Social Psychology - Research - Methods

Methods

Social psychology is an empirical science that attempts to answer questions about human behavior by testing hypotheses, both in the laboratory and in the field. Careful attention to sampling, research design, and statistical analysis is important; results are published in peer reviewed journals such as the Journal of Experimental Social Psychology, Personality and Social Psychology Bulletin and the Journal of Personality and Social Psychology. Social psychology studies also appear in general science journals such as Psychological Science and Science.

Experimental methods involve the researcher altering a variable in the environment and measuring the effect on another variable. An example would be allowing two groups of children to play violent or nonviolent videogames, and then observing their subsequent level of aggression during free-play period. A valid experiment is controlled and uses random assignment.

Correlational methods examine the statistical association between two naturally occurring variables. For example, one could correlate the amount of violent television children watch at home with the number of violent incidents the children participate in at school. Note that this study would not prove that violent TV causes aggression in children: it is quite possible that aggressive children choose to watch more violent TV.

Observational methods are purely descriptive and include naturalistic observation, "contrived" observation, participant observation, and archival analysis. These are less common in social psychology but are sometimes used when first investigating a phenomenon. An example would be to unobtrusively observe children on a playground (with a videocamera, perhaps) and record the number and types of aggressive actions displayed.

Whenever possible, social psychologists rely on controlled experimentation. Controlled experiments require the manipulation of one or more independent variables in order to examine the effect on a dependent variable. Experiments are useful in social psychology because they are high in internal validity, meaning that they are free from the influence of confounding or extraneous variables, and so are more likely to accurately indicate a causal relationship. However, the small samples used in controlled experiments are typically low in external validity, or the degree to which the results can be generalized to the larger population. There is usually a trade-off between experimental control (internal validity) and being able to generalize to the population (external validity).

Because it is usually impossible to test everyone, research tends to be conducted on a sample of persons from the wider population. Social psychologists frequently use survey research when they are interested in results that are high in external validity. Surveys use various forms of random sampling to obtain a sample of respondents that are representative of a population. This type of research is usually descriptive or correlational because there is no experimental control over variables. However, new statistical methods like structural equation modeling are being used to test for potential causal relationships in this type of data.

Regardless of which method is used, it is important to evaluate the research hypothesis using the results, either confirming or rejecting the original prediction. Social psychologists use statistics and probability testing to judge their results; these define a significant finding as less than 5% likely to be due to chance. Replications are important, to ensure that the result is valid and not due to chance, or some feature of a particular sample. False positive conclusions, often resulting from the pressure to publish or the author's own confirmation bias, are a hazard in the field.