Design science in public administration: producing both situational interventions and generic knowledge

Abstract

Design science is making a comeback in the discipline of public administration: inspired by the growing popularity of design thinking, various scholars have highlighted the value of design science for the public sector. This theoretical and methodological article aims to contribute to the academic value of design science and focuses on the “double burden” of design science: how can it be applied in such a way that it generates both (relevant) situational interventions and contributes to (robust) generic academic knowledge? We identify which types of generic academic knowledge can be generated through design science in public administration: theoretical knowledge, design exemplars, methodological knowledge, and normative knowledge. We use these four types of generic knowledge to develop a new perspective on design science. We also present guidelines for design research that aim to generate both situational interventions and generic knowledge. These guidelines emphasize the need to be more rigorous and systematic in the process of design research as to ensure the validity and reliability of the contributions to generic knowledge for the discipline of public administration.

THE “DOUBLE BURDEN” OF DESIGN SCIENCE

The application of design science in public administration is experiencing a new revival. Classic work has been conducted by Behn (1996) and Hood (1991), with newer studies applying design science to topics as different as public service design (Strokosch and Osborne 2023), organizational development (Van Aken and Romme 2009), policy design (Bason 2016), public participation (Moynihan et al. 2012), and new democratic arrangements (Romme et al. 2018) In the past 10 years, we have been seeing a new thrust of publications on design-oriented public administration (Barzelay 2019; Barzelay and Thompson 2010; Bason 2010, 2017; Brinkman et al. 2023; Dorst 2015; Dorst et al. 2016; Howlett 2014; Lewis, McGann, and Blomkamp 2020; Romme and Meijer 2020; Van Buuren et al. 2020). The current state of thinking about design in public administration has been summarized in a recent literature review by Hermus, Buuren, and Bekkers (2020), with this review highlighting the need to acknowledge the value of design science as one of the important research methodologies in public administration research.

Current approaches to design science often fall short of indicating precisely how this research contributes to generic academic knowledge for the public administration discipline (for exceptions: Romme and Meijer 2020; Van Aken and Romme 2009). Hevner et al. (2004: 81) write: “the key differentiator between routine design and design research is the clear identification of a contribution to the archival knowledge base of foundations and methodologies.” Argyris and Schön (1989: 613) refer to this as the “double burden” of design experiments: these experiments need to test hypotheses and contribute to academic knowledge but also effect some desirable change in the situation. This means our academic community needs strong design science methods to balance the ambition of producing useful knowledge for public administration practice—”field problems” (Van Aken and Romme 2009)—with the goal of developing generic knowledge for the public administration discipline—”pure knowledge problems” (Van Aken and Romme 2009). For this reason, design science can play a key role in bridging the “academic-practitioner divide” (Ancira, Rangarajan, and Shields 2022).

This article aims to tackle the “double burden” of design science by developing a systematic understanding of design science as a knowledge process, and presenting guidelines for design science in public administration research that tackles both “field problems” and “pure knowledge problems.” The importance of tackling this “double burden” is well illustrated with an example presented by Bason (2010). In examining design thinking, Bason (2010: 135) highlights how the City Council of Sunderland (UK) used design thinking to tackle the problem of unemployment: “building upon design approach such as ethnographic research, service journeys, fast experimentation and prototyping, [the researchers of] LiveWork helped the city council identify a range of possible solutions that could get people more efficiently on a path back towards work.” With this study, the contribution to “field problem”—unemployment in Sunderland—is clear but the contribution of this research work to “pure knowledge problems” is not discussed. One can imagine that these experiences can provide insights relevant to other councils in the UK or abroad, that lessons can be drawn about the normative assessment of various interventions and these experiences can also contribute to methods of design research. As is often the case, this “feedback” to pure knowledge problems for public administration is not realized. Thus, this example illustrates the need to develop a thorough understanding of design research as a knowledge utilization and knowledge production activity.

Currently, our academic understanding of the relations between the various knowledge types that play a role in design science is limited and fragmented, while this understanding is needed as a basis for debating the “double burden.” By elaborating on design science as a knowledge process, this article makes three contributions to the literature on research methods in public administration research. First, the article analyzes the literature to identify four different types of generic knowledge that can be both used for and constructed in design research. Second, it presents an overarching perspective on the process of building upon generic types of knowledge and constructing both new types of generic knowledge and specific situational interventions. Third, it presents methodological guidelines that complement current guidelines for design research by elaborating how the four types of generic knowledge can be used for and constructed in design research.

By providing both a conceptualization and methodological guidelines, this article contributes to the literature on methodologies for public administration research and presents design science as a methodology that is specifically suitable for doing research that aims to generate situational interventions and generic knowledge.

WHY DO DESIGN SCIENCE IN PUBLIC ADMINISTRATION RESEARCH?

