Why is the Milgram experiment banned



Political attitudes and prejudices




Group behavior


Cultural differences


The Greek word "êthos", on which the term ethics is based, has three meanings in its original form: (1) It means the usual place of life, (2) the habits that are lived in this place and (3) the way of thinking and the way of thinking (Character) that underlie these habits. This of course results in a great many areas of application of ethics. The Latin equivalent of "êthos" is "mores". "Creative" translation into German resulted in new meanings for these two words, so that Höffe (1999, p. 18) suggests the following definition: "Under "ethics" one understands a - partly more, partly less elaborated - doctrine of morals and manners, the moral philosophy, but under "morals" and "manners" their subject matter.“ 

The questions that concern us in ethics are (1) "What should I do?" (Rules and norms that are unreservedly valid, as well as obligations), (2) "Who do I want to be and how do I want to live?" (What makes a happy and happy existence?), (3) "In which social and political environment, in which (social) institutions do we want and should we live?"

The history of ethics can first be roughly divided into pre-philosophical ethics and philosophical ethics. The pre-philosophical ethics is formulated in wisdom, religious or divine commandments or in rules of prudence (see e.g. the ten commandments, ancient Egyptian wisdom). They are used to cope with everyday life. Especially for adolescents, these wisdoms give advice on how to manage a life successfully and without unnecessary difficulties.

Philosophical ethics is used to cope with critical situations and crises. It sees itself as a practical philosophy. Methodically, without referring to political, religious or historical authorities, attempts are made to make generally valid statements. These statements should be abstracted from the cultural background. There is a risk that one's own culture will be used as a yardstick, which can lead to ethnocentrism. A special form of ethnocentrism is Eurocentrism: Here other cultures that do not correspond to the European-American culture are devalued as primitive, barbaric or "underdeveloped". In Christian ethics, which adopt the four cardinal virtues of bravery, prudence, justice and wisdom from Greco-Roman antiquity, the theological virtues faith, hope and love are included as important components in the "conduct of life". Another focus is freedom. In Kant, the principle of freedom takes the place of the principle of happiness. The action is now carried out according to laws. The highest principle consists in the will, which gives itself the laws (autonomy: self-legislation).

The philosophical theories on ethics can be further divided into teleological, deontological and mixed theories.

Teleological theories judge actions only by the value of their consequences. Here, the individual theories must already define clear value standards. Most theories take pleasure or pleasure as a measure (Patry, 2002). The theories also distinguish for whom the consequences of the actions are supposed to be good. Consequences can be good (a) for the doer (selfish), (b) for someone else (altruistic), or (c) for the community (utilitarian).

Deontological theories judge actions on the basis of absolute and general principles that apply regardless of the consequences. An important representative of this direction is Immanuel Kant. The supreme moral principle according to Kant is:


Act only according to the maxim that you can at the same time want to become a general law.“


This categorical imperative applies "without any restriction [...], objective, general and necessary" (Höffe, 1997, p.150 f). It does not state what is good for a particular purpose, but rather what is good in and of itself. According to Kant, there can only be one categorical imperative. The practical imperative is derived from the categorical imperative:


Act in such a way that you use humanity both in your person and in the person of everyone else at the same time as an end, never just as a means.


In the context of psychological practice and research, three moral principles should be applied that can be derived from teleological and deontological theories (Patry, 2002). A moral principle is understood to be a final moral principle that cannot be derived from a general norm (Höffe, 1999). These three moral principles are (1) autonomy, (2) non-harm and benefit, and (3) justice.




Autonomy means self-determination. Kant postulates that free will acts reasonably according to its own maxims and laws. Without the autonomy of the will, all actions can at most be done according to duty, but not out of duty. John Stuart Mill (quoted from Patry, 2002, p. 36 f) means in this context: "If we are allowed to develop our own lives, that alone is beneficial. These benefits are further increased when we feel that our actions are truly ours". A society in which autonomy is respected is a happier society, "as their citizens have the opportunity to develop their skills to act as rational, responsible moral agents". However, if respecting autonomy harms the common good, personal autonomy can also be ignored.

