unreliable. On the other hand, the differences between the pairs of subjects starting with the individual sessions are considerably larger and in many cases statistically reliable.

Also in groups starting with the group situation the differences between the medians of the last (Session III) group session and the individual session (Session IV) of each individual (designated as Self-D : self-difference) are small and statistically unreliable in most cases. This is, we repeat, an important point for social psychology, suggesting that once an individual faces a stimulus situation in the group situation for the first time and reacts to it with the norm of the group, there is a tendency to continue to react to the same situ-


(32)

sherif_1935a_31.gif


(33)

sherif_1935a_32.gif


(34) -ation with the same norm established in the group, even when the subject is no longer in the group situation. No attempt has been made to make a careful analysis of the differences between the groups of 2 and 3 subjects. These two kinds of groups give essentially the same results.

TABLE III DIFFERENCES BETWEEN MEDIANS AND CRITICAL RATIOS OF DIFFERENCES OF PAIRS OF SUBJECTS
Groups of Two Subjects starting with the individual situation
  Session D 1 ole1.gif
First GroupI Individual 3.12 45.93
II Group 1.03   8.24
III Group 2.12   9.22
IV Group 0.60   5.50
Second GroupI Individual 6.81 32.72
II Group 1.50   9.61
III Group 0.41 3.04
IV Group 0.34 3.44
Third GroupI Individual 2.01 10.56
II Group 0.02 0.35
III Group 0.76 12.64
IV Group 0.77 15.43
Fourth GroupI Individual 5.59 19.07
II Group 1.19 7.14
III Group 0.54 5.15
IV Group 0.45 4.81
1 D represents the difference between the medians of the judgments of the two subjects in each group


(35)

TABLE IV DIFFERENCES BETWEEN MEDIANS AND CRITICAL RATIOS OF DIFFERENCES OF PAIRS OF SUBJECTS
Groups of Two Subjects Starting with the Group Situation
  Session D 1 ole2.gif 
First Group      
I Group .04 0.05
II Group .35 2.34
III Group .19 1.67
Self-D2 S1 (III-IV) .37 3.08
  S2 (III-IV) .17 1.31
IV Individual .73 5.75
Second Group      
I Group .05 0.24
II Group .38 1.80
III Group .20 1.00
Self-D2 S1 (III-IV) .15 0.75
  S2 (III-IV) .98 5.19
IV Individual .63 3.31
Third Group      
I Group .02 0.27
II Group .03 0.75
III Group .02 0.55
Self-D2 S1 (III-IV) .04 1.33
  S2 (III-IV) .24 6.00
IV Individual .26 8.67
Fourth Group      
I Group .07 0.63
II Group .35 3.05
III Group .05 0.52
Self-D2 S1 (III-IV) .10 1.00
  S2 (III-IV) .33 3.00
IV Individual .18 1.65
1 D represents the difference between the medians of the judgments of the two subjects in each group

2 Self-D (self-difference) represents the difference between the medians of the same subject in the last group session (Session III) and the individual session (Session IV)


(38)

TABLE V DIFFERENCES BETWEEN MEDIANS AND CRITICAL RATIOS OF DIFFERENCES OF PAIRS OF SUBJECTS
Groups of Three Subjects Starting with the Individual Situation
  D ole3.gif
First Group    
Session I — Individual    
S 1 - S 2 1 6.84 38.00
S 1 - S 3 5.67 31.50
S 2 - S 3 1.17 23.40
Session II — Group    
S 1 - S 2 1 1.89 9.94
S 1 - S 3 1.15 6.77
S 2 - S 3 0.74 5.36
Session III — Group    
S 1 - S 2 1 1.67 11.92
S 1 - S 3 1.14 8.27
S 2 - S 3 0.53 5.30
Session IV — Group    
S 1 - S 2 1 0.08 0.55
S 1 - S 3 0.19 1.42
S 2 - S 3 0.11 1.10
1 S 1 - S 2 stands for the difference (D) of the medians of judgments of Subject 1 and Subject 2 in the group, etc.

