The Rep Test and Other Sorting Tasks in ILS Research

Abstract

How people make sense of the world around them via the categories they use is a question that social science researchers frequently attempt to address through their investigations (e.g., Spradley, 1970). One prevalent approach in organizational research is the rep grid method (Reger, 1990). This technique, based upon Kelly's (1955) role construct repertory test, asks participants to sort items (e.g., people, recent events, or artifacts) within a three-member group into subgroups based on participant-defined similarity and dissimilarity. The researcher's main task, therefore, is to inquire about the characteristics or conditions that each participant uses to sort these itemsâ what is similar among the paired items and what makes the pair different from the item that is excluded. Reger (1990) remarks that such an approach creates conditions in which "the researcher's frame of reference and worldview would not be imposed upon the respondent" (301). In information and library science (ILS) research, individuals' uses of categories for sense-making are viewed as being closely coupled to their interactions with available information resources, services, or systems. Sorting tasks are one general class of methods used to uncover the categories that users employ during these interactions. In information systems research, Tan and Hunter (2002) discuss qualitative and quantitative applications of the rep grid method. While other sorting tasks in ILS research do not explicitly share the same intellectual lineage as the rep grid method, the general approaches and outcomes are largely consistent. Kwasnik (1991) studied how users classify personal documents based on criteria other than document attributes. She asked participants to provide a 'guided tour' of an office location along with a document sorting process, and, in reflection, she notes that "people are able to articulate the process by which...decisions were made, and the data produced by this articulation lend themselves to analysis at a level which can yield general rules about the behaviour" (389). The qualitative analysis from Kwasnik's (1991) study can be contrasted with quantitative analyses of sort data used for guidance in the design of information displays (Carlyle, 2001) and interoperable metadata schema (Tennis, 2003).The rep grip method and other sorting tasks represent alternative approaches to direct questioning through standard qualitative interviewing. In this presentation, the authors will review various ways in which sorting techniques have been employed in ILS research, highlight how these methods are applied in their research (e.g., Edwards, forthcoming), suggest strategies for the inclusion of these tasks in study designs, and describe unique challenges encountered during data collection and analysis.