Information Systems - The State of the Field

334 Cleaning the Mirror

Open source software

Figure 17.1 Tight and loose topic areas in information systems

Legend:

TAM = Technology acceptance model

DOI = Diffusion of innovations

TCE = Transaction cost economics

Tight areas are smaller shapes.

Loose areas are larger shapes.

potentially shifts economic activities away from hierarchical control and into market forms of governance.

The larger shapes in Figure 17.1 designate topic areas informed by a wider variety of theoretical approaches, research questions and methodologies. The three examples shown in Figure 17.1 are the organizational impact of information systems, open source software and business value of IT. With the exception of open source, these are not especially new topic areas. However, they have eluded paradigm consensus because researchers continue to explore contradictory findings and pursue new theoretical angles for investigating the same topic. New angles challenge older assumptions, such as the simple connection between IT and business transformation or business value. Researchers in loose topic areas are uncomfortable settling for obvious relationships, striving to explain complex phenomena with greater fidelity. Because new theoretical ideas are often drawn from other disciplines, loose topic areas maintain flexibility and adaptability.

Both tight and loose topic areas may be criticized, but on different grounds. Tight areas might be accused of wasting resources to strengthen accumulated findings about things that we already know. For example, in the presence of more pressing issues in IS, the

Cleaning the Mirror 335

established relationship in TAM between user perceptions and user intentions might be judged trivial and unimportant. Likewise, understanding the antecedents of technology adoption in DOI might be judged less relevant where users have less discretion to choose technologies. Loose topic areas may be criticized for failing to produce strong results (Lyytinen and King, 2004). If researchers continually challenge each other and produce new theories instead of replicating prior work, IS might achieve an agnostic state where nothing is known for certain. Loose areas of IS research may also be criticized for too much borrowing and not enough originality (Weber, 2003).

Under paradigm consensus, we clean the mirror and see researchers engaged in similar efforts. We thus create a ‘school’ of IS research where the rules are established and enforced through editorial policies, criteria for promotion and tenure, and other measures (in which the term ‘discipline’ takes on a different meaning). Small deviations from convention are tolerated but undisciplined departures are discouraged.

Under diversity, we clean the mirror to see an alarming variety of topics, theories and research methods. To some, the mirror may appear shattered, a mosaic reflecting images in multiple directions. On closer inspection, however, each facet of the field may offer something unique that is worth learning. A researcher trained in the social sciences may discover design research (Hevner et al., 2004); a scholar grounded in positivist philosophy may discover interpretive epistemology; an economist may derive new insights into electronic markets by learning about social institutions; and so on. As we entertain new perspectives on complex problems, we establish a different kind of identity, one that is flexible and adaptive in response to changes in the world. In this scenario, the field’s identity is constantly reshaped.

Careful reviews of tight topical areas reveal the potential to complicate and loosen them. For example, DOI studies may be complicated by asking research questions about patterns of IT use following adoption. Many technologies are adopted, but research suggests that users can appropriate similar or identical technologies in different ways. For example, theories of human agency and practice (Emirbayer and Mische, 1998; Orlikowski, 2000) have informed recent research on IS use following technology adoption or implementation (e.g., Boudreau and Robey, 2005; Schultze and Orlikowski, 2004). Likewise, TAM could be complicated by turning attention toward the patterns of usage involving ensembles of technologies that users have already accepted. Here the emphasis would shift from whether users accept technology to how users appropriate the technologies that they have accepted. By complicating the narrow

336 Cleaning the Mirror

focus on adoption decisions, these tight paradigm areas could be made looser and perhaps more interesting and relevant.

Careful reviews of research in loose topic areas are also needed to establish findings that are consistent across a wider range of studies. Rather than assuming complete diversity, such reviews should show regions within topic areas where a tighter consensus has emerged. For example, studies in open source software development appear to have produced a good understanding why developers contribute to open source projects when they receive no direct payment from generating public goods (Lakhani and Von Hippel, 2003). Prior to studies of this phenomenon, open source projects existed as interesting paradoxes. Reviews of loose areas should also seek areas where new research should be done. For example, Elgarah and her associates (2005) examined the theoretical underpinnings of research on EDI, identifying five conceptual gaps that required further research.

