Classification of Participants: A Guide to Appropriately Using Cluster Analyses

Clustering techniques have been applied to a wide variety of research problems. For example, a group of diners sharing the same table in a restaurant may be regarded as a cluster of people. In food stores items of similar nature, such as different types of meats or vegetables are displayed in the same or adjacent locations. We are usually familiar with the classification of dependent measures such as exploratory factor and confirmatory analyses. However, in the fields of health, physical education and exercise science, classifying participants in physical activity, health and behavior related research is an important step prior to evaluating dependent measures. Since Tryon (1939) first encompassed a method for grouping objects of similar kind into respective categories, different classification techniques have been developed and employed in our field research. The commonly used cluster analyses are: Median split method, k-means cluster analysis, hierarchical cluster analysis, and two step cluster analysis. In this project, the different cluster analyses are compared based on their strengths and weaknesses. The benefits and challenges of the current approaches in classification are discussed when each of the technique is introduced. A procedure for conducting the two step cluster analysis is introduced step by step with the SPPS executing diagrams. For demonstration purposes, the example data are from one of our published manuscripts and there is no conflict of interest involved. The SPSS executing procedure of two step cluster analysis includes: (1) Cleaning up the data, (2) hierarchical cluster analysis, (3) determine of clusters, (4) cluster centers, (5) K-means analysis, (6) label clusters, (7) validation, (8) following up tests MANOVA and ANOVA, and (9) reporting cluster analysis results.