Ruth’s Sociology Resources Blog

Experiments are not often used by sociologists, for ethical, theoretical and practical reasons. There are two different types of experiment:

Laboratory experiments are the normal means of conducting research in the natural sciences. Experiments test hypotheses.

In laboratory experiments hypotheses are tested by controlling variables – factors that can change. Results are usually quantitative and based on an objective system of measurement. Results are usually seen as being reliable because the experiment can be replicated.

Because data is quantitative, it is very easy to establish correlation between factors and effects, however, it is not always easy to establish whether correlations are also causal relationships – does the factor actually cause the effect?

However, laboratory experiments are not always seen as valid sources of sociological data about people – this is because people will react differently to situations which they see as artificial than they will to “real” situations.

Field experiments are seen as a possible way of getting round the problems of laboratory experiments. Field experiments are conducted in normal social situations.

An example of such an experiment, was a test of people’s attitudes towards people of different appearances by putting someone dressed as a businessman and then as a tramp outside a railway station to ask directions. Reactions to the questions were then gauged.

It is never possible, when dealing with people, to identify all the variables which may affect experiment results. People may well realise that they are the subject of an experiment and this may change their behaviour – this is known as the “Hawthorne Effect”. (To find out why it is called the Hawthorne Effect, see Sociology in Focus (1995), p.609) Another problem is Experimenter (or researcher) Bias – it is often difficult to know what effect and messages the experimenter is conveying to those being studied.

Experiments also raise ethical questions – is it right to experiment on people? Experiments could potentially be harmful to people. And should they know that they are being experimented on? The general consensus is that people should always know when they are being experimented on, unless this would significantly affect the results, in which case they should be told after the experiment has been carried out.

Instead of experiments sociologists tend to use comparative methods – they compare groups across society, cultures or time. Comparing social variables in different situations may make it possible to establish the causes and effects of social factors.