Social psychology is the science of studying the way in which people’s thoughts, feelings and behaviours are influenced by the imagined or real presence of other people (Allport 1985). It originated in the U.S. and developed within the last century. It combines elements from psychology, which focuses on the personality and behaviour of an individual, and sociology, which provides insight into rules and laws to make a group or society function. The focus of social psychology lies on finding out what situations or social influences make people behave or act the way they do.
As an empirical science, social psychology derives its findings from initial hypotheses and theories based on previous research and its results, on the observation of a phenomenon in everyday life, or they are completely new. Evidence must be gathered through a hypothesis-testing model. A social scientist cannot simply rely on logical thinking, common sense or folk wisdom as evidence, as he or she should be able to predict as accurately as possible, when, how, and why things occur. Therefore, scientific research must be conducted to produce the evidence required. If hypotheses and theories cannot be proven, the scientists will either redefine their theories, choose a different method to prove them, until satisfactory proof is obtained, or they might drop their theory altogether.
For the purpose of gathering evidence, the scientist has the choice between three well-developed methods used to study social psychology: the observational method, the correlational method and the experimental method. The following part of the essay will look at each of these three methods and describe their advantages and disadvantages.
The observational method describes behaviour of individuals within a group or of a group of people. It is often used in ethnography to describe different cultures and generate hypotheses about psychological processes (Adler & Adler 1994, Fine & Elsbach 2000). The advantages of the observational method are that the research can be conducted in real-life, natural, non-artificial settings by observing people either from outside or within a group to find out about certain phenomena and to record measurements or impressions of behaviours. In the so-called participant observation, the researcher joins the relevant group and interacts with its members without trying to alter any situation. As stated in Aronson, Wilson & Aitken (2002), Raphael Ezekiel (1995) did this, as he was interested in the behaviours of radical political groups like neo-nazi organisations and the Ku Klux Klan, by joining meetings of those groups without becoming a member himself. It is also possible to let two or more people carry out independent observations and record data relevant to the research. If all judges come to the same result independently, it provides more credible evidence for the scientist. One example for this interjudge reliability mentioned in Aronson et al (2002) is the observation of aggression in schoolchildren during recess of school.
In spite of the positive aspects, there are limits to the observational method. It is restricted to one particular group, setting or action, and results of observations cannot easily be generalised to other cultures, settings or activities. It cannot automatically be assumed that people in the U.S. act or behave in exactly the same way in a situation like people in a Muslim country or in Japan. A further negative point is that some behaviours do either not occur very often or they only occur in private. Research on domestic abuse of any kind would not be possible by means of the observational method.
Another tool within the range of the observational method is the archival analysis. Instead of observing people, this involves the scientist examining written documents of the relevant culture like diaries, novels, magazines and newspapers. Don Smith (1976), for example, used archival analysis to examine the content of pornography in adult paperback literature.
Advantages of having a variety of written materials at hand quickly lead to disadvantages. The available material could have gaps or be seen as unreliable. People, even historical chroniclers, write from different viewpoints. While a tabloid journalist of “The Sun” prefers to distort the facts slightly to ensure a high turnover in sales, a socialist journalist uses facts to show a convincing point against capitalism. These facts show that the archival analysis can be a good backup for other observations, but isn’t too reliable on its own.
The next method to be examined is the correlational method, which deals with the prediction of behaviour. The scientist sets two variables, which are systematically measured and tested. The measurements are then assessed with the help of a correlational coefficient; a statistical technique to show predictability of one variable from another and the result shows the relationship between aspects of behaviour in a number between +1.00 and –1.00.
The advantage of this method is the possibility to find out about the relationships between variables. It does not, however, show if one variable causes the other to occur, so the cause of behaviour cannot be determined, nor does it show the possibility of a third variable being involved as a cause for behaviour or that correlation doesn’t prove causation. An example from Aronson et al (2002) refers to research into the relationship between violence on TV and its effect on children. The results could not determine whether violence on TV caused violence in kids, whether violent kids were more likely to watch violent TV programmes, or whether there was actually no causal relationship between violent children and violence on TV, but a possible third variable as a cause, e.g. neglectful parents.
The main tool of the correlational method is the survey, a questionnaire to be completed by a random selection of people representing the general population, which is chosen to take part. Everyone in the population must be given an equal chance for selection, which is guided by characteristics like age, gender, religion, income level and educational background, depending on the research topic. The survey outcome enables researchers to examine relationships between variables, which are difficult to observe and determined by questions about people’s attitudes, beliefs and behaviour.