Like jazz, the IS field shares many characteristics of ‘modern’ social institutions. Modern social relations are plucked from other times and places and inserted into local environments (Giddens, 1990). As theories and research practices are taken from other disciplines, IS becomes more heterogeneous, in effect stretched across broader expanses of time and space. In its worst incarnations, modern institutions resemble fractured mosaics: post-modern fragments with no center. At their best, modern institutions extend their relevance and capabilities, building outward from traditions to create multiple centers of activity. This is a much different identity from the one implied by a single core paradigm. For me, it is a more attractive vision for the IS field, one that values and honors its inherent diversity.

NOTE

1 I thank Vanessa Liu for her feedback on an earlier draft of this commentary.

REFERENCES

Banville, C. and Landry, M. (1989) ‘Can the Field of MIS be Disciplined?’ Communications of the ACM, 32, pp. 48–60.

Benbasat, I. and Weber, R. (1996) ‘Rethinking “Diversity” in Information Systems Research’, Information Systems Research, 7(4) December, pp. 389–99.

Benbasat, I. and Zmud, R. W. (2003) ‘The identity crisis within the IS discipline: Defining and communicating the discipline’s core properties,’ MIS Quarterly, 27(2), June, pp. 183–94.

References 337

Boudreau, M-C. and Robey, D. (2005) ‘Enacting integrated information technology: A human agency Perspective’, Organization Science, 16(1), January–February, pp. 3–18.

DeSanctis, G. (2003) ‘The social life of information systems research: A response to Benbasat and Zmud’s call for returning to the IT artifact’, Journal of the Association for Information Systems, 4(7), December, pp. 360–76.

Elgarah, W., Falaleeva, N., Saunders, C. S., Ilie, V., Shim, J. T. and Courtney, J. F. (2005) ‘Data exchange in interorganizational relationships: review through multiple conceptual lenses’, The DATA BASE for Advances in Information Systems, 36(1), Winter, pp. 8–29.

Emirbayer, M. and Mische, A. (1998) ‘What is agency?’ American Journal of Sociology, 103, pp. 962–1023.

Galliers, R. D. (2003) ‘Change as crisis or growth? Toward a trans-disciplinary view of information systems as a field of study’, Journal of the Association for Information Systems, 4(6), November, pp. 337–51.

Giddens, A. (1990) The Consequences of Modernity, Stanford University Press, Stanford, CA.

Hentoff, N. (1996) Original Liner Notes to John Coltrane’s ‘Meditations’, Impulse Records, 1966.

Hevner, A. R., March, S. T., Park, J. and Ram, S. (2004) ‘Design science in information systems research’, MIS Quarterly, 28(1), March, pp. 75–105.

King, J. L. (1993) ‘Editorial Notes’, Information Systems Research, 4(4), December, pp. 291–8.

Lakhani, K. R. and von Hippel, E. (2003) ‘How open source software works: “Free” user-to-user assistance’, Research Policy, 32(6).

Lyytinen, K. and King, J. L. (2004) ‘Nothing at the center?: Academic legitimacy in the information systems field’, Journal of the Association for Information Systems, 5(6), 2004, pp. 220–46.

Orlikowski, W. J. (2000) ‘Using technology and constituting structures: A practice lens for studying technology in organizations’, Organization Science, 11, pp. 404–28.

Orlikowski, W. J. and Iacono, S. (2001) ‘Research commentary: Desperately seeking the “IT” in IT research: A call to theorizing the IT artifact’, Information Systems Research, 12(2), pp. 121–34.

Robey, D. (1996) ‘Diversity in information systems research: Threat, promise, and responsibility’, Information Systems Research, 7(4), pp. 400–08.

Robey, D. (2003) ‘Identity, legitimacy, and the dominant research paradigm: An alternative prescription for the IS discipline’, Journal of the Association for Information Systems, 4(6), June, pp. 352–59.

