Human Behavior Experiments Opinions and Group Pressure

About a human behavior experiment to determine the effects group pressure has on individual opinions.

HUMAN BEHAVIOR EXPERIMENTS

IS SEEING BELIEVING?

Title of Experiment: Opinions and Social Pressure

Conducted by: Solomon E. Asch and associates, at several colleges and universities

Reported in: Scientific American, Vol. 193, No. 5 (November, 1955), 31-35

Object: To investigate the effects of group pressure on individual opinion. Can majority opinion lead someone to see what he does not see--or at least say that he does?

The Experiment: Experimenters played variations on the following theme. A small group (2 to 15) of college students were told that they were taking part in an experiment in visual judgment. The experimenter showed them two cards--one bearing a single vertical line, the other bearing three vertical lines of differing lengths. The students' job was to choose the line on the second card which was the same length as the single line on the first card. One by one, the students announced their answers. A second set of cards was produced, then a third, fourth, and so on for 18 trials.

Initially, everyone chose the same matching line: the right one. But on the third trial, one student (the subject of the experiment) found that his judgment differed from the others'. Again and again he heard the other students give erroneous answers. Sometimes he was the only dissenter. Sometimes he had the support of one other person. Sometimes that one other dissenter disagreed both with the subject and with the majority. Sometimes the other dissenter finally joined the majority or left the room.

What the student subject did not know was that all other students, including his fellow dissenter--if he had one--had been told how to respond. He alone was exercising "independent" judgment. Or was he?

Conclusions: If all students had responded honestly, they would have made mistakes only 1% of the time. The 123 student subjects of this experiment yielded to the majority's wrong judgment 36.8% of the time. Although about a fourth of the subjects persisted in their independent answers for trial after trial, others were high conformers; once they gave in to majority opinion, they did not, as a rule, regain their independence. In post experiment interviews, the conformists all tended to underestimate the degree to which they had conformed.

Variation in the size of the group made a difference, up to a point. A single opponent had little effect on an individual's answers. With two opponents, subjects in the minority accepted the wrong answer 13.6% of the time. With three opponents, subjects' errors rose to 31.8%.

A dissenter with an ally was less likely to give in than a dissenter who stood alone. Subjects who had the support of one other person answered incorrectly only one fourth as often as those without any support. The number of incorrect answers dropped even when the fellow dissenter, like the majority, was giving a wrong answer--a different one. Asch concluded that "dissent per se increased independence and moderated the errors that occurred." If the subject's fellow dissenter joined the majority, the subject also tended to submit. But if instead the fellow dissenter left the room, the subject did a better job of holding on to his independence.

Throughout these trials, the subject's ability to stick with the right answer depended in part on how wrong the majority was. Yet "even when the difference between the lines was 7 in.," Asch said, "there were still some who yielded to the errors of the majority." However, Asch cautioned against drawing too pessimistic conclusions from the experiment. After all, he said, "those who participated in this challenging experiment agreed nearly without exception that independence was preferable to conformity."

You Are Here: Trivia-Library Home » Human Behavior Experiments » Human Behavior Experiments Opinions and Group Pressure
« Human Behavior Experiments Street Corner CrossingHuman Behavior Experiments Black Panthers and the Police »