Experiments

What are experiments?

• In the natural sciences such as chemistry and physics, scientists set out to discover scientific laws of cause and effect.
• For example, physicists have discovered that an increase in the temperature of a gas will cause it to expand.
• The method favoured by natural scientists for discovering these laws is the laboratory experiment.
• Sociologists have occasionally used the laboratory experiment as a way of studying human behaviour. In fact, however, this is just one of three different types of experimental method that sociologists sometimes use.
• The others are field experiments and the comparative method.
• 
• 

Laboratory Experiments

• Suppose we want to find out what causes plants to grow.
• One way would be to take a set of identical plants and randomly divide them into two groups – an experimental group and a control group. We then treat them differently:
• The control group - the amount of nutrients the plants in this group receive is constant, and their growth is carefully measured.
• The experimental group - the amount of nutrients they receive may vary and their growth will also be carefully measured.
• After comparing results, we will find that the plants in the experimental group grew more rapidly after receiving extra nutrients.
• We have thus found a cause-and-effect relationship; that receiving nutrients causes growth.
• In scientific terms, the nutrient is the independent variable (the causal factor) and the resulting growth is the effect or dependent variable (since it depends on the first variable, nutrition).
• The logic of the experimental method is that the scientist manipulates (alters) the variables in which they are interested, in order to discover what effect they have.
• By following this method, the scientist can establish a cause-and-effect relationship.
• In turn, this will allow them to predict accurately what will happen in the future underspecified conditions. In our example, the scientist will be able to predict what will happen when a certain quantity of nutrient is given to the plants.

Reliability

• Once an experiment has been completed, other researchers can replicate it.
• Researchers often give specific details of their method into their report so that others are able to do so.
• This means that experiments are highly reliable, for two reasons:
• The original experimenter can specify precisely what steps were followed in the original experiment so other researchers can repeat these in future.
• It is a very detached method: the researcher merely manipulates the variables and records the results. The scientist’s personal feelings and opinions have no effect on the conduct or outcome of the experiment.
• The laboratory experiment therefore has major advantages as the method used to identify cause-and-effect relationships in the natural sciences.
• For this reason, we might expect positivist sociologists to use laboratory experiments, since they favour a scientific approach.
• Despite this, however, there are several reasons why such experiments are rarely used in sociology, even by positivists.

Practical Problems

• Society is a very complex phenomenon. In practice, it would be impossible to identify, let alone control, all the possible variables that might exert an influence on, say, a child’s educational achievement or a worker’s attitude to work.
• It is impossible to use this method to research the past, as a sociologist can only manipulate the variables of what is happening in the present, and not in the past.
• Laboratory experiments can only be used to study small scale samples.
• This makes it very difficult to investigate large-scale social phenomena such as religions or voting patterns.
• The small-scale nature of laboratory experiments also reduces their representativeness.
• 
• 

Ethical Issues

• There are ethical (moral) objections to conducting experiments on human beings, at least under certain circumstances.
• LACK OF INFORMED CONSENT - As a general principle, the researcher needs the informed consent of the research participants. However, this may be difficult to obtain from groups such as children or people with learning difficulties who may be unable to understand the nature and purpose of the experiment.
• DECEPTION - It is also generally considered wrong to mislead people as to the nature of the experiment, as Stanley Milgram (1974) did in his famous studies of obedience to authority. Milgram lied to his research participants about the purpose of the research, telling them that they were assisting in an experiment on learning, in which they were ordered by the researcher to administer electric shocks when the learner failed to answer questions correctly.
• HARM - The research may also cause harm to the participants. In Milgram’s experiments, many research participants were observed to “sweat, stutter, tremble, groan, bite their lips and dig their nails into their flesh. Full-blown, uncontrollable seizures were observed for three subjects.”
• However, supporters of Milgram argue that his experiments can be justified ethically because they alert us to the dangers of blindly obeying authority figures.