Schultze, U. and Orlikowski, W. J. (2004) ‘A practice perspective on technologymediated network relations: The use of Internet-based self-serve technologies’,

Information Systems Research, 15, pp. 87–106.

Weber, R. (2003) ‘Still desperately seeking the IT artifact’, MIS Quarterly, 27(2), June, pp. iii–xi.

Whinston, A. B. and Geng, X. (2004) ‘Operationalizing the essential role of the information technology artifact in information systems research: Gray area, pitfalls, and the importance of strategic ambiguity’, MIS Quarterly, 28(2), June, pp. 149–59.

18

Designing Design Science

Salvatore T. March

Relevance, rigor and results are the trifecta of academic research. They are defined by the constituency that comprises and supports the discipline (Denning, 1997; Tsichritzis, 1999). In information systems (IS) this constituency includes IS academic researchers, organizations that develop and deploy information technologies (IT), organizations that procure and implement such technologies, IS managers within such organizations and, more and more commonly, general and upper level managers within such organizations. The results demanded by this constituency are (1) innovative IT artifacts that address business tasks and problems and (2) the knowledge of how to effectively choose, appropriate, implement and integrate them into the organization. Innovative IT artifacts extend the boundaries of known applications of IT, addressing important problems heretofore not thought to be amenable to computational approaches. Knowledge of how to effectively choose, appropriate, implement and integrate them into an organization unlocks their value.

Research in information systems lives at the interface of people, organizations and IT (Lee, 1999). People within organizations conduct business. Managers design organizational and inter-organi- zations structures within which their employees and partners develop and execute business plans, e.g., they market, produce and sell products, service customers, manage financial and human resources, and develop and manage partnerships. IT artifacts support people and organizations in accomplishing those tasks. The nature of that support is dependent upon the capabilities of the IT artifacts engaged. The success or failure of that support is dependent upon both the capabilities of the IT artifacts engaged and the designed and

Designing Design Science 339

emergent phenomena that occur when such capabilities are appropriated within a given organizational context.

Consequently, research in information systems has developed along two lines. One is engaged in understanding the nature of the phenomena that occur when information-technology artifacts are used (or misused) by people within organizations to accomplish human and organizational goals. This research has a behavioral focus. It develops and tests theories that explain or predict human and organizational phenomena that occur within the context of specified IT artifacts (capabilities). Such reasoned explanations and predictions ultimately enable people and organizations to more effectively and efficiently use IT artifacts. The second is engaged in developing innovative IT artifacts that address tasks and problems previously thought not to be amenable to computational solution. This research has a design focus. The artifacts developed in such research demonstrate new capabilities that can be appropriated by people and organizations to achieve their goals.

Organizations implement information systems to improve performance. Business managers commonly view performance through an economic lens, defining its goals relative to maximization of firm value, i.e., the present value of the difference between revenues and expenses. Given the high cost of information technology investments managers are understandably asking questions like: ‘Why do information technology investments not result in an increase in firm value?’ and ‘What can be done to maximize the likelihood that they will do so?’ These are the fundamental questions that IS researchers must answer (Benbasat and Zmud, 1999; 2003). The first is a causal question. The second is a problem-solving, decision making question. IS researchers must address both.

Answering the first question requires an understanding of phenomena that occur at the intersection of organizations, people and information technologies—the locus of the information systems discipline (Lee, 1999). Researchers must develop and justify theories, deep principled explanations of these phenomena (March and Smith, 1995; Hevner et al., 2004; Benbasat and Zmud, 2003). These are valuable in understanding what has happened and why it happened, and possibly in anticipating what will happen.

While posed theories may be explanatory rather than causal in nature their relevancy and value are determined by the degree to which they enable managers to create organizational work-systems that improve organizational performance (Alter, 2003). This is the focus of the second research question. Answering it is fundamentally a design task that requires managers to ‘shape artifacts and events to create a more desirable future’ (Boland, 2002). Organizations and the

340 Designing Design Science

information technologies that support them are both artificial, designed artifacts (Simon, 1996; Bunge, 1985). Their characteristics are molded to achieve specific purposes. They are ultimately evaluated by their effects on firm value. To the degree that organizations can shape behavior, e.g., through training and reward structures, people, although certainly not artificial, also have an element of design.