In the modern approaches to ethics, three concepts of autonomy are distinguished. The first definition of autonomy emphasizes the Freedom of action. An act must be done voluntarily. No external coercion may therefore be exerted on the agent. It must also be done on purpose; H. the agent only wants to carry out this action and no other. A second definition of autonomy deals with the freedom of choice. A person can always choose freely if he has all the information about the various decision alternatives. If this information is not available, one cannot speak of autonomous decisions. The third definition of autonomy refers to the purposeful consideration pointed out before an action. An individual is autonomous if he has the ability to make rational and informal decisions and to act on them (Patry, 2002). There are several prerequisites for this ability. (1) The individual must be able to formulate appropriate goals, (2) prioritize the goals, (3) determine the best means of achieving those selected goals, (4) act effectively, and (5 ) abandon or change selected goals if the consequences of the means used are undesirable or the means themselves are inadequate.

When is it ethically justifiable to restrict a person's autonomy? For this purpose, Mappes and Zembaty (1991) give several principles that limit freedom.

1.        Protection principle. A person's freedom may be restricted if they are prevented from harming other people.

2.        Non-triggering principle. A person's freedom may be restricted if this prevents them from causing shame, embarrassment or discomfort in others.

3.        Paternalism principle. A person's freedom may be restricted if this prevents them from harming themselves.

4.        Principle of extreme paternalism. A person's freedom may be restricted if it is thereby used.

5.       Principle of morality. A person's freedom may be restricted if it is prevented from doing wrong ethically.

6.        Welfare principle. A person's freedom may be restricted if one thereby benefits others.


The principle of autonomy also applies to psychological research. Here it is already important that the test subjects have all the relevant information before the experiment or the investigation, understand it and, on the basis of this information, can consciously and voluntarily decide whether they want to participate in the study or not. The information must be explained clearly and simply (if possible without technical jargon). The decision to participate in the experiment should be made explicit. Interesting results emerged in a study by Brody, Gluck, and Aragon (1997) who asked psychology students about their experiences with experiments. For 20% of the subjects, the information about the experiment was too vague. 15% of the students believed that by simply showing up, they had given their consent to the experiment. 15% said they did not want to take part in the experiments. But they were embarrassed to openly express their refusal to participate in front of the group.

The protection of privacy should also apply to psychological experiments. Not every behavior of the test person in psychological examinations is relevant to the questions. The taxonomy of behavior by Sieber, Iannuzzo and Rodriguez (1995) can be used for the ethical discussion. Five categories can be identified here.

  1. Behaviors that are public and performed for an audience.
  2. Behaviors that take place in public but are not intended to be observed by others.
  3. Behaviors that take place in a place with limited access and are only intended for those involved (e.g. doctor's office).
  4. Behaviors that take place in a private place and are only intended for those involved, but which are not secret (e.g. in your own apartment).
  5. Behaviors that take place in a private location, are only intended for those involved and are secret.

The less public the behavior observed in the experiment, the greater the invasion of privacy if it is observed nonetheless. In any case, however, the test person should be informed about the behavior that is being recorded. If the subject does not consent to the use of their data, they must be deleted from the data record.

Furthermore, the confidentiality of the data must be ensured. The subject must give their consent for their information to be passed on. This transfer can take place in different ways (Patry, 2002, p. 63 f):


  • A recording of the test subject appears in a documentation video or TV broadcast about the test.
  • The test director discusses the test subject's results with colleagues, students or with his family.
  • Verbatim quotation of test subjects' statements in texts.
  • The experimenter announces the results in other ways (by means of a notice, etc.).
  • Parents or relatives will be informed of the results.


