Kelly (1955; 1963) developed personal construct theory based upon his work as a clinical psychologist to help assess his patients’ interpersonal relationships. He determined that individuals would develop a personalised system for dealing with current or future situations that was based upon their own interpretations of their past experiences.
An individual’s personal construct system may be documented using the technique known as the Role Construct Repertory Test, or RepGrid. ‘[RepGrids] … provide a way of doing research into problems … in a more precise, less biased, way than any other research method’ (Stewart and Stewart, 1981). It is also suggested by these authors that the RepGrid technique ‘… enables one to interview someone in detail, extracting a good deal of information … and to do this in such a way that input from the observer is reduced to zero’ (Stewart and Stewart, 1981). The two main components of the RepGrid are elements and constructs. Elements are entities within the research domain upon which the research participant may be able to form an opinion. The constructs are the research participant’s interpretation of the elements within the same research domain.
The RepGrid technique has been employed in research areas beyond those for which it was originally designed. It has been used for general problem construction and market research (Bannister and Mair, 1968; Corsini and Marsella, 1983; Eden and Jones, 1984; Eden and Wheaton, 1980; Fransella, 1981; Shaw, 1980) and for knowledge acquisition for expert systems (Botten et al., 1989; Latta and Swigger, 1992; Phythian and King, 1992). Also, the RepGrid technique has been employed in a series of information systems research projects (Hunter, 1993; Hunter and Beck, 1996; Hunter, 1997; Hunter and Beck, 2000). The research question related to determining how members of various groups construe the skills and personal characteristics of ‘excellent’ systems analysts. That is, the research attempted to document the personal construct system of research participants within the domain of discourse relating to their experiences working with systems analysts.
RepGrids were used during the interview process because they bring structure to the interview while allowing flexibility and reducing researcher bias. It was considered important to determine the interpretations of the research participants. Thus, it was necessary to adopt a tool that emphasised gathering data from the research participant while allowing the participant to determine the subject matter and content of the data. This aspect is one of the advantages of the RepGrid technique.
RepGrids generate a large amount of rich, in-depth, qualitative and narrative data relating to a research participant’s explanation of an elicited construct. The documentation of the research participant’s explanations as interview notes forms the basis of the research data. Detailed comments were recorded for each pole of the elicited construct. The researcher determined a system of hierarchies for each construct, which depicted the relationships, within the interview notes, between an elicited construct at the RepGrid level and a detailed action statement, at the interview note level. The interview notes were obtained via the technique of Laddering whereby the researcher probes further regarding the research participant’s detailed interpretations of a general comment.
The initial project in this series was conducted in Canada. Subsequent replications were carried out in Singapore and again in Canada. This data supported comments regarding cross-cultural aspects of how the performance of ‘excellent’ systems analysts is interpreted. Hunter and his colleagues were able to determine, among a number of results, that information systems professionals perceived ‘excellent’ systems analysts as being process oriented, while business professionals viewed ‘excellent’ systems analysts as those who were able to deliver content. Further, in the cross-cultural replication, the Singapore research participants viewed ‘excellent’ systems analysts as experts, while the Canadian research participants regarded ‘excellent’ systems analysts as coaches.