The current attention for design science builds upon earlier calls for attention to design as an inspiration for public administration.¹ An important literature review that covers both the roots of current ways of doing design research and also the state of the art of this field was conducted by Hermus, Buuren, and Bekkers (2020). They highlight the variety of the field and identify no less than six design approaches, varying from traditional scientific and informational approaches to innovative, user-driven and what they call “more inspirational” approaches. The design science approach to research is gaining traction and is being developed at various academic institutes around the world, necessitating the need to reflect on it in relation to public administration and public management research.

For the discussion of the “double burden” central to this article, we focus on the question of how scholars have provided legitimacy to design science in public administration research, even though this approach does not fit within dominant approaches for empirical research (even if these are broadly conceived). Design research is not about producing a justified belief or understanding of a particular phenomenon (McCain 2016), but about producing solutions for problems. There has been, and is, a vivid debate about the question whether the production of solutions should indeed be the focus of academic work.

Herbert Simon (1969) is most often seen as the founding father of design science. In his book “The Science of the Artificial” he highlights that studying how the world can be—the artificial—is a valid ambition of academic work and requires a different research approach. Not much later, Vincent Ostrom (1974) stresses the validity of Simon’s argument and added that we need an appropriate theory of design to understand how modifications to a system will affect its performance. These two classical thinkers developed the foundations for an academic approach to prescription rather than only description and explanation. In doing so, they provided a strong academic basis for not only studying current situations (“as is”) but also studying how these situations can be improved through targeted interventions (“to be”).

A specific translation of these ideas about design science to the field of public administration was made 15 years after Simon’s foundational book by Shangraw & Crow (1989) (later re-visited: Crow and Shangraw 2016). They built upon the ideas of Simon and Ostrom and stressed that public administration should be understood as a design science since the ambition is not only to produce a theoretical understanding of public administration but also to intervene in social reality. In the nineties, Behn (1996: 97–120) argued public managers behave not like scientists but like engineers since they rely on their own judgement and spontaneity. They iterate, modify, experiment, test, and redesign through a process of trial and error, allowing managers to learn from what works, discard what does not work, and modify what can be improved. This calls for an approach to public administration that acknowledges the specific nature of the field and the differences with social sciences with no or limited prescriptive ambitions: a design-oriented public administration (Overman 1989; Walker 2011).

The case for public administration as a prescriptive discipline is made by Michael Barzelay and Fred Thompson (2010) and later Barzelay (2019). They do not only refer to Herbert Simon’s “sciences of the artificial” (1969) but also to Dwight Waldo’s (1952) emphasis on normative, deliberative reasoning and Eugene Bardach’s (2004) understanding of the process of designing practical interventions. Public administration in its early days was an administrative job—its aim was giving practical, useful advice for converting existing states into preferred ones. This meant that a very different approach was used than to that which we are now accustomed in our discipline. This approach seems more similar to what Flyvbjerg (2001) refers to as phronesis (defined as INSERT), which entails designing change mechanisms and artifacts to help practitioners make the most of their cognitive capabilities when facing practical challenges (Bason 2017; Dorst 2015). Instead of identifying general principles or laws, scholars of public administration worked towards crafting a set of heuristics for designing practical interventions. This strengthened the societal value of the research but does risk interfering with the ambition of also developing more generic knowledge.

Whether there is a tension between situational and generic knowledge depends on the social science perspective. Academics who focus on randomized clinical trials stress that they use individual design experiments to provide evidence for the effectiveness of generic interventions. In this perspective, generic knowledge forms the basis for generic interventions that can be applied over a range of different contexts (for an overview: Stoker and John 2009). In most approaches to design in the public sector, however, the importance of contextuality is highlighted and then the value of a specific design is highly dependent on the extent to which it fits the situational context (Barzelay 2019; Bason 2017; Dorst 2015,Hermus, Buuren, and Bekkers 2020; Romme and Meijer 2020). In this perspective, the relationship between situational and generic knowledge is less straightforward and more fundamental reflection on the nature of the knowledge produced in situational context is needed.

Situational knowledge refers to the (holistic) knowledge needed to act in a specific situation. Flyvbjerg (2001) builds upon Aristotle to argue that knowledge about how to act in a specific situation should also be considered as academic knowledge but that this type of knowledge is different from scientific theories that aims to identify generalizable patterns. Following Aristotle, he uses the term “phronesis” to refer to the cases and “episteme” as a label for what is generally considered to be scientific knowledge. Phronesis is a context-specific type of knowledge that is also integrated in nature and taps into different knowledge reservoirs. This situational knowledge plays a key role in design processes that aim to develop insights about the specific problem history, the nature and manifestation of the problem, and the empathy for the various actors in a specific problem situation as a basis for the design rather than focusing on issues of generalizability. At the same time, situational knowledge taps into different forms of generic knowledge and the use of generic insights is encouraged in design processes (Bason 2016).

