Table 2 shows the empirical papers in the sample coded for research paradigm. The papers were coded as positivist, interpretivist, critical, or mixed. We followed the approach of Chen and Hirschheim (2004) and only coded empirical papers for paradigm. Only one paper, in Personal DSS, was coded as mixed. Surprisingly, no DSS paper in the sample adopted a critical paradigm. The analysis period saw a significant increase in non-positivist research in IS with an increasing presence of interpretivist case studies in the highest quality journals. Table 2 shows that DSS research is overwhelmingly dominated by the positivist paradigm with 92.3% of empirical studies following that approach. Chen and Hirschheim’s (2004) study of general IS research from 1991 to 2001 reported that 81% of papers had a positivist orientation with 19% using an interpretivist approach. Thus, DSS research is more dominated by positivism than general IS research and DSS researchers have been more conservative than their general IS colleagues in embracing philosophical diversity.

Arnott and Pervan (2005) found that only 9.6% of DSS papers were of high or very high professional relevance. One strategy for improving the relevance of DSS research is to increase the number of case studies, especially interpretive case studies. Put simply, a field that is so removed from practice needs case study work to ensure that the questions it is addressing are both relevant and important. Interpretive case studies can illuminate areas of contemporary practice in ways that natural science-like studies such as laboratory experiments and surveys cannot (Cavaye, 1996; Eisenhart, 1989). Importantly, they can inspire researchers to focus on issues of current importance and build lasting links between academics and senior professionals, a process that will assist with grant funding as well. Table 2 shows that data warehousing and EIS/BI have the highest proportion of interpretivist studies (although the number of DW papers is probably too small to make firm conclusions), while intelligent DSS and personal DSS have almost ignored non-positivist paradigms. It is interesting that the more modern types of DSS are being researched with a more contemporary mix of paradigms than older types of DSS. Further analysis of the interpretivist studies reveals that almost all are focused on the theory development stage of research, thus confirming their importance in developing new theory in DSS.

Table 2 also shows the dominance of the oldest types of DSS in the agendas of researchers. While DW and EIS/BI have been mainstream in practice since the mid 1990s they only account for 9% of empirical papers (8.4% of all papers). Arnott and Pervan (2006) confirmed this dominance of the oldest aspects of the field.

Table 3 shows an analysis of the sample using the IS research classification developed by Alavi and Carlson (1992) and revised by Pervan (1998). The highest level of the classification divides papers into empirical and non-empirical categories. The lowest level addresses research designs and methods. Table 3 shows that around one-third (33.6%) of DSS research is non-empirical, with the remaining two-thirds (66.4%) being empirical. Chen and Hirschheim’s (2004) analysis of overall IS research reported a different split between non-empirical (40%) and empirical (60%) research, showing that DSS research has significantly more empirical research than general IS. The most popular single research method since 1990, using the Alavi and Carlson (1992) taxonomy, has been the laboratory experiment. This, in part, reflects the methodological focus of North American business schools.

What is noteworthy in Table 3 is the 21% of papers in the empirical-objects categories. DSS was founded on the development of experimental systems for managers and has a long history of publication of descriptions of DSS applications that are novel or important. This is part of what is now called design science (Hevner et al., 2004). There could also be a significant amount of design-science research in the ‘Tools, Techniques, Methods, Model Applications’ and ‘Conceptual Frameworks and Their Application’ article types. As a result, design science could be the largest major category of DSS research. DSS researchers have much to offer the current debate on IS design science methodologies; it may even be the most significant contribution that DSS can make to its parent discipline.