Non-harm to the test subject


Psychological research should not be harmful. Experiments can lead to behavior that is damaging to self-esteem in the short term or people can be encouraged to perform behavior that is discrepant with attitudes. In most experiments, these fluctuations and manipulations are cleared up in a follow-up discussion and the test subjects are compensated for the injustice they suffered. In some experiments, however, very strong manipulations are carried out or test subjects are exposed to very extreme stress or conflict situations. Three studies that do not meet modern ethical standards should now be briefly presented and critically explained. The first study is by Watson and Rayner (1920) on the classical conditioning of feelings (phobias), the second by Milgram (1963, 2001) for researching the belief in authority and the third by Haney, Banks & Zimbardo (1973) for the behavior of naive test subjects in a prison simulation. The experiments mentioned triggered a large number of ethical discussions when they were published, and they were taken as an opportunity to revise the ethical guidelines for psychological research.



Watson & Rayner's Experiment on Anxiety Conditioning (1920)


Watson and Rayner (1920) tested the emotional conditioning hypothesis on a 9 month old boy ("little Albert"). The boy was healthy from birth and showed no emotionally suspicious reactions. He was the son of a nurse who worked in the hospital where this experiment was being carried out, and he had spent most of the time in that hospital.

The experiment began by testing his emotional response to neutral stimuli. First he was presented with a white rat, a rabbit, a dog, a monkey, a ball of wool, etc. He showed no negative reactions to these stimuli. Nobody had seen him afraid until then.

The authors used a very unpleasant, loud noise as the aversive stimulus. They hit a large piece of metal with a hammer. First, a member of staff caught little Albert's attention, so he looked in his direction. Another adult worker was standing behind Albert who made the loud noise. After the third sound, Albert began to cry. This was the first time during the experiment that the boy started crying and showed signs of fear.

After a break of a few weeks, the actual conditioning started at the age of 11 months and 3 days. Little Albert was shown the rat and he was about to touch the rat when the loud noise was made behind him. Albert was startled and fell forward, but he did not cry. Now Albert grabbed the rat again and the noise came again. Albert fell over again and began to whimper. In order not to disturb the child further, the conditioning was interrupted for a week.

A week later, Albert was presented with the rat again. He watched her for a long time from a distance, but showed no sign that he wanted to approach her. Now the rat was pushed closer to him and Albert held out his left hand. Just before he could touch or stroke her, he jerked his hand back again. After this rejection he was allowed to play with building blocks for a while. Then the rat was presented to him again three times in a row and as soon as he approached the rat, the sound was triggered. Albert fell over again and turned away from the rat. In a further round, the rat was presented to him without making a sound. Albert then began to cry. Then the rat was presented to Albert twice in a row and the sound made. He started crying and turned away. One last time the rat was presented to him alone and the boy immediately fled.

Five days later, Albert was tested again. At first he played with his building blocks and he did not show any negative reactions to the laboratory or its staff. When he was shown the rat, he immediately started whimpering and turned away from the rat. Thereupon the building blocks were given to him again to play with, with which he happily occupied himself. After a while he was shown the rat again and fled from it.

The next step was to check whether his fear had also generalized to similar objects, or whether he had only learned a rat phobia through conditioning. The experimenters showed Albert a hare. He immediately began to whimper and when he should touch the animal he put his head on the mat and fled from the animal. After this situation he was allowed to play with his building blocks again and he calmed down again.Next a dog approached, this initially did not trigger any negative emotional reactions in Albert, only when the snout almost touched his head did he start to cry. Now he was allowed to play with his building blocks again. When he was shown a fur coat, Albert started crying again and tried to crawl away. The ball of wool, to which he had not reacted at the beginning of the experiment, also triggered negative emotional reactions in him. Thus there was a strong generalization of emotional behavior.

Five days later, his emotional response was tested again and he behaved the same way. This time the unpleasant noise was triggered once during the presentation of the dog. Until then, the experiment was carried out in the laboratory. Now a series of tests has also been started outdoors. Even in this new situation, Albert showed the negative emotional behavior. The unpleasant noise was triggered in the rat in one pass.

After a break of a few weeks, a final test was carried out without reinforcement. In all situations (rat, dog, fur coat and Santa Claus mask) Albert started to cry.