This bring us to the concept of generic knowledge. This article touches upon the fundamental debate on the nature of scientific knowledge: the demarcation problem (Resnik 2000). This debate is central to the philosophy of science and can never be adequately represented in a single article. A rich discussion of this topic can be found in McCain (2016), who argues that scientific knowledge can be regarded as a justified belief or understanding of a particular phenomenon based on evidence that meets scientific criteria. This type of knowledge is certainly important in public administration, with this knowledge in public administration characterized as theoretical knowledge, with well-known examples such as network theory, agenda-setting theory, and configuration theory. However, scientific knowledge is not the only type of knowledge that constitutes the body of generic knowledge in our discipline. Additional types of knowledge are methodological knowledge, normative theory, and the body of cases relevant to a discipline. These knowledge types are mostly separate from what is considered scientific knowledge since they do not present representations of the world as it exists but methods for generating knowledge, knowledge about the world as it could be and specific manifestations of the world in context. Therefore, generic knowledge as used in this article is a broader concept than scientific knowledge. Overall, generic knowledge—as a contrast to situational knowledge—needs to be understood as the set of generic theories, principles, insights, and methodologies that form the basis for the academic debates and teachings in a discipline. This generic knowledge plays an important role in situational interventions, but the methodological question addressed in this article is how situational interventions based on design research can also contribute to generic knowledge in the discipline of public administration.

The debate about the relations between practical interventions and generic knowledge links directly to the debate about improving and understanding public administration. Van Aken (2004: 220; 2005: 22) writes that design research in the management sciences can be seen as a quest for understanding and improving human performance. Hevner et al. (2004) make an argument in the Information Sciences Discipline that seems to equally apply to Public Administration, with two paradigms characterizing the research: empirical sciences (which we can understand as a combination of the neo-positivistic and interpretative perspectives above) and the design sciences (for a similar argument: March and Storey 2008: 725). The empirical sciences aim to develop an understanding of the world as it “is,” whereas the design sciences strive to develop knowledge about the world as it “ought” to be. They stress that these paradigms have different knowledge ambitions—truth versus utility—and that both are complementary (see also Van Aken (2005: 21) for a similar distinction between descriptive and prescriptive knowledge).

Even though there has been quite some epistemological debate about design science, current presentations of methods for design science in the public sector pay limited or no attention to these debates and the “double burden” or research. Bason’s (2017) work on leading public sector design provides a strong introduction of the topic to academics, students, and professionals and highlights key elements of design research: exploring the problem space, generating alternative scenarios and enacting new practices. Concrete methods are presented that include ethnographic research, visualization, patterns recognition, ideation, concept development, prioritization, prototyping, user testing, and realization. These methods are well presented but a discussion of how design research can contribute to “knowledge problems” is lacking.

Three examples of recent design research in public administration highlight how research has been dealing with the double quest of public administration—improving public administration practices in situational contexts and contributing to a more generic understanding of public administration—and they show the variety of topics addressed and theoretical perspectives used in this type of research. Ruijer’s (2021) research focuses on data collaboratives in the public sector. Data collaboratives are collaborations between various actors to share and process relevant information about a societal problem. Sharing data is not only a technical issue but also an organizational and political challenge that requires building collaborative relations of trust. Ruijer (2021) designed a data collaborative in a specific city in the Netherlands to share policy-relevant information and generate new insights as a basis for improved policies. This research was then used to generate generic insights about how public organizations work across boundaries and sharing data to address complex problems.

Romme et al’s (2018) design research aims to stimulate democratic innovation in a municipality in the Netherlands. The key problem they address is the high distrust in political institutions and a growing sense of powerlessness among many citizens. They use circular design research, understood as design approach for realizing a dynamic equilibrium, to strengthen the local democratic governance system’s capability for collective dialogue and learning. This research produces concrete interventions in this municipality—for example, by identifying bridge builders and organizing more informal meetings—to create a more cyclical practice of forming ideas, judgments, and decisions rather than only inviting citizens to participate when ideas have already been developed. In parallel, the research contributes to academic debates about local democracy by providing academic insights about the transformation of low trust situations into consent-based cultures of collaboration.

Finally, Van Hout et al.’s (2024) design research addresses the problem of scaling up public innovations. Scaling up innovations is a key problem for many public organizations since experiments have become ubiquitous but very often these experiments do not result in large scale changes. This research investigates this problem in the context of the Netherlands Ministry of Infrastructure and results in an instrument that this ministry can use for managing their process of scaling up innovations. At the same time, this research produces a range of generic insights into the complexities of scaling processes and thus contributes to theories about innovation in the public sector.

