Minitypologies from within-subjects designs: uncovering systematic individual differences in experiments

The present paper describes a bottom-up (data-driven) approach to the study of individual differences in experiments having a within-subjects design. The approach uses the similarities between the subjects' patterns of performance in the experiment as the basis for separating them into subgroups (labeled minitypes). It results in a breakdown of the hitherto undifferentiated lump of error variance into two components: systematic individual differences and residual error. When a minitypology is detected the experimenter is rewarded with: (a) finding out the different ways in which people respond in the experimental conditions under study and (b) an interesting set of questions concerning the variables which may account for the different performances. The paper discusses the technical questions involved in the application of the minitypologies technique and presents three examples of its use.