Information Systems Research Methods: The Technology Transfer Problem

The IS research community has moved away from concentration upon the technical issues associated with IS and now tends to focus on more behavioural issues. This paper discusses the evolution of IS research themes and summarises the methodologies which illustrate this shift in emphasis and the current trends in IS research. The paper then argues that the "accepted" taxonomies of IS research methodologies characteristic of these trends do not adequately support technology transfer and the diffusion of innovation, essential to effective applied IS research. The paper concludes that systems development should be considered a legitimate research activity within the technological domain of IS. We illustrate the technology transfer capabilities of systems development by means of two IS research projects currently underway. INTRODUCTION For some time now, there has been debate within the information systems (IS) research community over the choice of research methodologies and their suitability for different aspects of IS research. Traditionally the IS community focused primarily on the technical problems associated with the use of IS in industry, commerce and administration, possibly because systems developers had difficulties getting the computerised systems to work effectively (Keen, 1980; see also Backhouse et al, 1991). IS professionals have since acknowledged that political, organisational and social concerns have an impact on the effective use of IS within a company (Backhouse et al, 1991; see also Mumford et al, 1985). The field of IS, consequently, is a discipline in which practitioners must understand both the human and technological factors associated with IS with equivalent proficiency (Avison and Fitzgerald, 1991). This paper was published in the proceedings of AACIS=94" the 5 Australasian Conference on Information Systems , Melbourne.Victoria, September 27-29, 197-208. The IS research community has also adopted this "softer" philosophy because it now tends to focus on behavioural issues and, subsequently, has shifted away from the technical issues associated with IS (Keen, 1987; see also Backhouse et al, 1991). Despite this shift, Keen still argues that the primary "mission" for IS research "... is to study the effective design, delivery, use and impact of information technologies [IT] in organisations and society" (Keen, 1987:3). Each term used in Keen's definition of the IS research objectives provides a broad dimension of study for a researcher: effective: IS research should focus on improving practice through research into the design, delivery, use and impact of IT design: systems design is fundamental to IS, so that IS researchers must understand the design process to identify potential and unlikely applications of IT delivery: IS research should consider the requirement for IS to work both organisationally and technologically use and impact: IS research should consider the consequences and contributions of IT within the perspective of the entire organisation, as well as the development and implementation process information technologies: alternative approaches, such as executive information, group decision support and expert systems, offering additional facilities, can be considered in addition to traditional IT solutions (including software, data and telecommunications) in organisation and society: the application of IT within these contexts should be the focus of study, in contrast to studying IT itself. As Keen (1987) suggests, the IS discipline must have a sound theoretical base to support the study of practical issues and innovations. Consequently, there is a need for basic (pure) research to be undertaken, both in terms of cross-disciplinary research and of study within individual reference disciplines (such as Sociology, Management or Computer Science). Nonetheless, since the spiralling impact of IT on the business world has provided much of the impetus for IS research over the past two decades, IS research has tended to be (and remains) primarily applied in nature. While "... scientists [must] produce work that is original and is valued by others in the same field ... for the applied scientist at least, the idea or product must [also] be useful" (Smith, 1990:12; see also Leedy 1993). Smith's views can be generalised to all researchers and, in the case of applied researchers in IS, it is apparent that the business community is the group to which the product must be useful. It is clearly not sufficient, however, that a product of IS research be useful it must also be used. It is therefore crucial to the derivation of benefit from an applied research programme that a "technology transfer" process take place. In the examination of Information Systems research (and, particularly, of IS research methods) which follows, however, we find little evidence of support for a technology transfer programme or for research methods sufficient to underpin such a programme within the IS research community in a fully adequate manner. In this paper we discuss: ! the nature of Information Systems and IS research ! IS research methods; and ! the inability of "accepted" IS research methods fully to support technology transfer and the diffusion of innovation. We then consider some of the problems of technology transfer and suggest appropriate research methods to address these issues. Finally, we discuss two specific research projects currently underway (involving technology transfer) to illustrate the limitations of the existing IS research taxonomies and to demonstrate the way in which prototypical software development may be used as an IS research method (after Nunamaker et