These three examples all highlight how design research generated both knowledge to tackle specific problems—related to data collaboration, local democracy, and scaling public innovations—and contributions to theoretical debates in the scientific community—on datafication of local government, on institutional arrangements for local democracy, and on scaling processes of public innovations. These three design research projects resulted in presentations for local audiences, reports, and organizational and governance interventions, but also in academic presentations at international conferences and scientific papers in peer-reviewed journals. Somehow, these research projects managed to produce both situational knowledge for design interventions and generic knowledge for academic debates. In the next section, we analyze more specifically how we can understand the relations between the situational interventions and the different types of generic knowledge.

FOUR TYPES OF GENERIC KNOWLEDGE IN DESIGN SCIENCE

To analyze how design science can deal with the “double burden,” we need to make clear how knowledge is being used and produced in design research: various generic knowledge inputs and outputs of design science. The different types of knowledge are mentioned in scattered way in the literature and no overarching perspective on the process of building upon generic types of knowledge and constructing new types of generic knowledge has been presented in the literature. On the basis of our extensive reading of the literature on design research and extensive discussions with various international colleagues, we have identified four types of generic knowledge that are connected to design research: theoretical knowledge, methodological knowledge, normative knowledge, and a knowledge base of case studies. Inspired by Hevner et al.’s (2004: 80) Information Systems Research Framework, we have developed a heuristic model of the relations between the various knowledge types (see Figure 1).

This heuristic model can contribute to a more systematic approach to design research in public administration by explicitly highlighting the various choices in the use and production of generic knowledge in design science. For this reason, we introduce the various knowledge types in more detail and discuss the relations with the four different types of generic knowledge, illustrating each one of them with an example.

Building Upon and Contributing to Theories of Public Administration

A first type of generic knowledge that can be used in design science concerns theories from the field of Public Administration. Due to its focus on the object of study and not one perspective, Public Administration is a fragmented field of study with a variety of theories that Bryer (2021) categorizes in four groups: (1) theories on the role of public administration (e.g. social justice), (2) theories on the function of public administration (e.g. collaborative governance), (3) theories on the people in public administration (e.g. public service motivation), and (4) theories on the organization of public administration (e.g. strategic management). Even though a variety of theoretical frames from all across the discipline are used in design research, the epistemological relation between design research and theory—both in terms of utilizing theory for design interventions and of generating new theoretical insights on the basis of design interventions—is not always clearly discussed in publications on design research, which tend to focus more on contextual research.

While works such as Bason (2017) pay limited attention to the use of theory in design, Shangraw and Crow (1998) stress that theory is crucial to design efforts if we are to build upon earlier knowledge, experiences, and lessons. This type of generic knowledge—theoretical knowledge—can also be produced through design science. Hevner et al. (2004, 82–90), who write about the design of information systems, consider generic knowledge—theories and artifacts—as the objective of the design process. Translating specific cases to generic models is seen as a key activity in the design process. Similarly, Barzelay and Thompson (2010, S295–S296) suggests that dealing with this dilemma means that public administration research should grow to be a real applied science. They highlight that this means the ambition of the research should not be to aim at the advice and design but at “contemplative knowledge.” This means that the design should not be founded in comparative and historical arguments as was the common approach in the early days of public administration but in formal modelling and logical consistency: the design should be framed in a more general manner to facilitate its application in other contexts. In that respect, Shangraw and Crow (1998: 1066) stress the importance of strong evaluations of design interventions for developing a theoretical understanding.

In the literature, there is a long-standing discussion about the question whether situational designs can falsify generic models. Waldo (1952) indicates that situational designs can be regarded as falsifications—operational testing—of generic dogmas. He highlights that designs can show that generic dogmas prove to be unsuitable when they are tested in a specific situation. This is also the position that is taken by Stoker and John (2009), who focus on design experiments to test hypotheses. This position, however, is nuanced by Shangraw and Crow (1998: 1070) who stress that there is no testing in the Kuhnean sense since no natural phenomena are studied and only temporary ratified responses to environmental conditions are addressed. Similarly, Snowden and Boone (2007) highlight that situational interventions produce information about the broader, complex system. Probing is actually a way to obtain knowledge about the dynamics in this system by measuring how the system reacts to a situational intervention. Snowden and Boone’s (2007) argument is that simple and even complicated systems can be mapped but for a complex systems a targeted probe is the fitting way of developing knowledge about the system.

Ruijer, Dingelstad, and Meijer (2023) research on open data ecosystems presents an example of design research that built upon and contributed to theoretical knowledge in public administration. The situational problem addressed in the research was to find ways in which data about anti-social behavior could be shared among stakeholders to facilitate collaborative action. On the basis of theories about open data ecosystems (Dawes, Vidiasova, and Parkhimovich 2016; Zuiderwijk, Janssen, and Davis 2014), interventions were designed to strengthen the exchange of information with this open data ecosystem, such as opening up new datasets and organizing joint online work for using and interpreting datasets. These interventions helped to strengthen the data system but also highlighted limitations and key conditions for these interventions to be successful. On the basis of an assessment of the outcomes of the interventions, theoretical insights related the key role of open data champion, and the role of political power were reported back to the academic community in the form of a published research article.

