Computer-aided software engineering (CASE), a relatively recent technological innovation, is viewed by both researchers and practitioners as a potential means to increase the productivity (Banker and Kauffman, 1991; Norman and Nunamaker, 1988; Stamps, 1987; Swanson, et al., 1991) and quality (Howard, 1990) of information systems development activities, reduce costs and time spent in systems development (Feuche, 1989; Martin, 1989), and ease the software development and maintenance burden threatening to overwhelm information systems departments (Bachman, 1988; Banker and Kauffman, 1991; Swanson, et al., 1991). Actual experiences with CASE tools, however, have been mixed. While some studies have reported productivity gains (or perception of such gains) from the use of CASE tools (Banker and Kauffman, 1991; Necco, et al., 1989; Norman and Nunamaker, 1988; Swanson, et al., 1991), many others have found that the expected productivity gains are elusive (Card, et al., 1987; Yellen, 1990), or hampered by inadequate training and experience, developer resistance, and increased design and testing time (Norman, et al., 1989; Orlikowski, 1988, 1989, 1993; Vessey, et al., 1992). These contradictory experiences withCASE tools have been difficult to interpret and have puzzled both practitioners and researchers. The inadequacy of conceptual and theoretical foundation of organizational innovation diffusion, primarily based on the classical diffusion theory first espoused by Rogers (1962), have been cited as a prime reason for the contradictory empirical findings (Fichman, 1992). The classical diffusion theory, used in most studies of IT diffusion in general and CASE diffusion in organizations in particular, has many shortcomings. First, the theory operates under the assumption of an unchanging innovation (Brown, 1981). In reality, innovation is a continual process whereby the form and function of the innovation are modified throughout its life (LeonardBaron, 1988; Walton, 1989). Second, the theory emphasizes the demand aspect of diffusion, assuming that everyone has an equal opportunity to adopt; the supply side of the innovation is almost ignored (Brown, 1981). In fact, institutions that supply and market innovations determine to a certain extent who adopts them and when. Third, the classical diffusion theory considers the technological adoption decisions of individuals or organizations without taking into account community issues, assuming that individuals adopt innovations for their own independent use (Fichman, 1992). However, there is evidence that the technology can be subject to network externalities (Katz and Shapiro, 1986; Markus, 1987), which means that the value of use to any single adopter will depend on the size of network of other users. Fourth, the classical theory fails to distinguish between two types of communication involved in the diffusion process: signaling versus knowhow or technical knowledge (Attewell, 1991). It assumes that signaling information takes different lengths of time to get to different potential adopters (according to their centrality to communications networks and links to prior adopters), resulting in the early, middle, and late Scurve adopters, and is therefore viewed as central in explaining the diffusion process. However, one may question whether signaling information is a limiting factor in situations where information about the existence of new technologies and their benefits is widely broadcast by manufacturers' advertisements, by specialized business journals, and by trade associations (Burt 1987). The technical knowledge required to use a complex innovation successfully places far greater demands on potential users and on supplyside organizations than does signaling (Attewell, 1992). If obtaining technical knowledge is slower and more problematic, it can be posited that it plays a more important role in the diffusion of complex technologies than does signaling. Finally, most of the studies of supply-side institutions in innovation conceptualize the diffusion process in terms of knowledge transfer. Attewell (1992) argues that such studies treat the movement of complex technical knowledge under a model of communication most appropriate for signaling. Studies have, however, shown that although one can readily buy the machinery that embodies an innovation, the knowledge needed to use modern production innovations is acquired much more slowly and with considerably more difficulty (Arrow, 1962; Dutton and Thomas, 1985, Ray, 1969; Pavitt, 1985; von Hippel, 1988). Absorbing a new complex technology not only requires modification and mastery of the technology, but it also often requires (frequently unanticipated) modifications in organizational practices and procedures (Stasz, Bikson, and Shapiro, 1986; Johnson and Rice, 1987). Thus, implementing a complex technology requires both individual and organizational learning. Not surprisingly, the findings of past studies of IT diffusion show inconclusive support for the classical diffusion theory in the case of diffusion of complex information technologies (such as CASE) which exhibit user interdependency and impose knowledge burden on users (Fichman, 1992). (When the adoption decision of individuals or organizations depends on the dynamics of community-wide levels of adoption because of network externalities, innovation diffusion is characterized as exhibiting user interdependencies. Similarly, when technologies cannot be adopted as a "black box" solution but rather impose a substantial knowledge burden on potential adopters, innovation diffusion is characterized to exhibit high knowledge burden.) One interpretation of these findings is that classical diffusion variables by themselves may not be strong predictors of adoption and diffusion of complex technologies at the organizational level (Fichman, 1992). Fichman (1992) recommends that future research on IT diffusion at the organizational level consider other than classical or communications perspective, such as market and infrastructure, economic, and organizational learning perspectives, to account for these inconsistencies. In this study we complement the classical diffusion theory with an organizational learning perspective.
Sharma, Srinarayan, "Studying ComputerAided Software Engineering Diffusion in Organizations: Complementing Classical Diffusion Theory With Organizational Learning Perspective" (1995). AMCIS 1995 Proceedings. 146.