TABLE VI DIFFERENCES BETWEEN MEDIANS AND CRITICAL RATIOS OF DIFFERENCES OF PAIRS OF SUBJECTS
Groups of Three Subjects Starting with the Individual Situation
  D ole4.gif
Second Group    
Session I — Individual    
S 1 - S 2 1 1.78 14.81
S 1 - S 3 0.50 3.83
S 2 - S 3 1.07 21.28
Session II — Group    
S 1 - S 2 1 0.27 5.60
S 1 - S 3 0.18 2.54
S 2 - S 3 0.45 9.06
Session III — Group    
S 1 - S 2 1 0.49 6.13
S 1 - S 3 0.26 2.83
S 2 - S 3 0.23 2.50
Session IV — Group    
S 1 - S 2 1 0.34 5.80
S 1 - S 3 0.21 3.58
S 2 - S 3 0.13  2.17
1 S 1 - S 2 stands for the difference (D) of the medians of judgments of Subject 1 and Subject 2 in the group, etc.


(37)


TABLE VII DIFFERENCES BETWEEN MEDIANS AND CRITICAL RATIOS OF DIFFERENCES OF PAIRS OF SUBJECTS
Groups of Three Subjects Starting with the Individual Situation
  D ole5.gif
Third Group    
Session I — Individual    
S 1 - S 2 1 0.82 6.31
S 1 - S 3 2.42 14.24
S 2 - S 3 1.60 12.31
Session II — Group    
S 1 - S 2 1 0.09 0.61
S 1 - S 3 0.87 5.12
S 2 - S 3 0.78 3.96
Session III — Group    
S 1 - S 2 1 0.08 0.40
S 1 - S 3 0.66 5.07
S 2 - S 3 0.58 2.76
Session IV — Group    
S 1 - S 2 1 0.36 3.66
S 1 - S 3 0.00 0.00
S 2 - S 3 0.36  3.60
1 S 1 - S 2 stands for the difference (D) of the medians of judgments of Subject 1 and Subject 2 in the group, etc.


TABLE VIII DIFFERENCES BETWEEN MEDIANS AND CRITICAL RATIOS OF DIFFERENCES OF PAIRS OF SUBJECTS
Groups of Three Subjects Starting with the Individual Situation
  D ole6.gif
Fourth Group    
Session I — Individual    
S 1 - S 2 1 2.74 14.42
S 1 - S 3 1.37 7.61
S 2 - S 3 1.37 6.75
Session II — Group    
S 1 - S 2 1 0.04 0.27
S 1 - S 3 0.15 0.98
S 2 - S 3 .019 1.14
Session III — Group    
S 1 - S 2 1 0.13 0.57
S 1 - S 3 0.90 5.00
S 2 - S 3 0.77 4.52
Session IV — Group    
S 1 - S 2 1 0.33 2.06
S 1 - S 3 0.06 0.41
S 2 - S 3 0.27  1.71
1 S 1 - S 2 stands for the difference (D) of the medians of judgments of Subject 1 and Subject 2 in the group, etc.



(38)


TABLE IX DIFFERENCES BETWEEN MEDIANS AND CRITICAL RATIOS OF DIFFERENCES OF PAIRS OF SUBJECTS
Groups of Three Subjects Starting with the Group Situation
  D ole7.gif
First Group    
Session I — Group    
S 1 - S 2 1 0.65 4.64
S 1 - S 3 0.01 0.06
S 2 - S 3 0.64 4.92
Session II — Group    
S 1 - S 2 1 0.22 2.20
S 1 - S 3 0.10 1.25
S 2 - S 3 0.12 1.33
Session III — Group    
S 1 - S 2 1 0.10 4.44
S 1 - S 3 0.15 1.87
S 2 - S 3 0.25 2.77
Session III-IV (Self-D)    
S 1 III - S 1 IV 0.47 4.27
S 2 III - S 2 IV 0.11 1.10
S 3 III - S 3 IV 0.19 2.37
Session IV — Individual    
S 1 - S 2 1 0.98 8.16
S 1 - S 3 0.81 6.70
S 2 - S 3 0.17 1.88
1 S 1 - S 2 stands for the difference (D) of the medians of judgments of Subject 1 and Subject 2 in the group, etc.