Building Upon and Contributing to the Knowledge Base of Case Studies in Public Administration

The second type of generic knowledge concerns case descriptions of other cases that can provide insights for design research. Knowledge obtained in a specific situation can be communicated through theoretical analyses but also through specific case descriptions that provide an in-depth understanding of mechanisms in context (Flyvbjerg 2001). Flyvbjerg (2001) builds upon Aristotle to argue that these cases should also be considered as academic knowledge but that this type of knowledge is different from scientific theories that aims to identify generalizable patterns. Following Aristotle, he uses the term “phronesis” to refer to the cases and “episteme” as a label for what is generally considered to be scientific knowledge. The knowledge contained in case descriptions plays a key role in case-based learning approaches but also in design science. A second type of knowledge that can be used for design science concern the body of cases considered to be relevant to the specific design situation. These can be examples that are like the situational design context but also quite different cases that provide new insights for a situation. The decline of Kodak, for example, is quite often used to convey knowledge to a host of different organizations about the need to adapt to changing environmental conditions. Schön (1983) refers to these exemplary case as “exemplars” and Blomkvist and Holmlid (2012) stress: “exemplars can be used to inspire, explore, and analyse possible design solutions and are used extensively by design practitioners.”

In the literature on design science, these cases that convey relevant insights and that can inspire design processes in other contexts are often referred to as design exemplar. The special issue of Management Information Science Quarterly edited by March and Storey (2008) presents a nice example of the generic value of design exemplars. Design exemplars fit the idea of building upon situational knowledge but presenting this in such a way that actors in other context can draw lessons from these cases (Raadschelders 2008: 929–931). Pries-Heje and Baskerville (2008: 733) stress that this should be understood as a heuristic application of exemplars rather than a deterministic one: the holistic knowledge presented in the case is not reduced to mechanism but directly translated to other contexts.

Cases can also be the output from design science: a case presents generic knowledge, but not as theory but rather in the form of thick descriptions. In an argument on methods of extending knowledge from single interventions, Bardach’s (2004) article on the extrapolation/adaptation problem argues that extrapolation is not “strict and faithful replication” but borrowing of wisdom and practice of others, leading rather towards adaptation, customizing, and localization, instead of replication (Bardach 2004, 205). Similarly, Argyris and Schön (1989: 613) highlight that the generalization from this type of research do not uncover general laws but rather result in thematic patterns that need to be tested or confirmed in other settings. Van Aken (2004: 232) adds that through β-testing—replication research—researchers can develop further knowledge about how certain design intervention work out in other contexts (see also Denyer, Tranfield, and Van Aken 2008: 396).

Van Geldere, Steenks, and Meijer (2023) research into the use of digital twins for citizen participation provides an example of use of and contribution to the knowledge base of case studies in public administration (although this article also provides theoretical insights). The research aimed to strengthen both the collaborations within local government and between government and citizens as a basis for developing a digital twin that actually had value for social processes. The design intervention built upon insights from other case studies about organizational collaboration and citizen participation. The research resulted in contribution to the situational development of the digital twin but also in a rich description of this situation as a basis for other local government work on the use of digital twins for citizen participation.

Building Upon and Contributing to the Methodological Knowledge on Design Research

The third type of generic knowledge in our model is methodological knowledge about doing design research in public administration research. In public administration research, there is a growing body of methodological knowledge that forms the basis for training and guiding researchers in their work. Methodological knowledge can be distinguished from scientific knowledge since it does not refer to justified understandings of phenomena in the world but to the methods that can be used to produce these justified understandings. Methodology operationalizes an epistemological position into concrete research strategies to produce new evidence. This body of knowledge is crucial for public administration as a sound academic discipline which can defend its knowledge claims. Important work has been done over the previous years on methodological approaches as diverse as experimental research (Bouwman and Grimmelikhuijsen 2016; James, Jilke, and Van Ryzin 2017), living lab research (Dekker, Franco Contreras, and Meijer 2020), and design experiments (Stoker and John 2009). This methodological knowledge is continuously updated through new research and by testing and re-developing proposed methods.

The renewed attention for methods in public administration research has also impacted the field of design research. Methods testing and further development forms a key output from design science. Dorst (2015) highlights how the method he has developed—the focus on re-framing the problem situation as a basis for design interventions—grew from many experiences with design interventions in different contexts. Building upon Krippendorff (2006), Jelinek, Romme, and Boland (2008: 319) refer to this as ‘science for design’ in their introduction to a special issue of Organization Science: through its systematic assessment and analysis, science keeps the design discourse viable and productive. The special issue presents a variety of examples of how science can strengthen design by providing insights in aspects as diverse as the effectuation of design, the design attitude, and the interaction between design and organizational structure.

