هشت مسئله کلیدی برای انضباط سیستم های پشتیبانی تصمیم گیری

This paper integrates a number of strands of a long-term project that is critically analysing the academic field of decision support systems (DSS). The project is based on the content analysis of 1093 DSS articles published in 14 major journals from 1990 to 2004. An examination of the findings of each part of the project yields eight key issues that the DSS field should address for it to continue to play an important part in information systems scholarship. These eight issues are: the relevance of DSS research, DSS research methods and paradigms, the judgement and decision-making theoretical foundations of DSS research, the role of the IT artifact in DSS research, the funding of DSS research, inertia and conservatism of DSS research agendas, DSS exposure in general “A” journals, and discipline coherence. The discussion of each issue is based on the data derived from the article content analysis. A number of suggestions are made for the improvement of DSS research. These relate to case study research, design science, professional relevance, industry funding, theoretical foundations, data warehousing, and business intelligence. The suggestions should help DSS researchers construct high quality research agendas that are relevant and rigorous.

Decision support systems (DSS) is the area of the information systems (IS) discipline that is focused on supporting and improving managerial decision-making. Essentially, DSS is about developing and deploying IT-based systems to support decision processes. DSS has been an important area of IS scholarship since it emerged in the 1970s. It has also been a major area of IT practice and the decisions made using IT-based decision support can have a significant effect on the nature and performance of an organization. The current DSS industry movement of business intelligence (BI) is one of the most buoyant areas of investment despite the IT downturn of the early to mid 2000s. The market in new BI software licences grew 12% from 2003 to 2004 and is expected to have compound growth of 7.4% to 2009 [34]. DSS is not a homogenous field and over its 35-year history a number of distinct sub-fields have emerged. The history of DSS reveals the evolution of a number of sub-groupings of research and practice [6]. The major DSS sub-fields are: • Personal Decision Support Systems (PDSS): usually small-scale systems that are developed for one manager, or a small number of independent managers, to support a decision task; • Group Support Systems (GSS): the use of a combination of communication and DSS technologies to facilitate the effective working of groups; • Negotiation Support Systems (NSS): DSS where the primary focus of the group work is negotiation between opposing parties; • Intelligent Decision Support Systems (IDSS): the application of artificial intelligence techniques to decision support; • Knowledge Management-Based DSS (KMDSS): systems that support decision making by aiding knowledge storage, retrieval, transfer and application by supporting individual and organizational memory and inter-group knowledge access; • Data Warehousing (DW): systems that provide the large-scale data infrastructure for decision support; • Enterprise Reporting and Analysis Systems: enterprise focussed DSS including executive information systems (EIS), business intelligence (BI), and more recently, corporate performance management systems (CPM). BI tools access and analyze data warehouse information using predefined reporting software, query tools, and analysis tools [63]. Of these sub-fields, PDSS, Enterprise Reporting and Analysis Systems, and DW have had the most presence in practice. This paper integrates a number of strands of a long-term project that is critically analysing the academic field of DSS. The foundation of the project is the content analysis of 1093 DSS articles published in 14 major journals from 1990 to 2004. The first, descriptive, results were presented in Arnott, Pervan and Dodson [8]. Pervan, Arnott, and Dodson [72] presented a critical analysis of group support research from 1990 to 2003, while Arnott, Pervan and Dodson [7] analysed the funding of all types of DSS research. Pervan and Arnott [71] examined data warehousing and business intelligence research and Dodson, Arnott and Pervan [22] analysed the role of the system's client and user in DSS research. The major publication from the project to date, Arnott and Pervan [6], analysed published research in a number of dimensions including journal publishing patterns, research paradigms and methods, decision support focuses, professional relevance, and judgement and decision-making foundations. The aim of this paper is to integrate the findings of the strands of the project into a set of key issues that can assist DSS researchers in the development of research agendas that are important for both theory and practice. This paper is structured as follows: first, the project's research method and design is described. This is followed by the identification of the key issues that have emerged from the various strands of the project. Each key issue is discussed in turn, following which suggestions for the improvement of DSS research are made, the limitations of the research outlined, and the future directions of the project are discussed.