Survey results are, however, only useful, if they reflect a response of the general population, not only that of the participants. There are often doubts about the accuracy of survey results or how representative the sample selection was. One example mentioned by Aronson et al (2002) regarding that problem refers to a survey conducted by the Literary Digest magazine in the autumn of 1936 before the presidential election. The results predicted a win of the Republican candidate Alf Landon. However, Franklin D Roosevelt, the Democrat, became president. It turned out afterwards that the magazine’s selection of participants was taken from car registers and telephone directories, but most ordinary people at that time neither owned a car nor a telephone. Other survey errors can occur, when people are asked to predict future behaviour in a hypothetical situation or to explain past behaviour (Schuman & Kalton 1985), Schwarz, Groves & Schuman 1998). There is also a phenomenon of people telling more than they know (Nisbett & Wilson 1977). Surveys are continually improved, but their results probably remain less reliable than results from the next method investigated, the experimental method.
The experimental method looks at the causality of behaviour, examines whether one variable causes the other to occur and can help to draw conclusions of cause and effect. Experiments are systematically set up so that all randomly-selected participants experience the conditions the researcher assigns to them. Each participant gets an equal chance to take part in any condition; all different personalities and backgrounds should be evenly distributed. As experiments could harm the participants’ health and welfare in experiences of unnecessary stress or discomfort, scientists must get the participants’ informal consent prior to the experiment, when the purpose of the study is explained. Ethical issues can contradict or interfere with “good science”. In critical cases, participants are told a cover story rather than the true purpose of the experiments as a means of deception. However, the truth is revealed in the debriefing after the completion of tests.
The scientist sets a dependent and an independent variable with the hypothesis that the independent variable depends on the level of the dependent variable. The independent variable representing one aspect of the situation is the one the researcher changes to check for the effect this has on the continuously measured dependent variable representing the cause of specific behaviour. Aronson et al (2002) refer to research of Latané and Darley (1968) after the Kitty Genovese murder case of the early 1960s. Kitty Genovese was brutally assaulted and murdered in an alley in Queens, New York. The discovery that 38 neighbours had watched the 45-minute ordeal without any of them calling for or attempting to help inspired Latané and Darley to test whether the number of bystanders (as independent variable) had an effect on the people and their readiness to help (dependent variable).
Experiment results should have both a high internal and external validity. A high internal validity is achieved by only the independent variable affecting the dependent variable and all other extraneous variables being controlled. A good probability level, the p-value, calculated by a statistical technique, is another contribution. If the value is less than 5 in 100, the result can be seen as significant. It is, however, more complicated to gain a high external validity (the extent to which experimental study results can be generalised to other situations or people), as laboratory experiments are carried out in an artificial and entirely controlled environment. Some people behave different, if they know they take part in an experiment or if they get a reward for their participation. The random selection of people is often impractical and expensive. Different processes can occur for different cultures. In that case, experiments have to be replicated with another subject population in different settings.
Cross-cultural studies help to determine universal psychological processes unique to humans and the cultural influences on basic social psychological behaviour. The scientists must not impose their own viewpoints and definitions from their own culture onto the unfamiliar culture, but the two variables are to be understood in the same way in the different culture (Bond 1988, Lonner & Berry 1986). One result of cross-cultural research in Aronson et al (2002) shows western cultures focus on individualism and independence, while eastern cultures on collectivism and interdependence (Kitayama & Markis 1994, Markus & Kitayama 1991, Triandis 1989). To achieve a generalisation from cross-cultural studies, meta analysis is used to get an average result from two or more studies.
While laboratory experiments can produce results of a high psychological realism, (psychological processes in the experiment are similar to everyday life), a low mundane realism (similarity of the experiment’s situation with real life) is most likely. A scientist’s dilemma is the trade-off between internal and external validity (Aronson & Carlsmith 1968). To gain a higher external validity the scientist should carry out both laboratory experiments and additional field experiments in natural settings, which are also controlled for extraneous influences. Field experiments help to ensure a higher external validity, but they also mean that results are not always gained by one experiment alone.
Despite all advantages and disadvantages of the main methods in studying social psychology as described in depth in the previous paragraphs of this essay, the valuable contribution they provide towards gaining a better insight and understanding of people’s behaviour must be acknowledged. This applies both to finding out more about people’s behaviour in connection with other people or in various situations as well as to more attempts to discover similarities or differences of behaviours within different cultures.
Copyright Elke Wallace, 2002
Aronson, Wilson & Aitken, Social Psychology, Fourth Edition, Prentice Hall, New Jersey, 2002
This essay was written by my wife, Elke, for her social psychology evening class, and therefore represents her debut on E2.