The Hawthorne Effect

• Clearly, the setting of laboratory experiments is artificial and unnatural.
• As the situation and set-up of the experiment is fabricated by the researcher, people may behave in different ways than they would in a more natural setting.
• This difference in behaviour decreases the validity of the results.
• If people know they are being studied, they may behave differently; for example, by trying to second-guess what the researcher wants them to do and acting accordingly. This will ruin the experiment, which depends on the subjects responding to the variables that the researcher introduces into the situation, not to the fact that they are being observed.
• This problem has become known as the ‘Hawthorne Effect’ or ‘experimental effect’.
• In 1927, Elton Mayo began research into factors affecting workers’ productivity at the Western Electric Company’s Hawthorne plant. Working with five female volunteer workers who knew he was conducting an experiment, Mayo altered different variables such as lighting, heating, rest breaks and so on to see what effect they had on the volunteers’ output. Surprisingly, not only did output go up when he improved their working conditions, but it continued to rise even when conditions were worsened. Mayo concluded that the workers were not responding to the changes he was making in the experimental variables (such as the lighting),but simply to the fact that they were being studied and wished to please the experimenter.

Free Will

• Interpretivist sociologists argue that humans are fundamentally different to plants or rocks.
• Our behaviour should thus not be explained in terms of cause and effect.
• Instead, human behaviour should be viewed as the result of free choices made by individuals.
• In this view, the experimental method, with its search for causes, is therefore not an appropriate method for studying human beings.
• Given these problems, sociologists have two alternatives to laboratory experiments.
• These are field experiments and the comparative method or ‘thought experiment’.
• 

Field Experiments

• A field experiment has two features that distinguish it from a laboratory experiment:
• It takes place in the subject’s natural surroundings rather than in an artificial laboratory environment.
• Those involved are generally not aware that they are the subjects of an experiment, in which case there is no Hawthorne Effect.
• The researcher manipulates one or more of the variables in the situation to see what effect it has on the unwitting subjects of the experiment.
• For example, in Rosenhan’s (1973) ‘pseudopatient’ experiment, researchers presented themselves at 12 California mental hospitals, saying they had been hearing voices. Each was admitted and diagnosed as schizophrenic.
• Once in hospital, they ceased to complain of hearing voices and acted normally. Nevertheless, hospital staff treated them all as if they were mentally ill. None was found out.
• This suggests that it was not the patients’ behaviour that led to them being treated as sick, but the label ‘schizophrenic’ itself that led staff to treat them in this way.
• Rosenhan’s study shows the value of field experiments. They are more ‘natural’, valid and realistic, and they avoid the artificiality of laboratory experiments.
• However, the more realistic we make the situation, the less control we have over the variables that might be operating.
• If so, we cannot be certain that the causes we have identified are the correct ones.
• Some critics also argue that field experiments are unethical, since they involve carrying out an experiment on their subjects without their knowledge or consent.

The Comparative Method

• Unlike other experiments, the comparative method is carried out only in the mind of the sociologist.
• It is a ‘thought experiment’ and it does not involve the researcher actually experimenting on real people at all.
• However, like laboratoryand field experiments, it too is designed to discover cause-and-effect relationships.
• How to use the comparative method:
• Identify two groups of people that are alike in all respects, apart from the one variable we are interested in.
• C0mpare the two groups and see if this difference has any effect on the people.
• 
• 

Emile Durkheim

• An example of the comparative method is Emile Durkheim’s (1897) classic study of suicide.
• Durkheim’s hypothesis was that low levels of integration of individuals into social groups caused high rates of suicide.
• He argued that Catholicism produced higher levels than Protestantism. From this, he therefore predicted that Protestants would have a higher suicide rate than Catholics.
• Durkheim then tested his prediction by comparing the suicide rates of Catholics and Protestants who were similar in all other important respects (e.g., in terms of where they lived, whether they were married or single etc).
• His prediction was supported by the official statistics, which showed Catholics to have lower suicide rates.
• In seeking to discover cause-and-effect relationships, the comparative method has three advantages:
• it avoids artificiality;
• it can be used to study
• it poses no ethical problems, such as harming subjects past events
• However, the comparative method gives the researcher even less control over variables than do field experiments, so we can be even less certain whether a thought experiment really has discovered the cause of something.