Albert could not be relieved of his fear because his mother no longer took him to the hospital and it was therefore no longer possible for the experimenters to carry out an extinction of the fear. "Unfortunately Albert was taken from the hospital the day the above tests were made. Hence the opportunity of building up an experimental technique by means of which we could remove the conditioned emotional responses was denied us. Our own view, expressed above, which is possibly not very well grounded, is that these responses in the home environment are likely to persist indefinitely, unless an accidential method for removing them is hit upon.”(Watson & Rayner, 1920, p. 12).



Milgram's (1963, 2001) experiments on obedience


In his experiments, Milgram wanted to test the framework conditions for obedience to authorities. Shortly after the Second World War or after the end of the German Reich, this research question became the focus of research interest, as a large number of previously innocent citizens participated in the torture and murder of innocent people and took an active role in the execution of these atrocities.

To test the hypotheses, a relatively simple experiment was designed, which ostensibly should serve a completely different purpose in the eyes of the test subjects. Test subjects for a learning experiment were searched for by means of a newspaper advertisement. All participants were from the city of New Haven and were between 20 and 50 years old. No students were admitted as test subjects for this experiment because Milgram was of the opinion that they could see through the experiment and thus no valid results could be obtained. For the effort, the test subjects were promised $ 4 an hour and a travel cost reimbursement of 50 cents.

The original experiment took place in an elegant laboratory at Yale University. It consisted of two separate rooms. The student (one of the experimenter's staff) should sit in one room and the test subject (who was to play the teacher) and the experimenter in the other room.

At the beginning of the experiment, two people entered the laboratory. One person was an insider who pretended to be a test person, the other person was an actually unsuspecting test person (in the first experiment only male test persons were tested). They were welcomed by the experimenter and informed about the supposed content of the experiment: The aim was to investigate the effects of punishment on learning. For this experiment, one subject must be the teacher and the other the student. For this purpose, the distribution of tasks was raffled. First, the actual test subject had to choose between two slips of paper containing the distribution of roles (both slips said “teacher”, but the actual test person was not aware of this). The second "test person" (the employee of the Vl) then had to play the pupil and take a seat on a wooden chair in the study room. Both hands were fixed to the backrests and connected to an electrode (the student should receive electric shocks as a punishment for mistakes). Furthermore, an electrode ointment was applied to the wrist “to avoid blisters and burns”.

The pupil had to learn pairs of associations (e.g. blue - box, nice - day, wild - bird, etc.). In the test sequence, the student was given a word and had to choose the correct answer from four alternative answers (e.g. blue: (1) sky, (2) ink, (3) box, (4) lamp). He had to announce the answer by pressing a button. The answer was transmitted electronically to the next room where the teacher was sitting and displayed on the shock generator.

The shock generator consisted of a control panel with 30 toggle switches in a horizontal arrangement. Each toggle switch was labeled with a volt, with a scale ranging from 15 to 450 volts. In addition, a verbal description was attached for every four toggle switches (light shock, moderate shock, severe shock, severe shock, very severe shock, danger: threatening shock. Two switches were only marked "XXX".) When a toggle switch was operated, it lit up A red control light came on above him and an electrical hum could be heard. Then an electric blue light flashed very briefly and the pointer of the voltmeter swung to the right.

In order to underline the “authenticity” of the apparatus, the actual test person was given a 45-volt electric shock in a test run.

A shock had to be preceded by certain actions.

  1. An electric shock was only given to incorrect answers.
  2. The amperage was increased for each wrong answer.
  3. Before administering the shock, the test subject had to announce the current strength loudly and clearly.
  4. When it reached the highest voltage level, it should continue with this maximum level (the attempt was aborted after two further applications of the maximum level).

The test subjects kept turning to the test director and asking him for feedback on their behavior or on how to proceed. The experimenter made a series of "encouraging" remarks that were hierarchical. If the incentive of the first stage was not enough to encourage the subject to continue, encouraging remarks of the second stage were used. There were four levels in total:

Step 1: "Please continue!" Or "Please continue!"

