Laboratory Experiments in sociology

Last Updated on October 8, 2026 by Karl Thompson

Laboratory experiments take place in an artificial, controlled environment set up by the researcher. This post covers:

1. What a laboratory experiment is, and how it differs from a field experiment
2. Why sociologists hardly ever use them
3. Four laboratory experiments you can use as examples in sociology: Milgram, the Bobo doll study, the Stanford Prison Experiment and Chaikin et al.’s teacher expectations study
4. The theoretical strengths and limitations of the method
5. The practical strengths and limitations
6. The ethical strengths and limitations
7. How laboratory experiments come up in the AQA exams

Laboratory experiments are part of the research methods component of A-level sociology, examined mainly on Paper 3 and also on Paper 1.

What is a laboratory experiment?

An experiment measures the effect of an independent variable (the cause) on a dependent variable (the effect). The researcher changes one thing, keeps everything else the same, and measures what happens. If you need the key terms first — hypothesis, independent and dependent variables, extraneous variables, control group — see Experiments in Sociology: An Introduction.

A laboratory experiment does this in a setting the researcher has built and controls completely. It does not have to be a room full of test tubes. A laboratory is any artificial environment where the researcher decides who is there, what happens to them, and in what order. Milgram’s laboratory was a room at Yale University; Zimbardo’s was a mock prison in a university basement.

Laboratory experiments are the main method of the natural sciences, and they are common in psychology. A psychologist measuring the effect of alcohol on reaction time, for example, can bring participants into the lab, give some of them alcohol and some a placebo, make sure they all eat and do the same things in between, and then test their reaction times. Because everything except the alcohol was held constant, any difference in reaction times can be put down to the alcohol.

That is the whole appeal of the method: control. And control is exactly what you give up in a field experiment, which takes place in a real social setting such as a school or a workplace. The field experiments post has a table comparing the two side by side. The short version is that a laboratory buys control by giving up realism, and a field experiment does the opposite.

Why sociologists rarely use laboratory experiments

Sociologists hardly ever use laboratory experiments, for three reasons that are worth knowing separately because each one is a different kind of criticism.

Most of what sociologists study cannot be put in a laboratory. Social class, a subculture, a school, a labour market or a religion cannot be brought into a room. A laboratory can only hold a handful of individuals at a time.

Most of the variables sociologists care about cannot be manipulated. The logic of an experiment is that the researcher changes the independent variable. You cannot assign someone a social class, an ethnicity or a gender at random. Those are the variables sociology is mostly about.

The results may not apply outside the lab. People in an artificial setting, who know they are being studied, may behave quite differently from how they would in real life. This is the problem of external validity, discussed below.

HOWEVER, laboratory experiments still matter for A-level sociology. They are on the specification, the AQA does set questions on experiments, and several of the most famous studies in the social sciences are laboratory experiments that sociologists use all the time — usually as examples of what can go wrong.

Four laboratory experiments you can use in sociology

A named example is worth far more in an exam than a general statement. These four between them cover every strength and limitation in the rest of this post.

Milgram’s obedience experiment (1963)

Stanley Milgram recruited 40 men aged 20 to 50 from the New Haven area, paid them $4.50 to take part in what they were told was a study of memory and learning, and asked them to give increasingly strong electric shocks to a “learner” every time he got an answer wrong. The learner was an actor and the shocks were fake. 26 of the 40 (65%) went all the way to the final switch, marked 450 volts, and every one of them continued to at least 300 volts.

Milgram concluded that ordinary people will go a very long way in obeying an authority figure they see as legitimate. It is a sociological finding as much as a psychological one, because it is about the power of a social role rather than the personality of the person in it.

Why it matters for this post: it is the classic example of a laboratory experiment with high reliability, serious ethical problems and doubtful external validity. See my full post on Milgram’s obedience experiment.

Bandura, Ross and Ross’s Bobo doll experiment (1961)

Albert Bandura and his colleagues studied 72 children, 36 boys and 36 girls, aged roughly three to six, from the Stanford University nursery school. Some watched an adult attack an inflatable “Bobo doll”, hitting it with a mallet and throwing it around; some watched an adult play quietly; a control group saw no adult at all. Each child was then mildly frustrated and left in a room with toys, including the doll, while researchers watched through a one-way mirror.

Children who had seen the aggressive adult were more likely to imitate the specific aggressive acts they had seen. Bandura took this as evidence for social learning theory: children learn behaviour, including aggression, by watching others.

Why it matters for this post: it shows the method’s strength in establishing cause and effect very clearly, and its weakness in external validity just as clearly — a doll designed to be hit is not a person. It is also relevant to media effects and to crime. See my full post on the Bobo doll experiment.

Zimbardo’s Stanford Prison Experiment (1971)

Philip Zimbardo randomly assigned 24 male students to be prisoners or guards in a mock prison in the basement of Stanford University’s psychology building. The study was planned to last two weeks and was stopped after six days because of the guards’ abuse and the prisoners’ distress.