Stanley Milgram

• Stanley Milgram (1974) conducted a famous study of obedience to authority.
• Milgram lied to his research participants about the purpose of the research, telling them that they were assisting in an experiment on learning, in which they were ordered by the researcher to administer electric shocks when the learner failed to answer questions correctly.
• In reality, however, the purpose of the experiment was to test people’s willingness to obey orders to inflict pain.
• Unbeknown to Milgram’s research participants, no electric shocks were actually used. Milgram found that 65 per cent of them were prepared to administer shocks of 450 volts.
• In Milgram’s experiments, many research participants were observed to “sweat, stutter, tremble, groan, bite their lips and dig their nails into their flesh. Full-blown, uncontrollable seizures were observed for three subjects.”
• However, supporters of Milgram argue that his experiments can be justified ethically because they alert us to the dangers of blindly obeying authority figures.
• Moreover, the great majority of his participants (74 per cent) said afterwards that they had learned something of lasting value.

Elton Mayo

• In 1927, Elton Mayo began research into factors affecting workers’ productivity at the Western Electric Company’s Hawthorne plant.
• Working with five female volunteer workers who knew he was conducting an experiment, Mayo altered different variables such as lighting, heating, rest breaks and so on to see what effect they had on the volunteers’ output.
• Surprisingly, not only did output go up when he improved their working conditions, but it continued to rise even when conditions were worsened.
• Mayo concluded that the workers were not responding to the changes he was making in the experimental variables (such as the lighting), but simply to the fact that they were being studied and wished to please the experimenter.
• https://www.youtube.com/watch?v=IxZoxN5IjFE (Incorporar)
• Legenda: ?

Rosenhan

• In Rosenhan’s (1973) ‘pseudopatient’ experiment, researchers presented themselves at 12 California mental hospitals, saying they had been hearing voices. Each was admitted and diagnosed as schizophrenic.
• Once in hospital, they ceased to complain of hearing voices and acted normally. Nevertheless, hospital staff treated them all as if they were mentally ill. None was found out.
• This suggests that it was not the patients’ behaviour that led to them being treated as sick, but the label ‘schizophrenic’ itself that led staff to treat them in this way.
• Rosenhan’s study shows the value of field experiments. They are more ‘natural’, valid and realistic, and they avoid the artificiality of laboratory experiments.
• However, the more realistic we make the situation, the less control we have over the variables that might be operating.
• If so, we cannot be certain that the causes we have identified are the correct ones.
• Some critics also argue that field experiments are unethical, since they involve carrying out an experiment on their subjects without their knowledge or consent.

Harvey and Slatin

• Harvey and Slatin (1976) examined whether teachers had preconceived ideas about pupils of different social classes.
• Harvey and Slatin used a sample of 96 teachers.
• Each teacher was shown 18 photographs of children from different social class backgrounds. To control other variables, the photographs were equally divided in terms of gender and ethnicity.
• The teachers were asked to rate the children on their performance, parental attitudes to education, aspirations and so on.
• Harvey and Slatin found that lower-class children were rated less favourably, especially by more experienced teachers. Teachers based their ratings on the similarities they perceived between the children in the photographs and pupils they had taught.
• This study indicates that teachers label pupils from different social classes and use these labels to pre-judge pupils’ potential.
• Such expectations may be passed on to pupils through non-verbal communication.

Charkin et al

• Charkin et al (1975) used a sample of 48 university students who each taught a lesson to a ten-year-
• old boy.
• One third (the high expectancy group) were told that the boy was highly motivated and intelligent
• One third (the low expectancy group) were told that he was poorly motivated with a low IQ.
• One third were given no information. (the control group)
• Charkin et al videoed the lessons and found that those in the high expectancy group made more eye contact and gave out more encouraging body language than the low expectancy group.

Mason

• Mason (1973) looked at whether negative or positive expectations had the greater effect.
• Teachers were given positive, negative or neutral reports on a pupil.
• The teachers then observed video recordings of the pupil taking a test, watching to see if any errors were made.
• Finally, they were asked to predict the pupil’s end of year attainment.
• Mason found that the negative reports had a much greater impact than the positive ones on the teachers’ expectations.

Rosenthal and Jacobson

• Rosenthal and Jacobson’s (1968) ‘Pygmalion in the Classroom’ illustrates the difficulties of using field experiments to study teacher expectations.
• They carried out their research in a California primary school they called ‘Oak School’.
• Pupils were given an IQ test and teachers were told that this had enabled the researchers to identify the 20 per cent of pupils who were likely to ‘spurt’ in the next year.
• In reality, the test did no such thing and the pupils were, in fact, selected at random.