Design science research is a well-established research framework to structure research work. There are a couple of variants of this framework, but they all share the same idea that we should solve current and relevant problems by constructing novel solutions to these problems while using appropriate research methods in both the identification of the problem and the evaluation of the proposed solution.
While this may sound benign, design science represents a significant advancement over both the old (pre-2000) years of innovation in software engineering, where folks often would not or only poorly evaluate their proposed concepts, tools, and methods, and the last twenty years, where folks often would only analyse existing data to draw conclusions, using appropriate research methods, but not really address hard problems with innovative solutions. Design science does both: Address socio-economic and technical problems of relevance, yet do so by providing a framework for using appropriate research methods.
My research group settled on Peffers et al. (2007) [1] as the variant of choice for design science research. The five main activities in design science research according to Peffers et al. are (1) problem identification, (2) objective definition, (3) solution design, (4) demonstration, and (5) evaluation.
If you don’t know much about design science, you might be tempted to assume that most of the work and correspondingly the value being created happens during activity 3, solution design. This is not so. To understand this, we first need to distinguish between research and innovation. Innovation is something new that promises, typically, to solve a problem with a new approach, the innovation. An innovation, becomes research, if it gets evaluated using appropriate research methods. Or, in other words: The innovation itself is a toothless tiger, it only becomes real when some research shows that it works. In design science research, the innovation happens during solution design, and the evaluation or validation happens during activity 5, evaluation. Often, the evaluation consumes significantly more effort (costs in terms of time and money) than the solution design.
You might also be tempted to think that solution design then is at least the second most important part of design science research. After all, that’s where the innovation happens. It may not have scientific merits, but it is where the socio-economic and technical value is being created. This is also not true. Most of the hard work of theory building happens earlier, during activity 1, problem identification. Here, the researchers enter the domain of interest, build up understanding, and codify it. Once they analysed the problem domain, they can select and define their research objective (the research question to answer). At this stage, in all our work, the possible solutions already had taken shape in the minds of the researchers. What’s left for activity 3, solution design, is to work out the details. A paper publishing only the solution design should be rejected, because it has no scientific merits of its own. It is not clear where it is coming from and whether the described solution does anything good.
There are two main types of publications in design science research. Problem identification often results in a useful (proposed) theory of the domain, and this can be published. Methods of choice are systematic reviews, interview studies, descriptive surveys, and the like.
Evaluation provides either the scientific evaluation of the theory or selected validation of hypotheses. In our line of work, the qualitative work of evaluation often uses action research or case study research, and the quantitative work of validation uses hypothesis-testing surveys or controlled experiments. Then another set of publications presents slices from the extended theory that combine problem identification, solution design, and evaluation of that slice.
To summarize: The scientific value in design science research is created during the problem identification and evaluation activities. The value is captured in publications either only about the proposed theory from problem identification, or about the evaluated theory using problem-solution-evaluation slices.
[1] Peffers, K., Tuunanen, T., Rothenberger, M. A., & Chatterjee, S. (2007). A design science research methodology for information systems research. Journal of management information systems, 24(3), 45-77.
Leave a Reply