This sound methodological knowledge is of great importance for generating situational interventions. Askew, John, and Liu (2010) highlight the value of design experiments for producing situational knowledge about policy interventions and stress the interaction between developing useful designs and strengthening our knowledge about design methodologies. As we have already indicated, there is also an established body of knowledge about design research in public administration (Hermus, Buuren, and Bekkers 2020). Generic knowledge about methods for design research forms a key input in the situational process of designing an intervention. The key focus in most of the literature on design is about the methodology and cases are presented to illustrate how the method works rather than to convey information about the content of the case (Bason 2017; Dorst 2015).

Meijer and Ettlinger’s (2023) work on the role of designs in transdisciplinary collaborations is a good example of how research taps into and contributes to methodological knowledge. The problem addressed in this article is how to prepare local and regional governments for the information age. The article discusses a variety of designs that were developed, tested, and implemented in these local and regional governments: an instrument for systematically reflecting on the competences in data teams, a manual for working with dashboards attuned to organization processes, and a web-based tool to support councilors in asking questions about the datafication of local and regional government. This research builds upon insights about transdisciplinary research (Jahn, Bergmann, and Keil 2012; Mobjörk 2010) and also contributes to the academic knowledge on this research methodology by highlighting the need to address the public sector transdisciplinary capacity.

Building Upon and Contributing to Normative Theories in Public Administration

Finally, normative theories are used to reflect on the desirability of certain situational interventions. This type of knowledge can also be distinguished from scientific knowledge since it refers to the world “as it could be” rather than “as it is.” This type of knowledge cannot be tested in empirical situations but challenged on the basis of arguments referring to underlying value choices and consistency of the argumentation. The importance of not using only empirical theories but also normative theories has been identified for a long time, one could even say since Woodrow Wilson ([1956–1924], 2006), in public administration research. Classical normative themes include the responsibility and accountability of civil servants, the role of the law in public administration, and the relationship between administration and politics. Referring to Dwight Waldo’s (1955) classical work, Stivers (2000) argues public administration as an academic discipline builds upon two intellectual dimensions: the scientific search for truths and the philosophical search for the good. Normative theories play a key role in translating outcomes of research—an understanding of the situation—into recommendations for improving the situation.

The orientation on both the “true” and the “good” is clearly present in design research. An analysis of the problem space has a strong focus on presenting a rich and comprehensive understanding of the problem, but the designer has to make choices related to the questions of whose problems and which aspects of the problem are presented. In addition, translating the problem into solutions requires not only creativity but also a normative focus. The attention for being explicit about this normative focus, however, is rather limited in key texts such as Bason (2017) and Dorst (2015). Debates on design focus on issues such as empathy and re-framing rather than highlighting and debating the role of the designer in normative choices.

In the process of building upon normative theory, new insights about these normative theories can also be produced: design science produces knowledge outputs that can contribute to the further development of normative theories. Van Aken (2005) refers to this relationship as field testing of what he refers to as technological rules (see Bunge 1967). A technological rule can be regarded as a building block of a normative theory. The rule is based on a solution concept and indicates how you need to act in a certain situation to realize the desired outcomes. Thorough assessments of design interventions help to strengthen knowledge about these technological rules, and this can help to develop what Gregor (2006: 620) refers to as a theory for design and action. This line of argument has been developed in organizational science but thus far has not been applied much in public administration to make contributions to normative theory.

Meijer and Ruijer’s (2021) work on a code for good governance in an information age in the Netherlands illustrates how design research can contribute to normative theory. In their work, they investigated a variety of normative frameworks proposed for the use of digital technologies, such as algorithms, in the public sector. Their analytic work resulted in a categorization of types of values they grouped under three headings: democracy, rule of law, and government capacity. This framework forms the basis for a method that is now being used by various government organizations in the Netherlands to reflect on public values and identify measures needed to safeguard these values. This illustrates how they build upon various normative perspectives and used design research to integrate and “package” these notions into a method for establishing good governance in an information age.

Design research: building upon and contributing to various knowledge domains

The identification of the four types of generic knowledge can now be used to develop a new understanding of design science as a process of building upon and contributing to various knowledge domains. A key distinction is made between situational knowledge about the contextual situation (about the problem situation, design requirements, user needs, and the like) and generic knowledge (about social and administrative mechanisms, methods for design research, and the like). The situational knowledge flows, mostly based on an investigation of preferences, practices, and perceptions of users, citizens, and professionals, are well presented in the design science literature (e.g., Bason 2010, 2017; Dorst 2015) while the generic knowledge flows are also mentioned in the literature but often implicitly and not in a systematic manner. On the basis of our discussion of the various knowledge types and their relations, we can now present an extended heuristic model with descriptions of the various relations.

