Categorization Constructionist Assessment with Software-Based Affinity Diagramming

Affinity diagramming is a cheap and widely used knowledge elicitation technique in human–computer interaction (HCI). However, empirical methods for evaluating user performance in conducting affinity diagrams have remained relatively static. Despite the fact that often the main value of such affinity diagramming sessions lies in the group-based discussions and debates that take place during their construction, what is being captured is often only the final categorizations (the affinity diagram) rather than the process of constructing them. In this article, we propose the concept of categorization constructionism, which we describe as optimized when affinity diagrams are facilitated in groups that have a considerate input of activity in categorization decision-making. We describe how we used this rule to model the temporal nature found within affinity diagram categorizations as they are constructed. To help us test our approach, we utilized participatory design (PD) sessions in developing three TabletPC-based software tools (CATERINE, SAW, and MATE) that would record, allow manipulation of, and evaluate the organization of affinity constructs over time programmatically with digital inking processes. We then used these tools to conduct an experiment that would explore our concept of measuring constructionistic activity over time in practice through the use of our tools.