Technology clusters: Using multidimensional scaling to evaluate and structure technology clusters

related products that form a technology cluster is significantly better than the attributes of an innovation at predicting adoption. The treatment of technology clusters , however, has been ad hoc and study specific: Researchers often make a priori assumptions about the relationships between technologies and measure ownership using lists of functionally related technology, without any systematic reasoning. Hence, the authors set out to examine empirically the composition of technology clusters and the differences, if any, in clusters of technologies formed by adopters and nonadopters. Using the Galileo system of multidimensional scaling and the associational diffusion framework, the dissimi-larities between 30 technology concepts were scored by adopters and nonadopters. Results indicate clear differences in conceptualization of clusters: Adopters tend to relate technologies based on their functional similarity; here, innovations are perceived to be complementary, and hence, adoption of one technology spurs the adoption of related technologies. On the other hand, non-adopters tend to relate technologies using a stricter ascendancy of association where the adoption of an innovation makes subsequent innovations redundant. The results question the measurement approaches and present an alternative methodology. The diffusion of innovations paradigm provides explanations for when and how a new idea, practice, or technique is accepted, rejected, or reevaluated over time in a given society (Rogers, 2003). The strength of the theory is in its ability to coherently structure and predict the rate of adoption of an innovation. According to Rogers (2003), the decision to adopt an innovation is predicted for the most part, by the perceived attributes of an innovation, and to a lesser extent by the personality of the potential innovator. In most research on adoption, the perceived attributes explains from around 49 to 87% of the variance in the rate of adoption (Rogers, 2003). Hence, a great deal of research attention has focused on measuring the key attributes of an innovation that influence adoption decisions. Recent empirical evidence, however, suggests that the adoption of technological innovations is better predicted by the ownership of related innovations. This suggests that innovations are not viewed singularly, but rather as interrelated bundles of new ideas (Rogers, 2003). Following this, a number of researchers have begun including a list of related technologies in their predictive models for explaining the likelihood of adoption of a new technology. For the most part, however, the approach has been ad hoc, with lists of potential clusters or related technologies being formulated without …