ABSTRACT
Traditionally, research in sport and exercise has used methods based on the assumptions of the normative or the interpretive paradigm (Cohen et al., 2007: 7–26). The human nature assumptions of these paradigms motivate different approaches to data analysis. The normative paradigm assumes that human behaviour is deterministic and can be characterized by an average human being. The interpretive paradigm, however, assumes that human behaviour is voluntaristic with individual responses to situations. There are occasions where neither of these approaches is satisfactory due to the nature of the behaviour being investigated. There may be too many participants to analyse as individuals, but there may be different types of individual that we wish to identify and analyse rather than imposing a model of an average person. Sometimes different types of participant can be identified from existing variables such as positional role, gender, socio-economic group, level of participation and age category. However, there are other classifications based on combinations of variables collected during a study. These classifications may be based on behaviour, attitudes or other variables. Cluster analysis is a technique that allows such non-obvious groupings of participants to be identified.