2 Self-D (self-difference) represents the difference between the medians of the same subject in the last group session (Session III) and the individual session (Session IV)


(39)

TABLE X DIFFERENCES BETWEEN MEDIANS AND CRITICAL RATIOS OF DIFFERENCES OF PAIRS OF SUBJECTS
Groups of Three Subjects Starting with the Group Situation
  D ole8.gif
Second Group    
Session I — Group    
S 1 - S 2 1 0.59 2.70
S 1 - S 3 0.67 3.20
S 2 - S 3 0.08 0.38
Session II — Group    
S 1 - S 2 1 0.27 2.07
S 1 - S 3 0.05 0.38
S 2 - S 3 0.22 2.00
Session III — Group    
S 1 - S 2 1 0.24 2.17
S 1 - S 3 0.07 0.63
S 2 - S 3 0.17 1.55
Session III-IV (Self-D)    
S 1 III - S 1 IV 0.61 5.55
S 2 III - S 2 IV 0.45 5.00
S 3 III - S 3 IV 0.07 0.63
Session IV — Individual    
S 1 - S 2 1 0.82 9.11
S 1 - S 3 0.61 5.55
S 2 - S 3 0.21 2.21
1 S 1 - S 2 stands for the difference (D) of the medians of judgments of Subject 1 and Subject 2 in the group, etc.

2 Self-D (self-difference) represents the difference between the medians of the same subject in the last group session (Session III) and the individual session (Session IV)


(40)

TABLE XI DIFFERENCES BETWEEN MEDIANS AND CRITICAL RATIOS OF DIFFERENCES OF PAIRS OF SUBJECTS
Groups of Three Subjects Starting with the Group Situation
  D ole9.gif
Third Group    
Session I — Group    
S 1 - S 2 1 0.03 0.16
S 1 - S 3 0.33 1.94
S 2 - S 3 0.30 2.00
Session II — Group    
S 1 - S 2 1 0.66 4.13
S 1 - S 3 0.16 1.00
S 2 - S 3 0.50 3.60
Session III — Group    
S 1 - S 2 1 0.16 1.25
S 1 - S 3 0.13 1.00
S 2 - S 3 0.29 2.41
Session III-IV (Self-D)    
S 1 III - S 1 IV 0.06 0.55
S 2 III - S 2 IV 0.37 3.70
S 3 III - S 3 IV 0.72 6.00
Session IV — Individual    
S 1 - S 2 1 0.27 3.00
S 1 - S 3  0.53 4.82
S 2 - S 3 0.80  8.00
1 S 1 - S 2 stands for the difference (D) of the medians of judgments of Subject 1 and Subject 2 in the group, etc.

2 Self-D (self-difference) represents the difference between the medians of the same subject in the last group session (Session III) and the individual session (Session IV)


(41)

TABLE XI DIFFERENCES BETWEEN MEDIANS AND CRITICAL RATIOS OF DIFFERENCES OF PAIRS OF SUBJECTS
Groups of Three Subjects Starting with the Group Situation
  D ole10.gif
Fourth Group    
Session I — Group    
S 1 - S 2 1 1.10 2.90
S 1 - S 3 1.73 4.80
S 2 - S 3 0.63 1.90
Session II — Group    
S 1 - S 2 1 0.02 0.09
S 1 - S 3 0.58 2.31
S 2 - S 3 0.60 2.42
Session III — Group    
S 1 - S 2 1 0.19 1.10
S 1 - S 3 0.16 0.72
S 2 - S 3 0.03 0.14
Session III-IV (Self-D)    
S 1 III - S 1 IV 0.17 0.94
S 2 III - S 2 IV 0.20 0.95
S 3 III - S 3 IV 0.02 0.09
Session IV — Individual    
S 1 - S 2 1 0.56 2.66
S 1 - S 3 0.31 1.61
S 2 - S 3 0.25  1.19
1 S 1 - S 2 stands for the difference (D) of the medians of judgments of Subject 1 and Subject 2 in the group, etc.

2 Self-D (self-difference) represents the difference between the medians of the same subject in the last group session (Session III) and the individual session (Session IV)

CONCLUSIONS FROM INDIVIDUAL AND GROUP RANGE EXPERIMENTS

We may summarize these results in a few words : When individuals face this new and unstable situation first individually and then in a group, each establishes a range and a norm (standard) within that range ; the range and the norms tend to converge when the subjects come into a group situation. But the convergence is not as close as when they start with the group situation first.

