The Use of Fuzzy Clustering to Examine End User Segmentation

Citation data:

Vol: 56, Page: 605-615

Publication Year:
Usage 138
Abstract Views 96
Downloads 42
Repository URL:
Sethi, Vikram; Govindarajulu, Chittibabu; King, Ruth; Eakin, Mark
article description
A key research theme in the field of end-user computing (EUC) is learning more about end users and their needs and designing support strategies for assisting, managing, and controlling end-user activities. Typology maps, such as those by Rockart and Flannery (1983), have been used to categorize end users into different groups based on criteria such as the skill and sophistication of EUC activity. In most such studies, users self- select themselves into one of several groups based on generic definitions provided in the study. As in most pure taxonomies, a user becomes a part of one and only one group. In this study,we provide an alternative analytical mechanism to self-selection and unitary membership: the use of fuzzy clustering, which allows for gradual membership in different groups, with membership values indicating the probability or degree of membership within a specific group or cluster. Cotterman and Kumarís (1989) classification scheme is operationalized while allowing for overlapping membership probabilities into one of three clusters: user- developer-controller (UDC), user-developer (UD), and user (U). Further, the study also examines the proposition that end users vary in their use of available support sources and also in the type of support that they require. The expectation that different categories of users have different support needs is examined at three levels: cluster level, individual level, and cluster-conjunctive level.