Much of the IS literature has explored characteristics of organizations and people while ignoring salient characteristics of information technologies (Orlikowski and Iacono, 2001; Weber, 2003). As a result much of this research is equivocal. For example, Orman (2002) eloquently describes the folly of posing a theory to explain or predict the effects of information technology on organizational structure without including characteristics of the implemented IT artifact. Any such theory must include the combined characteristics of people, organizations and information technology.

Social scientists and organizational theorists have long studied characteristics of people and organizations. Cognitive, behavioral, organizational and economic theories that explain and predict human and organization behavior abound. It is not in the realm of information systems research to pose new theories in these areas, although certainly there are IS researchers who are also interested in the foundational disciplines such as psychology, sociology, anthropology, economics and linguistics. IS researchers use and specialize such theories in the context of extant IT artifacts that have specifically designed capabilities to further our understanding of these phenomena.

Similarly, it is the role of computer scientists and engineers to create new and innovative IT capabilities. It is not in the realm of IS research, although there are IS researchers who are also interested in computer science and engineering. However, IS researchers must engage in characterizing those capabilities and using them in the context of improving firm performance. Doing so requires a combination of design science and behavioral science paradigms in crafting a research agenda for the IS discipline.

The science of design, as posited by Simon (1996), builds upon the mathematical and computational modeling capabilities developed in applied mathematics and operations research. Fundamentally it is a modeling discipline that develops and utilizes constructs, models, methods and implementations to understand, evaluate and optimize the characteristics of a system. In IS research the system under study includes organizations, people and IT artifacts. Unfortunately IS researchers using this paradigm often ignore salient characteristics of organizations and people and focus exclusively on characteristics of the IT artifact. Furthermore, in the ken of ‘Comprehensive Anticipatory Design Science’ as posited by R. Buckminster Fuller (1992) this

Designing Design Science 341

research may appear to ascribe simplistic and nearly magical capabilities to technological artifacts.

The function of what I call design science is to solve problems by introducing into the environment new artifacts, the availability of which will induce their spontaneous employment by humans and thus, coincidentally, cause humans to abandon their previous problem-producing behaviors and devices. For example, when humans have a vital need to cross the roaring rapids of a river, as a design scientist I would design them a bridge, causing them, I am sure, to abandon spontaneously and forever the risking of their lives by trying to swim to the other shore.

R. Buckminster Fuller from Cosmography (http://www.bfi.org/designsc.htm)

Introducing IT artifacts into an organizational environment may not cause people to ‘abandon their previous problem-producing behaviors and devices’. One of the important IS research problems is understanding if and how well people will, spontaneously or under duress, appropriately or inappropriately, use IT artifacts introduced into an organizational environment (Venkatesh et al., 2003; Goodhue and Thompson, 1995). However, the capabilities of IT artifacts enable managers to design work systems, organizational structures and job characteristics, and develop human resources to accomplish organizational goals. Managers must be cognizant of the capabilities of existing and emerging IT artifacts. Design science as a research paradigm in IS focuses on building and evaluating innovative IT artifacts that extend the boundaries of known applications of IT, addressing important problems heretofore not thought to be amenable to computational approaches (Markus et al., 2002; Walls et al., 1992).

IT artifacts are broadly defined as constructs, models, methods and instantiations (March and Smith, 1995). Constructs are vocabulary and conceptualizations that enable communication and description of problems (phenomena), solution components, constraints and objectives. Models use these constructs to represent problems and solutions. Methods are algorithms or guidelines that enable the construction of instantiations—computer-based systems implemented within an organization. Each may constitute a contribution to research knowledge. Constructs, models and methods are evaluated with respect to their ability to improve performance in the development and use of information systems. Instantiations or implementations demonstrate the feasibility of utilizing information technology artifacts for a given task. They often embody newly developed

342 Designing Design Science

constructs, models and methods. They are evaluated with respect to their effectiveness and efficiency in the performance of the given task.