When individuals face this new, unstable situation as members of a group for the first time, a range (a scale) and a norm (standard) within that range are established which are peculiar to the group, and afterwards when they face the same situation alone they stick to the range and norm established in the group.


( 42)

INTROSPECTIONS

The introspections from these group experiments verify the points that came out in the introspections in the individual range experiment as to the experience of the lack of reference points.

In closing this section we cannot help giving a very interesting case of a group of 3 starting with the group situation. The subjects in the fourth group of three starting with the group situation gave an unusually large number of zero judgments in Session I with modes at zero for all three subjects. After filling the introspection blank one subject asked. "Was the light moving really?" The experimenter, not knowing what to say, asked, "Why?" The subject answered that in some cases he tried to put his finger between his eye and the light, and the light did not cross the finger.

Before Session II started with the same group the experimenter told them not to move any part of their bodies during the experiment. The mode rose to 4 in all three cases and kept that level throughout, including the individual sessions for each of the three subjects. The first session is definitely a case of polarization around one person. After all the experiments were over the experimenter asked this particular subject to write down his position in the educational institutions with which he was connected. He reported that he was president of his college fraternity, manager of the football team, etc. As can be seen readily, the case of this particular group contains good suggestions for new "prestige" experiments.

To the question : "Do you think you were influenced by the judgments of the other persons in the experiments?" appearing on the sheets at the end of the last session, comparatively few (about 25%) answered that they were. People do not have to be aware of the fact that they are being influenced by the group situation.

The last point has to do with the time relationships of the group influence. Are the subjects influenced by each judgment at the moment it is given by some other member of the group, or does the effect arise gradually throughout the experiment? The serial inspection of our results shows that it is largely a temporal affair, not a question of one particular exposure, though at times this may happen.

III. SUGGESTION AS A FACTOR IN PERCEPTION

We had occasion to refer to the fact that a single point has no definite locus ; it cannot be located at any definite place. It is experienced to move in an erratic way because it has no definite direc-


( 43) -tion. But when a definite attitude is taken it is experienced to move in the direction given by the attitude. In the previous experiment the influence of a group on the perception of the extent of the autokinetic movement for a short time was studied. In the present experiment our task is to study the influence of suggestion on the direction of the movement.

It is reported by those who have worked on autokinetic movement that the direction is variable, and that voluntary control is possible, (Bourdon, Charpentier, Carr, Adams). In this experiment we tried to induce this voluntary control by suggestion. Adams has already tested the influence of suggestion on three naive subjects:

"These subjects, who were much interested in abnormal psychology, were told that the illusion was a new form of planchette which would form any simple figure or letter which they thought of while fixating the light. With one of the subjects, the suggestion worked perfectly, only one failure being made in 62 trials. The letters and figures were not always in their proper positions, sometimes being inverted, or suffering other displacements, but in all but one case, the desired letter or figure was formed. The subject was much surprised at this failure.

"The other two subjects had difficulty in getting the illusion at first, but when they succeeded in obtaining it, the desired letter was formed in 80% of the cases." (1, pp. 41-42.)

In our preliminary experiments in 1932, different directions were suggested at different times during the same experimental session. It was found that the subject has the tendency to stick to the direction suggested in the first part of the experiment. Therefore in the main experiments only one direction was used throughout the whole session. The general set-up and procedure of this experiment were the same in general outline as described in the previous range experiment. Right and left directions were used exclusively (up and down might just as well be used). The instructions ran as follows:

FOR RIGHT INSTRUCTION

When the room is completely dark I shall show you a point of light. After a short time it will start to move to your right. As soon as you see the light move, press the key. A few seconds later the light will disappear. Then tell me the distance it moved. Try to make your estimate as accurate as possible.

For left direction left was substituted for the word right. We intentionally led the subject to think that the experiment was concerned with his sensitivity in estimating the distance which the light


( 44) moves, because one of the essential conditions in a suggestion experiment is not to give the idea that one is being given a suggestion. Fifty exposures were made in each session. A longer period in a pitch-dark room might produce negativism. The subjects were not asked to report the direction in addition to distance at each exposure, to avoid arousing any suspicion on their part that the direction is important. At the end of 50 trials, among other questions the subject was asked to make a diagram of the direction, and approximate frequency distribution if he drew more than one direction. During the experiment the subjects did not say anything about direction if they "perceived" the movement in the expected direction. Almost invariably they reported spontaneously if they "perceived" the movement in any other direction.