The top part of Figure 2 forms a simplified presentation of design science as often presented in the literature (e.g., Bason 2010, 2017; Dorst 2015; Lewis, McGann, and Blomkamp 2020; Van Buuren et al. 2020). The specific situation is analyzed to obtain an in-depth understanding of the problem and then, after a creative process of re-framing and developing solutions for the problem, a design intervention is developed for the situational context. This intervention is further developed through a series of iterations and results in an intervention that has demonstrated its value for tackling the problem in the situational context. This process is well presented in the literature on design science in the public sector (Hermus, Buuren, and Bekkers 2020).

The bottom part of Figure 2 shows how design science in public administration builds upon four different types of generic knowledge: (1) the knowledge base of theories and models in public administration, (2) insights, examples, and design exemplars from other contexts, (3) the knowledge base of methods, instruments, and operationalizations from design science, (4) the knowledge base of normative theories from public administration about “good” interventions. At the same time, rigorous design science can also contribute to these four types of generic knowledge: (1) through design probes, it adds to the theories on public administration, (2) by providing a new design exemplar, it produces knowledge for other contexts, (3) by experimenting with and assessments of the methodology, it adds to the methodological knowledge base, and (4) by highlighting how normative assumption about interventions play out in situational contexts.

Figure 2 highlights thus how design science connects the two challenges of Argyris and Schön’s (1989) “double burden” by emphasizing that the arrows between design science and generic knowledge go both ways: design science requires knowledge about situations to generate interventions in situational contexts but also builds upon a variety of types of generic knowledge. In design research, this generic knowledge is used to design situational interventions, but also new types of generic knowledge are produced. Design research builds upon four reservoirs of generic knowledge—theoretical knowledge, cases, methodological knowledge, and normative knowledge—and also contributes to these four reservoirs. In the next section, we provide methodological guidelines for dealing with this rich process of knowledge utilization and knowledge production.

METHODOLOGICAL GUIDELINES FOR DESIGN SCIENCE IN PUBLIC ADMINISTRATION

On the basis of the discussion of design science as a process of building upon and contributing to generic knowledge, we can now develop guidelines for an academic approach to design science in public administration when the objective of the research is not only to develop useful interventions but also to produce generic knowledge. These guidelines are based on a broad understanding of design science (most importantly: Barzelay 2019; Bason 2017; Dorst 2015; Hermus, Buuren, and Bekkers 2020; Romme and Meijer 2020; Van Aken and Romme 2009) combined with an understanding of public administration research as presented in standard texts in the discipline (McNabb 2017; Van Thiel 2014), specific guidelines for engaged research (Dekker, Franco Contreras, and Meijer 2020), and an analysis of various published examples of design research in public administration already been presented above (Romme et al. 2018; Ruijer, Dingelstad, and Meijer 2023; Van Hout et al. 2024).

The methodological guidelines concern the choices and actions of design researchers regarding the use of generic knowledge for setting up the design intervention and on the production of generic knowledge based on the evaluation of a design intervention. Please note that these guidelines focus specifically on the interface between generic knowledge and situational design interventions (i.e., the bottom part of Figure 2) and not on the exploration of problem and solution space or on the creative and abductive process of developing design interventions (i.e., the top part of Figure 2). The question how to develop situational interventions has been answered—in different ways—in the literature and important methods are reported in Bason (2017), Dorst (2015), and Brinkman et al. (2023). The objective of these guidelines is to create an add-on to those methods to make them suitable for producing generic knowledge for the discipline of public administration. In Table 1, various guidelines are provided for the four types of generic knowledge.

The basic premise of the methodological guidelines is that a rigorous approach to design science requires (a) well-considered and explicit choices about how the design process builds upon the four types of generic knowledge and (b) systematic and comprehensive analysis of the resulting outcomes of the design process to assess how these findings can contribute to generic knowledge. The requirement of explicit choices and systematic assessments applies to the four different types of generic knowledge and thus Table 1 presents a figure with eight cells where each cell provides some concrete guidelines.

The underlying logic of the guidelines for building upon generic knowledge—upper row of Table 1—is that choices in the process of building upon theoretical knowledge, design exemplars, design methods and normative frameworks need to be made in a well-considered manner and they need to be recorded as to make these choices transparent. In Table 1, 1a variety of suggestions are provided for the four different knowledge types such as “Build upon techniques for formal modelling to develop a model for the intervention that can be translated to other contexts” (Theoretical Knowledge), “Explicitly record which design exemplars are used for the intervention and which elements are used and translated to the context” (Design Exemplars), “Indicate what criteria and which context characteristics are considered for choosing a specific design research method” (Methodological Knowledge) and “Indicate how the design intervention is based on an explicit consideration of these normative foundations” (Normative Knowledge).

