The main methods of investigation available to psychologists

A case study can be used to investigate the behaviour of individuals although they it is not a research method itself. These various non-experimental techniques can be used alongside the study of personal and official documents such as letters, photographs and clinical notes to gather data. The advantages of these non-experimental forms of investigation include allowing the researcher to gather information on variables that are not easily observed. They also make it possible to collect information on a scale that would simply not be feasible by the experimental investigation method.

This means, for example, that a telephone, on-line or paper questionnaire or survey can be sent to hundreds or thousands of people at a time, including, if appropriate, internationally. However, the results from these non-experimental investigations rely upon the capability of the participants to accurately and consistently record responses about their own behaviour and feelings. This puts a premium on the design and compilation of the questionnaire to be used so that questions are clear and unambiguous.

This means giving careful thought to exactly what information is required; how to ask the appropriate questions to elicit that information and how to word the questionnaire that all recipients, who may cover a range of groups with variable educational attainments, understand what the question is asking. Such questionnaires are often best tested on a small pilot group, representative of the larger target group, to check that they are going to work well. This is all time consuming but unless such preparation is done, great problems can occur in trying to analyse and interpret the responses.

There are limitations about what can usefully be covered in such non-experimental methods, for instance, people are not able to predict the future; interesting questions do not always give accuracy (Sherman, 1980) In some case studies, problems arise when trying to find appropriate participants, for example where a psychologist is interested in the effects of brain damage, it is only possible to study an individual patients, who already has brain damage due to a disease or serious accident, whereas it would be useful to have information about their capabilities before the disease or accident occurred.

An example of this would be a patient known as HJA who sustained damage to certain areas of his brain as a result of a stroke. Humphreys and Riddoch (1987) found that HJA now had difficulty perceiving objects. In such a case the researchers had to estimate the person's ability before the incident even though this is not an ideal or accurate way to obtain required information. An additional disadvantage to this non-experimental approach is that cause and effect cannot be determined, so there is a limit as to what can be concluded.

The third and final main method used in psychological research is through observational studies. This basically involves observing animals or people in their natural environment. Observation work can take place in the laboratory but psychologists tend to prefer looking at real world behaviour. There are two ways of doing this: firstly, there are naturalistic observations where participants are studied, by the researcher, unobtrusively (often from a distance) in their natural surroundings.

Another form of non-laboratory work is through active interactions between the researcher and the participant(s), where the idea is to, in effect, become part of their world or group, that is to become 'one of them' in order to acquire the data. Observational studies tend to be required when a psychologist is interested in a variable that cannot be manipulated, for example, gender in such cases experiments are not appropriate. Correlation studies are observationally based and are often used when trying to investigate things like the effect of personality variables on behaviour.

The cause of the observed behaviour is determined by statistically examining relationships between these measurements, that is how one correlates to the other. Observations do not rely upon a participants' ability to describe their behaviour, which can be of advantage to the researcher, who is responsible and can control the reliability of the data obtained. A disadvantage of observational studies is that they do not allow cause and effect to be established, even when dealing with correlations.

An example of this is an investigation into the effect that previous membership in the Boy Scouts has on men's ' participation in community affairs, carried out by Chapin (1938). A group of men who had been in the Boy Scouts was compared with a group of men who had not. The findings were that men who had been members of the Boy Scouts were more likely to join community organisations later on in life. Merely observing a correlation between Boy Scout membership and later involvement in community organisations would be establish a cause and effect relationship between the two – it may be that there are other causative factors in play.

The cause and effect of this can may only be found through a controlled experiment, so observational studies have their limitations. A lot of detail and care (and more importantly time) must be taken when designing an observational experiment, to avoid problems with inaccurate results. One example of this is the case of training guards to treat prisoners more like guards (Musty, 1972). All the observers had to be given correct descriptions of the typical guard behaviour which was to be recorded. The experiment was carried out over a long period of time.

Observational studies tend to be rather time consuming and quick results are not always obtained. In the case of observational studies which involve active interaction with the participants, there can be a risk of provoking behaviour that would not have happened in the absence of the researcher but it has been suggested by Weick (1968) that through active interactions with the participants by the observer, a greater control and precision can be achieved and such 'provocation' does not necessarily forgo the naturalness of social behaviour.

Trained psychologists are able to evoke behaviour in social setting, without actually invalidating the outcome. For example, an accomplice who requires help can be used and the response by individuals can be recorded. A black woman and a white woman were used in an experiment carried out by Bickman and Kamzan (1973) to see whether race was an issue when someone is asking for help (money), no racial differences were found. A real life social setting was used (a supermarket).

This experiment was relatively easy to carry out and the results were easy to interpret. In conclusion, there are three main methods of investigation: experiments; surveys; and observations. They all have distinct advantages and disadvantages as described above. In identifying which method would be most appropriate in any particular case, psychologists would need to consider these factors and reach a judgement about which method would produce the most valuable and useful results, depending on what they are investigating.

In some cases, it may be appropriate to use a combination of investigation methods, for example using a survey technique to gather data and then an experimental approach to test different hypotheses as to what cause and effect linkages there might be. It may also be, that in some circumstances, external factors, like time and money, dictate which method is to be used or which have to be eliminated. In such cases, it is important to recognise the limitations of the method adopted and consider how the disadvantages might need to be reflected in the design of the investigation or in the interpretation of the results.