The analysis of the eight key issues constitutes a cause for reflection, revision, and evolution of DSS research agendas. Before suggesting some directions for DSS research, a word of caution about the findings is warranted as the eight key issues can be viewed in a slightly negative way. This is because the intention of the analysis was to illuminate problems in the field so that we may change our research behaviour in a way that significantly improves our work. As Tversky and Kahneman [85] have found, a negative frame can bias the perception of a decision or task. It should be remembered that despite its current problems, DSS has a long history of success in scholarship and practice. BI and PDSS systems are now an integral part of most managers' work. The idea that computers can be used to support rather than replace humans is as important today as it was in the 1970s. DSS scholars have contributed significantly to IS theory in areas such as evolutionary systems development, the incorporation of AI into business systems, multi-dimensional data structures, critical success factors, group processes, and managerial information behaviours. Nevertheless, the eight issues identified in this paper should be given careful attention. The key issues are summarized in Table 11. 4.1. Suggestions for improving DSS research Our suggestions for improving DSS research are framed by a long-term issue in IS research: the tension between academic rigor and professional relevance [3], [4], [10], [15], [29] and [44]. For most of our analysis period the emphasis of IS research has been on achieving rigor. This emphasis was appropriate for a new discipline and much has been written about the need for IS to be accepted as a valid discipline in universities. Benbasat and Zmud [10] argue that the IS discipline is now relatively mature and it “can afford to shift attention to relevance without undue concern about being criticised by others…” (p. 7). We start this section with suggestions for increasing research relevance because we believe that relevance is the area in most need of improvement. The first strategy for improving the relevance of DSS research is to increase the number of case studies, especially interpretive case studies. As was discussed in Key Issue 2, case studies are the research papers with the highest proportional relevance scores. Case studies can illuminate areas of contemporary practice in ways that studies such as laboratory experiments and surveys cannot [14] and [23]. A field that is as removed from practice as is evident from the analysis of Table 2 needs case study work to ensure that the questions it is addressing are both relevant and important. Researchers need to select problems with a consideration for professional relevance and interest, in addition to considering the recommendations of previous academic research. When considering professional-related problems, researchers need to think about the likely relevance of their work three to five years after the start of a project, a common time period from initiation of a project to publication in a good journal. Lee [48] argues that the dominant positivist approach in IS research has adversely affected the relevance of the field. He argues that interpretive and critical social theory investigations are needed to develop deep understandings of professional practice. Because DSS research is more dominated by positivism than general IS research, Lee's call for broadening the approaches to case study research is particularly important for DSS scholarship. Further, in these new areas, practice can (and perhaps should) lead research and provide opportunities for researchers to inductively build new theories using approaches based on the interpretive and critical paradigms. Importantly, contemporary case studies can inform DSS education. By its nature, case study research can build lasting links between academics and senior professionals and executives. This can in turn assist researchers with obtaining funding from industry. Industry funding is essential for two reasons. First, as mentioned above, researchers can no longer rely on non-competitive internal university funds to fully support their research. Further, major competitive grant funding is becoming more difficult to win in many countries. The second reason why industry funding is desirable is that it increases the commitment of all parties to a research project. When an industry partner contributes funds they treat the project more seriously and often researchers have access to more senior personnel than is the case for non-funded projects. The additional pressure to perform that is placed on researchers by accepting industry funding can improve the quality of research. As a result, successfully seeking industry funding is our second strategy for improving DSS research relevance. The third strategy for improving discipline relevance is the conduct of high quality design science research, for example [19], [41], [47], [55] and [69]. As was mentioned in the discussion of Key Issue 2, DSS has a long tradition of design science, although the proportion of design science research has been declining. Design science represents a significant escalation of industry involvement over case studies and requires different skills in researchers. Importantly, design science has the potential to influence, even lead, industry practice in ways that other research methods can't emulate. Lyytinen [50] identifies two types of relevance—the first is research that is quickly and easily digestible by a CIO, and the second is research that can “elevate and reshape professionals' thinking and actions in a longer perspective” (p. 26). Design science research, when it is properly grounded in relevant high-quality theory, has the potential to achieve the deeper concept of relevance associated with reshaping professional ideas. DSS design science also has the opportunity to embrace non-positivist approaches to design science. The early IS design science papers assumed a positivist stance [51] and [86]. More recent contributions