Genetic algorithms: a pragmatic, non-parametric approach to exploratory analysis of questionnaires in educational research

Data from a survey to determine student attitudes to their courses are used as an example to show how genetic algorithms can be used in the analysis of questionnaire data. Genetic algorithms provide a means of generating logical rules which predict one variable in a data set by relating it to others. This paper explains the principle underlying genetic algorithms and gives a non-mathematical description of the means by which rules are generated. A commercially available computer program is used to apply genetic algorithms to the survey data. The results are discussed.