The criteria adopted for the tabulation of results were as follows The area around a point of origin was divided into four quadrants. If the diagram drawn by the subject fell within the quadrant of the suggested direction, this result was considered positive (showing the effect of the suggestion). The following diagram will make the point clear:

sherif_1935a_33.gif

If right suggestion was given the diagram or diagrams falling within the quadrant at the right were considered positive ; that is, showing the influence of suggestion. Diagrams falling outside that quadrant were considered negative. The opposite 90Ί degree quadrant was used as the criterion for the left suggestion ; diagrams falling within that 90Ί degree area were considered positive. Knowing the fact that the light may move in any direction, this three-to-one criterion is a fair one.

The first part of the experiment was carried on in 1932 in the psychological laboratory of the Gazi Teachers College, Ankara, Tur-


( 45) -key. There were 20 subjects, male students in the Institute. Each subject was used for only one session. Ten subjects were given right directions and 10 subjects left. The results are given in Table XIII.


TABLE XIII DIRECTION RESULTS IN MOVEMENT EXPERIMENT
Subject Suggested Direction + –
1 Right 45 5
2 “ 41 9
3 “ 31 19
4 “ 50 0
5 “ 43 7
6 “ 7 43
7 “ 39 11
8 “ 47 3
9 “ 50 0
10 “ 42 8
11 Left 49 1
12 “ 48 2
13 “ 35 15
14 “ 50 0
15 “ 50 0
16 “ 50 0
17 “ 50 0
18 “ 10 40
19 “ 50 0
20 “ 50 0

Out of 1,000 judgments, 837 were reported in the direction suggested and 163 outside of the suggested direction.

The second part of the experiment was carried on at Harvard Psychological Laboratory in 1933. The set-up, procedure and instructions were essentially the same. There were originally 8 subjects. Each subject was used in two separate sessions, at least one week apart, to do away somewhat with the effect of the previous direction. If right direction was suggested in the first session, left direction was suggested in the second session. One subject, a graduate student, could not come for the second session. Therefore his results are discarded. The results are presented in Table XIV.

The third part of the experiment was conducted at the Columbia Psychological Laboratory in the summer of 1934. Set-up, procedure, instructions and the arrangement of sessions were essentially the same as in the Harvard experiments. Ten subjects were used, two sessions with each, with one right and one left direction as indicated in the table. The results are given in Table XV.


( 46)

TABLE XIV
Subject Session I Session 2
Suggested Direction + – Suggested Direction + –
1 L 40 10 R 48 2
2 R 50 0 L 50 0
3 R 45 5 L 32 18
4 L 50 0 R 50 0
5 R 47 3 L 30 20
6 L 50 0 R 50 0
7 R 50 0 L 0 50
    332 18   260 90

The sum of the positive and negative results (as defined above) was taken. The data were not subjected to further statistical treatment, because each judgment was not given separately. At the end of the experiment the subject made a general statement about the number of times the light moved in each direction.

TABLE XV
Subject Session I Session 2
Suggested Direction + – Suggested Direction + –
1 L 12 38 R 50 0
2 R 49 1 L 50 0
3 L 50 0 R 9 41
4 L 50 0 R 0 50
5 R 50 0 L 50 0
6 L 49 1 R 48 2
7 L 50 0 R 50 0
8 R 47 3 L 0 50
9 R 50 0 L 1 49
19 R 0 50 L 37 13
    407 93   295  205

On the basis of these results one may conclude

In general, suggestion is effective in giving definite direction to an indefinite, unstable stimulus situation. The number of negative judgments in the second sessions, when a direction opposite to the first is suggested, is definitely larger than in the first sessions. This may be due to the influence of the first session. When some subjects perceive direction once in a definite way as determined by suggestion, they keep on perceiving the movement in that direction in spite of the change of the instructions in the new session, thus showing perseveration.

Notes

No notes

Valid HTML 4.01 Strict Valid CSS2