The underlying logic for contributing to generic knowledge—bottom row of Table 1—is the idea that researchers need to systematically assess the relation between the design interventions and outcomes in context for causal inference. The guidelines for the use of evaluations of design interventions to generate new generic knowledge stress that researchers need to systematically develop interventions and assess what these outcomes mean for the generic knowledge that formed the basis for these interventions. Developing the interventions is crucial to the design research and, if the objective is to also produce generic knowledge, a critical reflection on the assumptions behind the intervention is required. The interventions are based on creative processes, but their knowledge basis needs to be questioned throughout the design process, such as pre-testing the intervention in (simulated) test settings to test the validity and reliability of the intervention. In addition, pre-assessment of the intervention by independent outsiders helps to ensure the validity and reliability of the intervention and to facilitate the subsequent impact analysis.

The literature highlights the need to apply rigorous methods for the assessment of design interventions to disentangle intervention effects, context effects, and, most complicated, the interaction effects between design and context. A key element in the production of robust knowledge is the evaluation of design interventions. In these evaluations, it is crucial to assess the impact of intervention and context and their interactions to produce knowledge about both the design intervention and the complex system. In this respect, Van Aken (2004: 234) indicates that the method of the realistic evaluation (Pawson and Tilley 1997) is a fitting approach since the basic realist formula is mechanism + context = outcome. This approach enables the researcher to reconstruct the various factors at hand and formulate conclusions about intervention and context.

The argument we present here does not imply that researchers should always aim to generate these four types of generic knowledge: one can choose a specific focus on any of the four types of generic knowledge that we identified. Depending on the ambition of the research, the design research approach needs to be adjusted to ensure that it will generate the desired knowledge outputs. A researcher does not have to produce all four kinds of generic knowledge but can choose a focus in their design and then follow the guideline that fits the ambition of the research.

DISCUSSION AND CONCLUSIONS

This article tackles the “double burden” of producing situational interventions and generic knowledge in public administration. We argue design research has the potential of responding to this challenge but requires an appropriate methodology. On the basis of an analysis of the various knowledge types both used for and potentially produced in design science, we formulated methodological guidelines for design science in public administration research that aims to produce both situational interventions and generic knowledge. The message of this article is not that all design science necessarily needs to have the ambition to produce academic knowledge. Many design interventions will be used to develop new organizational models, policies, services, institutions, and the like without having the ambition to publish about these interventions in academic journals. The focus of this article is on design science that has the ambition to contribute to the body of knowledge in our discipline. To produce generic academic knowledge, design science in public administration needs to follow rigorous methodological guidelines.

The article makes three important contributions to the literature: two conceptual and one methodological contribution. The first conceptual contribution concerns an understanding of the different types of generic knowledge used and produced in and through design science. The current literature acknowledges the value of generic knowledge but fails to present a specific understanding of the types of generic knowledge that are used and produced in and through design science. The second conceptual contribution is an understanding of design science as a process of knowledge use and knowledge construction. The various relationships between the knowledge types have been identified and presented in a clear heuristic overview. The methodological contribution concerns a specific list of guidelines that can be used in design research to ensure this research does not only result in useful interventions but also contributes to the valid and reliable production of generic knowledge.

A first point of criticism of the argument is the exclusive focus on the methods of research. Doing research, however, is not only about applying methods but also about having a certain creative and inquisitive attitude. This aspect does not feature much in the literature but is developed thoroughly by Michlewski (2008), who has explored empirically what constitutes a design attitude. The challenge for researchers who want to tackle the “double burden” is to develop a hybrid attitude of both a designer and a researcher. More work into the synergy but also tensions between a design attitude and an empirical research attitude could help to extend our analysis and to strengthen our understanding of the “soft side” of doing academic work.

A second issue is the absence of more critical perspectives. Raadschelders (2008) identifies relativist/critical perspectives as a separate academic tradition in public administration. These perspectives do not aim to develop instrumental solutions or generate generic theories but deconstruct dominant perspectives and ways of seeing and talking about reality. This tradition, however, seems to be notoriously absent in publications on design science. For a full-fledged academic approach to design science in public administration, this perspective seems to be a crucial addition to the instrumental and theorizing approaches.

Finally, we would like to note that this article builds upon the assumption that design science in public administration can generate useful interventions and academic knowledge simultaneously. However, there may be reasons why design interventions cannot be reported or reports of the design interventions are expected to negatively influence relations between key actors. In those case, producing generic knowledge may be more difficult due to restrictions on access to information and permissions to publish findings. One should wonder, however, whether academic researchers should engage in those types of designs or whether they should leave them to consultants.

In sum, this article shows how design science in public administration can have a place in the repertoire of academic researchers who aim to tackle “pure knowledge problems.” The current focus on situational relevance does not form a barrier to more rigorous contributions to generic knowledge in our field but it does require us to rethink the methodologies used for design research in public administration. The guidelines presented in this article can form a starting point for developing more rigorous approaches to ensure that design science can become a valuable addition to our toolkit of instruments for doing public administration research.

References

Footnotes