al, 1990-91). THE NATURE OF INFORMATION SYSTEMS "An information system is a system which assembles, stores, processes and delivers information relevant to an organisation (or to society), in such a way that the information is accessible and useful to those who wish to use it, including managers, staff, clients and citizens. An information system is a human activity (social) system which may or may not involve the use of computer systems." (Buckingham et al, 1987:18) Information Systems has attracted researchers from a variety of disciplines, including computer science, management science, economics, sociology, psychology and behavioural science (see Bariff and Ginzberg, 1982; Keen, 1987; Avison and Fitzgerald, 1991; Backhouse et al, 1991). The multi-disciplinary background of the community has provided a rich domain having a plethora of views, experiences and backgrounds on which to base its understanding of the development, use and impact of IS. It has also, however, resulted in contention regarding the desirable nature and boundaries of IS. That the field of IS has no well defined domain is evident from the number of articles beginning with a "working definition" (see, for example, Cooper, 1988) or with a statement of the perspective from which the article is written (for example, management and organisational, as opposed to technical), or which discuss the very nature of Information Systems itself (Mumford, 1991; Banville and Landry, 1992). Keen (1987:13) comments: "The IS research field has found it difficult to build coherence and visibility. Yet it continues to attract interest and commitment. The main reason is surely that the people who view themselves as part of this community are concerned with organisational effectiveness rather than just technology." Agreeing with Keen's views, Avison and Fitzgerald (1991:7) note that: "It seems to be the case that for a subject [such as IS] to be considered a legitimate discipline of knowledge there must be a general agreement that its object of study is unique and possesses social utility... Clearly information systems possesses social utility, though there is no firm agreement on what is the object of study." The existence of multiple "reference disciplines" (Keen, 1980), particularly as some disciplines are considered to be more relevant than others (Teng and Galletta, 1990), adds to the confusion caused by the lack of an agreed definition for IS. The diversity of opinion on the nature of IS is not surprising considering the discipline's history. Nevertheless, this is a source of difficulty for IS researchers who must decide how to map out future directions for research in the area and to justify a research topic in an effort to make it academically respectable. An empirical study carried out by Farhoomand (1992) has ranked the popularity of research "themes" in IS (originally reported by McKenney in McFarlan, 1984:337) during the years 19771985 (see Table 1). Ranking of IS Research Themes Key IS Issues 1. Databases, systems, software design Improved IS planning 2. Management of and user planning for IS Facilitation and management of end-user computing 3. Human-computer interface Integration of data processing, office automation and tele-communications 4. Implementation Improved software development and quality 5. DSS, decision theory Measuring and improving IS effectiveness/productivity 6. Data management, information resource management Facilitation of organisational learning usage of IT 7. Information requirements analysis Aligning the IS organisation with that of the enterprise 8. Organisational design Specification, recruitment and development of IS human resources 9. Technology transfer Effective use of organisation's data resources 10. Impact Development and implementation of DSS Table 1 IS Research Themes 1977-1985 (Farhoomand, 1992) It is clear from this ranking that IS development (including information requirements analysis, software design and software implementation) was considered to be an important and researchable sub-domain of IS at that time. Elsewhere in the IS literature, software development is frequently reported as a research area (Teng and Galletta, 1990; Watson and Brancheau, 1992). Land (1992) and Galliers (1992; 1993), however, in their discussions of the IS research domain, take a very different view. While acknowledging the underlying technology of an information system, they consider this to be a negligible component of the Information Systems domain. There is considerable support for this more recent view and, in fact, the present trend in the Information Systems research community is towards a view of IS as a social system, almost to the exclusion of the technical component (Cooper, 1988; Liebenau and Backhouse, 1989). Indeed, there have been calls to regard Information Systems from a non-technological viewpoint (Liebenau and Backhouse, 1989; Galliers, 1993). Perhaps the most telling expression of this view is expressed by Bjørn-Andersen (1985:274): "For better or for worse, it [IS] must be seen as a social science discipline." The views of those "middle-of-the-road" commentators who have proposed a socio-technical perspective (Sandberg, 1985; Avison and Fitzgerald, 1991) which recognises the social and technical aspects of information systems and the need for both to co-exist have been dismissed as representing the "technical camp". Klein and Hirschheim (1987:287) comment: "Here, IS are perceived as so-called socio-technical systems with the emphasis placed on the technical rather than social side. However, little more than lipservice is paid to social aspects other than quality of working life concerns related to job satisfaction needs." Galliers (1993b:119), reporting