Artifacts are intentionally designed for a specific purpose within a specific context. The purpose defines the criteria against which the performance of the artifact is evaluated. The context or environment defines the constraints in which the artifact must operate. Artifacts have inherent constraints that define their capabilities. The process of design is adapting the internal environment of the artifact to the external environment of the task domain to achieve its purpose (Simon, 1996). Difficulties in designing IT artifacts lie in the complexity, uncertainty and evolving nature of the external environment and in the relative inflexibility of the internal environment (Brooks, 1996). Computational systems rely on software that implements specified algorithms and on data acquired from the environment. When contingencies arising in the environment are not adequately addressed by the algorithms or when data are outside the design capabilities of the artifact, the system typically fails. While fault-tolerance and adaptive-systems are areas of current research, in general, software systems are simply not reliable outside their design parameters (Leveson, 2005). They may be designed to ‘fail gracefully’, but fail they will.

Computational modeling and the implementation of computer programs are fundamental to this paradigm. In modeling and constructing innovative IT artifacts, researchers gain significant new knowledge and demonstrate the feasibility of the studied capability to address heretofore unsolved organizational problems. These, combined with an understanding of people and organizations, hold the promise of research contribution for the IS discipline.

Consider, for example, the AIS Special Interest Group on AgentBased Information Systems (SIGABIS). It purpose, to advance ‘knowledge in the use of agent-based IS .. . to improve organizational performance’ (SIGABIS, 2005) has elements of both lines of research and must engage both behavioral science and design science methods. It is crucial for researchers engaged in this area of research to recognize this dual research focus.

Such research is analogous to traditional research in management science. Neither is specifically concerned with developing new computational methods or solution procedures. These are left to computer scientists and applied mathematicians. However, both are concerned with solving heretofore unsolved management problems using computational models and methods. Solving managerial problems engages both design and behavior. Managers create business strategies and design the business processes needed to implement them. The designed business processes combine people, organizational

References 343

structures and IT artifacts in ways aimed at accomplishing organizational goals.

Representing the emergent behavior of inter-organizational business systems is extremely complex as is representing design decisions and evaluating their consequences. IS researchers engage behavioral and economic theories within the process of generating, predicting and evaluating the performance of alternative system designs, specifically studying the role and affects of IT artifacts.

The agent metaphor, for example, is well-suited to the representation of the varying and possibly conflicting objectives of internal and external stakeholders, enabling managers to better understand the consequences of design decisions reflected in its likely emergent behavior. These representations or models are essentially computational theories—reasoned explanations or predictions of phenomena that occur when IT artifacts are used to construct artificial worlds in support of real-world activities. However, building an agent-based model of organizational phenomena is not necessarily a research contribution. A research contribution must (1) identify the uniqueness presented by an agent-based approach for the real-world problem, (2) clearly explain why there is any question about its applicability to the problem and how the approach extends existing conceptualizations of computational capabilities, and (3) demonstrate that the approach more faithfully represents the real world and that this increased modeling fidelity significantly affects decisions taken by managers. Lacking these agent-based models become ‘exercises in programming’, perhaps very useful to practitioners but not constituting a research contribution. Adequately addressing these enables the researcher to formulate and address important research questions adding to the breadth and depth of knowledge in the IS discipline.

REFERENCES

Alter, S. (2003) ‘18 reasons why IT-reliant work systems should replace “the IT artifact” as the core subject matter of the IS field’, Communications of the AIS (12), October, pp. 365–94.

Benbasat, I. and Zmud, R. W. (1999) ‘Empirical research in information systems: The practice of relevance’, MIS Quarterly, 23(1), March, pp. 3–16.

Benbasat, I. and Zmud, R. W. (2003) ‘The identity crisis within the IS discipline: Defining and communicating the discipline’s core properties’, MIS Quarterly, June, pp. 183–94.

Boland, R. J. (2002) ‘Design in the punctuation of management action’, in R. Boland (ed.), Managing as Designing: Creating a Vocabulary for Management Education and Research, Frontiers of Management Workshop, Weatherhead School of Management, 14–15 June 2002 (available at http://design.cwru.edu).