Obedience to Authority: Milgram (1963)

Aims:

To determine how people respond to a legitimate authority figure who asks that they act against a third individual. Prompting this was the attempt to verify the "Germans are different" hypothesis which suggested that the behaviour of the Germans towards the Jews during WW2 showed that they were uniquely predisposed to follow orders.

Procedure:

A participant is invited to help in a study of memory at the psychology department of a prestigious university. On attending, he is told that he is a teacher, and as such his task is to teach a learner a list of word pairs, and to administer punishment whenever the learner makes a mistake. Punishment takes the form of an electric shock, with the level of shock increasing each time the learner makes a mistake. To administer each shock, the learner uses a machine on which a series of switches is each labelled with a voltage rating, with the ratings starting at 15 volts and rising in increments of 15 volts to a level of 450 volts. Additionally, each group of four switches is labelled with text such as Slight Shock, Moderate Shock, up to Danger: Severe Shock. The final two switches are marked XXX.

Present in the room with the teacher is an experimenter, dressed in a lab coat, who ensures that the teacher continues with the experiment. When the teacher shows reluctance to apply punishment, the experimenter prompts him to continue, using statements such as:

Unknown to the participant, the learner is a confederate of the experimenter, and does not actually receive electric shocks. However, he provides feedback in the form of increasingly vocal objections to higher "shock" levels, demanding to be released from the experiment, and then refusing to answer questions (which the teacher is told must be treated as incorrect responses) before becoming ominously silent.

The experiment is repeated many times, under many different conditions, including:

  1. (Remote Feedback) teacher in a separate room from learner, but able to hear learner thumping on the wall
  2. (Voice Feedback) teacher in a separate room from learner, but able to hear learner's responses through a dividing wall
  3. (Proximity) teacher in the same room as the learner
  4. (Touch Proximity) teacher in the same room as learner, and required to hold the learner's hand on a metal plate to ensure he receives a shock
  5. (Alternate Venue) experiment carried out in a venue not associated with the university
  6. (Closeness of Authority) experimenter not present in the room with the teacher, but delivering "prompts" over a telephone
  7. (Women) experiment carried out with female "learners"
  8. (Peers rebels) two confederates of the examiner are described as other teachers, but show resistance to the examiners authority when told to apply shocks

In all cases, once the experiment has finished, the participant is interviewed and debriefed; the true nature of the experiment is disclosed to him a. He meets the "learner", and is reassured that no dangerous electric shocks were in fact applied. The participlant is assured that his behaviour is entirely normal and that any feelings of upset or tension were shared by other participants. All participants received a follow-up questionnaire which invited them to explain their feelings about the experiment.

Findings:

Contrary to what Milgram (and virtually all the people he consulted prior to the experiment) expected, the experiments showed a consistently high level of obedience. During the pilot studies for the experiment, in which no verbal protests were made by the "learner", virtually all subjects were happy to apply "shocks" up to the maximum level indicated on the screen.

During the experiments listed above, the corresponding rates of conformance (that is, the percentage of participants who carried on applying "shocks" up to the maximum level indicated) was found to be:

  1. (Remote Feedback) 65%
  2. (Voice Feedback) 62.5%
  3. (Proximity) 40%
  4. (Touch Proximity) 30%
  5. (Alternate Venue) 47.5%
  6. (Closeness of Authority) 20.5%
  7. (Women) 65%
  8. (Peers rebel) 10%

The figures shown above are only part of the story. As well as noting the rate of obedience, the debriefing session also allowed the participants to assess their level of stress during the study. While such ratings are subjective, they nevertheless showed that participants felt a high level of stress during the experiments. This tension was evident during the experiment, as it could be seen that "teachers" behaved in a way that showed they were experiencing stress, including sweating, trembling, and, in some cases, anxious laughter

Conclusion:

The findings indicate that the pressure to be obedient is be very strong, but their are some factors that consistently influenced obedience in this study:

Strengths:

  1. This experiment provides evidence that undermines the "Germans are Different" hypothesis
  2. Milgram repeated his experiment hundreds of times, and by making small variations in the scenario, and obtained a set of results which gives a very clear indication of the conditions that affect obedience rates.
  3. Although many of the participants found the experiment very stressful, their responses to the follow-up questionnaire indicated that overwhelmingly they were pleased to have been involved: 92% of participants returned questionnaires; of these, only 1.3% said they were "sorry" or "very sorry" to have been in the experiment. Additionally, Milgram received unsolicited letters from participants. For example:

Weaknesses:

Milgram's experiment was subject to criticism, but provided a very robust defence to such arguments. Namely:

  1. Ethical issues Diana Baumrind (1964) expressed concern for the welfare of participants, questioning whether adequate measures had been taken to protect them.

    The issue of "freedom to withdraw" is the one point where I think Milgram is on dodgy ground here. There were many cases in the experiment where participants were told "you must carry on", which violates this freedom.

    However, given the nature of the experiment, such a situation was unavoidable, and as Milgram says "the central moral justification for allowing my experiment is that it was judged acceptable by those who took part in it" ("The Individual in a Social World", Milgram 1977)

  2. Lack of experimental realism, Orne and Holland (1968) suggested that the results of Milgram's experiments could not be generalised, since participants behaved the way they did because they did not believe the "story" they had been told, and were instead trying to fulfil the experimenter's expectations.

    Milgram (1972) rebuts to this suggestion by saying:

    Additionally, Milgram's experiments do appear to have ecological validity, and their findings have been reproduced, for example Hofling (1966)

See class notes for 14-Nov.


Study sheets index