on the UK Doctoral Consortium 1992, comments: "One issue that did arise this year, of which supervisors should be aware, is the different philosophies underpinning doctoral research in Software Engineering vis à vis IS. The former group is clearly being counselled into what might be described as R&D activity, with the major contribution being some new software or other. Conversely, the latter see their contribution as being more theoretical than tangible." We are concerned to see that there appears to be a move to expel, from the domain of IS, research into the development of information systems particularly if that research should focus on an instance (or instances) of software development. We have argued (Swatman and Swatman, 1992) that there are domain-dependent aspects of systems development. That is, the process of systems development within the IS domain is, in many respects, significantly different from the process of systems development within the technical/scientific domain, which is the traditional territory of Software Engineering. The tacit assumption that systems development research is not legitimate as a research topic/approach in IS would imply the need for a separate IS software engineering domain concerned with technology transfer from research to industry. Currently, no such domain exists. 2For example, software engineers operating in a technical or scientific environment are often mathematically sophisticated, while systems developers working in a commercial context are more skilled in interaction with technically unsophisticated clients. IS RESEARCH METHODOLOGIES There is an abundance of literature on IS research methodologies and although there is no consensus on which research methodology is applicable within which problem context, there is general agreement on the plurality of research methods in this field. Several taxonomies have appeared in the literature (Van Horn, 1973; Vogel and Wetherbe, 1984, Avison and Fitzgerald, 1991; Galliers, 1985; 1992; Shanks et al., 1993). In Table 2, we set out a comparison of these taxonomies. Research Approach Van Horn V&W A&F Galliers Shanks Theorem Proof X X Engineering X Forecasting X Mathematical Modelling X Laboratory/Adaptive Experimentation X X X X X Field Experiment/Test X X X X X Simulation X X Survey X X X X X Case Study X X X X X Phenomenology/Hermeneutic/ Descriptive/Interpretive X X X Action Research X X X Futures Research X Role/game playing X Subjective/Argumentative (Conceptual Study) X X X X Table 2 A Comparison of Taxonomies of IS Research Methods The most noticeable trend in the evolution of these taxonomies is the increase in the number of interpretivist approaches. This is a reflection of the general shift towards interpretivism, but it has also put Information Systems research in something of a dilemma requiring a trade-off between relevance and rigour (Keen, 1991). It is, for example, not uncommon to find calls for more qualitative research in IS alongside calls for more empirical work, which is often viewed as being more rigorous. If, however, one examines the literature which reports the published research within the IS domain, we find that: ! the period 1970-1980 was characterised by mainly non-empirical studies (Hamilton and Ives, 1992) ! this was followed by calls for more empirical studies in the early 1980's (Farhoomand, 1992) ! the late 1980's and early 1990's, however, have seen criticism of the preponderance of empirical studies in IS research and, consequently, calls for more qualitative research (Klein et al, 1991). A further, detailed study by Alavi and Carlson (1992), covering the years 1968-1988, provides further evidence of the continuing move towards the use of empirical research methods within the IS domain. While the literature concerning published IS research shows a move towards greater use of empirical research methods over this period (Cooper, 1988; Farhoomand, 1992), an examination of the IS research methodological literature shows calls for greater emphasis to be laid on interpretivist research (Boland, 1985; Klein and Hirschheim, 1987; Galliers, 1993). A disturbing aspect of this trend is noted by Baroudi and Orlikowski (1989), who report that the statistical treatment in much of the published IS empirical work is inadequate. One could interpret this apparent disagreement within the IS community as suggesting that the methodologists' call for a move away from scientific/positivist research methods towards interpretivism has been interpreted by the IS research community as a call for research into sociological, rather than technological, issues in IS. That is, the call for softer research methodologies appears to have resulted in the study of predominantly socio-organisational rather than technological research topics. These softer topics have been investigated using increasingly empirical methods, so as to add rigour to what is perceived as being insufficiently "scientific" research. CATERING FOR TECHNOLOGY TRANSFER We have argued above that Information Systems is currently considered to be primarily an applied discipline. Further, to be effective as an applied discipline, it must not merely produce useful products but those products must become used within the target community. An examination of the summary of IS research taxonomies presented above allows us to exclude the following research methods from consideration, since they are suitable for: ! basic research:$ theorem proof $ mathematical modelling ! examination of the current or previous state of the subject matter:$ subjective/argumentative $ phenomenological/hermeneutic research. $ survey $ case study ! examination of trends within the target community:$ forecasting