It is a laboratory experiment — more precisely a simulation — not a field experiment, however often it turns up on lists of field experiments. The researchers built the setting and controlled it entirely.

Why it matters for this post: it is now the standard example of how a laboratory experiment can go wrong. Later analysis of the study’s own archives (Le Texier, 2019) found that the guards were coached on how to behave, which means the study may show participants doing what the researcher wanted rather than anything about prisons. See my full post on the Stanford Prison Experiment.

Chaikin, Sigler and Derlega’s teacher expectations experiment (1974)

This is the one to use if a question is about education. Chaikin and colleagues had student teachers give a short lesson to a boy whom some had been told was bright, some had been told was not, and some had been told nothing. The lessons were videoed and the teachers’ behaviour analysed.

Teachers who expected a bright pupil behaved differently towards him — more eye contact and more encouraging body language — than teachers who expected a weak one.

Why it matters for this post: it shows what a laboratory can do that a field experiment cannot. Rosenthal and Jacobson claimed teacher expectations were passed on through classroom interaction but never observed a classroom. A laboratory let Chaikin and colleagues watch the interaction directly. HOWEVER, one lesson with one pupil in an artificial setting is a long way from a year in a real classroom, so it supports labelling theory without proving that the same thing happens in schools.

The comparison between Chaikin et al. and Rosenthal and Jacobson is a neat way to show the trade-off between the two types of experiment in a single paragraph.

Theoretical strengths of laboratory experiments

Control over variables, and therefore cause and effect. This is the method’s defining strength. Because the researcher controls the setting, extraneous variables can be held constant, and any change in the dependent variable can be put down to the independent variable. Bandura could show that the children who saw aggression became more aggressive because that was the only thing that differed between the groups. No survey or observation can show cause this cleanly.

Precise, quantitative measurement. Lab experiments produce numbers — volts reached, aggressive acts counted, reaction times recorded. That allows exact comparisons between groups and between studies.

High reliability. The procedure is standardised, so another researcher can repeat it exactly. Milgram’s study has been repeated many times. The most important repeat is Burger (2009), who ran a modified version with 70 adults, 29 men and 41 women. Burger stopped the procedure at 150 volts for ethical reasons; 70% of his participants were willing to go beyond that point, against 82.5% in the comparable Milgram condition, a difference Burger found was not statistically significant.

Detachment and objectivity. The researcher follows a fixed script and keeps interaction with participants to a minimum, which leaves little room for the researcher’s own values to shape the result.

This is why Positivists favour them. Positivists want sociology to be scientific — to establish cause and effect through objective, quantitative, replicable methods. The laboratory experiment is the method of the natural sciences, and it ticks every box on that list.

Theoretical limitations of laboratory experiments

Low external validity. This is the main criticism, and the one sociologists care about most. The setting is so artificial that the results may tell us little about how people behave in real life. When were you last asked by a scientist in a lab coat to give electric shocks to a stranger? Authority in real life comes from teachers, police officers and managers, in settings people understand, and people may respond to it quite differently. Ecological validity is the term for whether findings apply to real-world settings; Bandura’s doll, built to be knocked over, is the textbook case of low ecological validity.

Demand characteristics. Participants know they are in an experiment, and they try to work out what it is about. Orne (1962) argued that they then tend to behave in the way they think the researcher wants. Zimbardo’s guards are the clearest example: some later said they were acting the part they thought was expected.

The Hawthorne effect. Simply knowing you are being studied can change behaviour, whatever the independent variable is. The term comes from the Hawthorne factory studies, and it applies with even more force in a laboratory, where nobody can forget they are part of a study.

Small, unrepresentative samples. Milgram’s 40 participants were all men from one American town. Bandura’s 72 children came from a single university nursery. Zimbardo’s 24 were male students who answered an advert for a study of prison life. Generalising from samples like these to “people” or “children” is a big leap.

Interpretivists object to the whole approach. Interpretivists argue that people are not objects reacting to stimuli. They act on the meanings they give to situations, and an experiment measures what people do without finding out why. Milgram’s numbers tell you how far people went; they do not tell you what the participants thought they were doing. This is the reductionism criticism — reducing complex human action to a simple cause-and-effect relationship.

Practical strengths of laboratory experiments

Once set up, they are relatively easy to run. Participants come to the researcher, the procedure is the same every time, and there is no need to negotiate access to a school or a workplace. A laboratory has no gatekeeper.

They are attractive to funders. Their scientific, quantitative nature carries a prestige that can make funding easier to obtain than for small-scale qualitative research.

Results can be produced quickly. A field experiment often has to run for months or a whole school year. A laboratory session can be over in an hour.

Practical limitations of laboratory experiments

You cannot get society into a laboratory. Large groups, communities, institutions and social processes that unfold over years are outside the reach of the method. This is the main reason sociologists use field experiments instead when they use experiments at all.

Many sociological variables cannot be manipulated. Class, gender, ethnicity, religion and age cannot be assigned at random, which rules out an experiment on most of the questions sociologists ask.

Large samples take a long time. Participants are usually tested one at a time or in small groups. Building up a large sample means running the experiment again and again.