have added to the understanding of design science. For example, Carlsson [13] proposed a design science method based on critical realism and McKay and Marshall [54] proposed a design science process that extends IS researcher's experience with action research. A large part of the improvement of DSS research relevance could come from shifting research agendas towards the effective development and deployment of data warehouse and business intelligence systems. This shift may not be as radical as may be first thought. Most of the research problems that are the focus of other DSS types can be made relevant to BI and DW. These problems include development methodology, system usability, organizational impact, technology adoption, project success and failure, project evaluation and approval, and IT governance. The benefits of increased relevance, funding, quality, and professional influence far outweigh the cost of the agenda change. The second theme in improving DSS research that arises from the analysis in this paper concerns the rigor of our work. While relevance is the theme that needs greatest attention, academic rigor needs to be central to research designs and the average rigor of DSS research needs to be improved. Further, academic rigor is what many professionals value in IS research, particularly when academic studies are compared with commercial research reports and vendor white papers. One indicator of a rigor problem is the field's relatively low success rates in ‘A’ journals other than DSS. The low success rate could be biased by the influence of Management Science in the sample. Management Science is a multidisciplinary journal that publishes a relatively large number of papers of which only 2.1% are DSS. Another reason why DSS may be underrepresented in ‘A’ journals is its large proportion of design science research. While design research is prominent in DSS research it is a small part of overall IS research [14]. Minority areas often find it difficult to make headway in ‘A’ journals. The recent decline in DSS design science publication could be due to researchers changing their research agendas to target projects that will achieve the prestigious publication they need for academic reputation. However, the prospects for design science research in ‘A’ journals are improving. The specification of quality guidelines for design science research by Hevner et al. [38] is an important step in this improvement. DSS design science research is being published in ‘A’ journals (for example, Markus, Majchrzak and Gasser [52] in a US ‘A’ journal, and Arnott [5] in a European ‘A’ journal). We encourage DSS researchers to maintain their interest in design science. However, this interest will only be successful if that work is rigorous and well executed. The second element of improving academic rigor identified in the analysis (Key Issues 3 and 8) concerns the theoretical foundations of the field. Around half of the papers in the sample are not founded on judgement and decision-making research, and those that are tend to be based on relatively old references. In general, DSS research needs to be based on more contemporary behavioural decision theory [for example, [31], [32] and [37]]. Other theoretical aspects of judgement and decision making could be imported from management and related fields to provide a stronger theoretical basis for projects. The current narrow base of reference theory may have acted to overly constrain what projects have been tackled by DSS researchers. In summary, we suggest that to improve DSS research, researchers should: 1. Undertake more case studies, particularly using an interpretive approach, 2. Continue the design science tradition of the field but pay greater attention to the rigor of projects, 3. Select research problems based on genuine long-term professional relevance, 4. Seek more industry funding for projects, 5. Pay greater attention to the effective development and implementation of data warehouse and business intelligence systems, 6. Update and broaden the theoretical foundations of projects with respect to judgement and decision-making. 4.2. Limitations No research study is free of limitations and this project has at least three areas of possible concern. First, this study reviewed a finite set of DSS articles (1093) but it could be argued that this number is large enough to support the validity of our conclusions. Second, conducting a literature review and coding the content on various dimensions is, of necessity, rather subjective. However, the rigor of the coding and analysis procedures used and the research experience of the researchers ensured that the data was fairly reliable. We believe that other researchers using our protocol would produce similar results. Finally, any large study of journal papers is dependent on the set of journals chosen. We chose a mix of general management science, information systems, and decision support systems journals. This set should be sufficiently representative of the field. We also included five European journals to provide an international mix that is generally absent from other studies. We did not include professional journals as our focus was on DSS research. 4.3. Further research We plan to continue the content analysis of DSS articles and produce a further overall analysis of the field around 2010. Based on the suggestions for improving DSS research made above, two further investigations of the intellectual foundations of DSS are well under way. The first is a critical review of DSS design science research using the guidelines developed by Hevner et al. [38]. The aim of this analysis is to provide prescriptions for improving the rigor and relevance of DSS design science research. The second project is a more detailed analysis of the judgement and decision-making foundations of DSS research with a special emphasis on the role of Simon's theory of behavioral decision-making has played in shaping the field. A third project that we wish to pursue is to investigate the management and organization theory foundations of DSS research.