Analysing Data of Interacting Groups of People

Social psychologists and others are often interested in comparing scores of people in different groups of people who interact within these groups. A classic example is interaction between students in school group. These situations present a really big problem for classical statistical approaches which assume that the people being analysed are independent of each other. There are often ways to get around this but they involve extremely complex statistics.

Researchers in the School of Psychology at The Australian National University have developed an extremely straightforward method for analysing group data like these. The technique developed by Craig McGarty and Michael Smithson relies on extremely simple statistical methods (the binomial test) and can be conducted quickly in field research with the aid of nothing more than a pocket calculator and a table of values.

The method has been published in a paper in the European Journal of Social Psychology, and a workshop given by Craig McGarty at ANU to advanced undergraduates which compares the method with other methods is available as a pdf of a powerpoint presentation.