Ethical strengths of laboratory experiments

Participants usually know research is taking place, even when they are not told its real purpose. That is more than can be said for many field experiments, where people often have no idea they are being studied at all.

They can be debriefed. Because the researcher knows exactly who took part, everyone can be told afterwards what the study was really about and offered support. You cannot do that with passers-by in a field experiment.

They can withdraw. In a modern laboratory study, participants are told they can stop at any point.

The findings may justify the costs. Milgram argued that what his study revealed about obedience was important enough to justify the distress it caused. Many of his participants agreed with him afterwards. Whether that argument works is itself a good evaluation point.

Ethical limitations of laboratory experiments

Deception and lack of informed consent. If participants know what is being tested, demand characteristics and the Hawthorne effect undermine the result, so researchers often deceive them about the real purpose. Milgram’s participants believed they were taking part in a study of learning and that the shocks were real. They could not give informed consent to something they had been misled about.

Harm to participants. Milgram himself described participants sweating, trembling, stuttering and digging their fingernails into their flesh, and three had seizures. Baumrind (1964) criticised the study for exactly this, arguing that the distress was not justified. Zimbardo released five prisoners early because of their distress.

Research on children. The Bobo doll study deliberately frustrated young children and exposed them to aggressive behaviour. It is hard to see it passing an ethics committee today.

Most of the classics could not be run now. Ethical review is far stricter than in the 1960s and 70s; the British Sociological Association’s statement of ethical practice sets out the current expectations. Burger’s 2009 Milgram study shows the consequence: to get ethical approval he had to stop at 150 volts, screen out people likely to be badly affected, and remind participants repeatedly that they could withdraw. That is a good point to make in an exam, because it means the famous findings can never be fully repeated.

Laboratory experiments in the AQA exams

Experiments are examinable as part of Theory and Methods, which appears mainly on Paper 3 and also on Paper 1. Questions about the method itself ask for its strengths and limitations, or how theoretical perspectives view it. The reliable structure is PET: practical, ethical and theoretical, a point in each, each supported by a named example.

“Laboratory experiments lack validity” earns very little. “Bandura’s children were reacting to a doll designed to be hit, in a room set up by the researchers, so we cannot assume they would behave the same way towards other children in a playground” earns a good deal more. Every point in this post has one of the four examples attached for exactly that reason.

Experiments also come up in the Methods in Context question on Paper 1, which asks you to apply a method to a specific issue in education. Field experiments are usually the better choice there, but if a question is about laboratory experiments, use Chaikin et al. and apply the points to pupils, teachers and schools specifically — the ethics of research on children, the artificiality of a one-to-one lesson in a lab, and the fact that a school’s real dynamics cannot be recreated. See my post on research methods in context: experiments and education.

One trap to avoid: do not offer the Stanford Prison Experiment as a field experiment. It is a laboratory experiment, and examiners know it.

Theory and Methods A Level Sociology Revision Bundle

Theory and Methods A-level sociology revision bundle cover

If you like this sort of thing, then you might like my Theory and Methods Revision Bundle – specifically designed to get students through the theory and methods sections of A-level sociology papers 1 and 3.

Contents include:

1. 74 pages of revision notes
2. 15 mind maps on various topics within theory and methods
3. Five theory and methods essays
4. ‘How to write methods in context essays’

Signposting and Related Posts

This post has been written primarily for students studying research methods as part of the AQA’s A-level sociology course.

Before this topic, most students cover Experiments in Sociology: An Introduction, which explains hypotheses, independent and dependent variables and the basic logic of experimental design.

The natural companion post is Field Experiments in Sociology, which covers the type of experiment sociologists actually use, with seven worked examples.

For the four examples in full, see Milgram’s Obedience Experiment, the Bobo doll experiment, the Stanford Prison Experiment and Research Methods in Context: Experiments and Education.

For condensed revision notes covering both types, see Experiments in Sociology – Revision Notes.

For everything else on this topic, see the main research methods page.

Sources: Bandura, A., Ross, D. and Ross, S. A. (1961) ‘Transmission of aggression through imitation of aggressive models’, Journal of Abnormal and Social Psychology. Baumrind, D. (1964) ‘Some thoughts on ethics of research: after reading Milgram’s “Behavioral Study of Obedience”‘, American Psychologist. Bryman, A. (2016) Social Research Methods, 5th edn, Oxford University Press. Burger, J. M. (2009) ‘Replicating Milgram: would people still obey today?’, American Psychologist. Chaikin, A. L., Sigler, E. and Derlega, V. J. (1974) ‘Nonverbal mediators of teacher expectancy effects’, Journal of Personality and Social Psychology, 30. Le Texier, T. (2019) ‘Debunking the Stanford Prison Experiment’, American Psychologist, 74(7). Milgram, S. (1963) ‘Behavioral study of obedience’, Journal of Abnormal and Social Psychology. Orne, M. T. (1962) ‘On the social psychology of the psychological experiment’, American Psychologist.

Please click here to return to the homepage – ReviseSociology.com