Stage 2: "The experiment requires you to continue!"

Level 3: "You absolutely have to continue!"

Level 4: “You have no choice, you have to keep going!"

In addition, the test subjects heard the reactions of the "student".


The victim's responses were a key element of the situation and therefore had to be carefully worked out. In preliminary investigations, no acoustic feedback from the victim was initially used. We believed that the word and volt readings on the generator would create enough pressure to reduce the subject's obedience. However, this turned out to be wrong. Since there was no protest from the student, almost every subject happily continued to the end of the volt scale as soon as they were told to, without apparently being impressed by the word names."(Milgram, 2001, p. 39).


In the main experiment, the student's screams were played on tape to create a standardized situation:


  • Up to 75 volts, the student gave no response from discomfort. At 75, 90 and 105 volts it gave a slight growl.
  • At 120 volts, the student shouted that the electric shocks were now uncomfortable.
  • A painful groan was audible at 135 volts.
  • At 150 volts, the student shouted, “Experimenter, get me out of here! I don't want to take part in this experiment any longer! I refuse to continue! ”.
  • At 180 volts, the student shouted, "I can't take the pain!"
  • At 270 volts the answer was a definite roar.
  • At 300 volts, he yelled that he would not give an answer to the memory test.
  • At 315 volts, he again stated that he would no longer take part in this experiment.
  • After 330 volts no one heard from him and he gave no more answers.


After the experiment was over (either voluntarily by the test subject (the teacher) or by abortion by the test director), a detailed discussion took place between test subject, test director and pupil.


“As a minimum, it was made clear to each subject that the victim had not received dangerous electric shocks. Each test subject had the opportunity for a friendly reconciliation with the unharmed victim and for an extended conversation with the test director. The experiment was given to disobedient subjects [i.e. H. those who voluntarily no longer wanted to give stronger electric shocks, Note d. A.] explained in a way that positively underscored their resolve to disobey the experimenter. Obedient test subjects were assured that their behavior was normal and that other participants had the same feelings of conflict or tension as they did. The subjects were told that they would receive a full report when the test series was completed. In some cases, even more extensive and detailed discussions about the experiment were held with individual test subjects"(Milgram, 2001, p. 41).


The subjects' willingness to obey differed greatly from the test conditions to which they were exposed and differed significantly from the prognoses of uninvolved persons who were asked before the experiment what maximum shock level they would apply. Table 1 gives an initial overview of the results. Two psychologists and three adults stated in the prognostic study that they would not give a single electric shock. The highest shock that would be delivered is 300 volts. Most of the respondents assumed that the test subjects would not administer more than 150 volts, since at this point the student clearly and loudly asks to be allowed to stop the experiment. These statements are based on several assumptions. First, it is assumed that the subjects are decent and well-behaved people who do not want to hurt anyone. Secondly, most of the respondents thought that the cause of the action was to be found in the agent himself, and that he was in control of the termination at all times. At each point in time he weighs up how much benefits and costs arise for him and also for the student as a result of his behavior. Only if the person were forced to do so under intense pressure or physical violence would they cause harm to someone else.


Actual behavior painted a very different picture: no one refused to give an electric shock. The first termination took place at 105 volts in the condition in which the student sat in the same room as the teacher (3rd test condition). More than half of the test subjects in the first two test conditions applied the maximum penalty. In both experimental conditions, the student sat in a different room. In the first test condition, only a loud knock was heard; in the second test condition, the student could be heard screaming. In the third condition, 40% of the test subjects applied the maximum stimulus. In the fourth condition, the student received a shock only when his hand was on a shock plate. At 150 volts, the student asked to be allowed to stop the experiment and refused to

Hand on the shock plate. The subject was then ordered by the experimenter to force the student's hand on the shock plate. After all, 30% of the test subjects in the fourth test condition used the maximum shock.




Table 1. Forecasts and actual time of cancellation

Shock level

Label and voltage


actual behavior




Adult middle class

free space

Acoustic feedback






Slight shock








































Moderate shock