Towards methodological diversity in sustainability transitions research? Comparing recent developments (2016­2019) with the past (before 2016)

The sustainability transitions research field continues to broaden rapidly, engaging scholars from a wide variety of disciplinary backgrounds. As highlighted in the STRN research agenda, transitions research relies on a broad range of methodological approaches; yet the methodological diversity of transitions studies has remained somewhat limited. In this article, we investigate if this verdict still applies to transition studies published since 2016. While our findings suggest that the recent literature became methodologically more diverse especially with regard to theoretical frameworks and research questions, other methodological dimensions continue to suggest a lack of diversity. Based on our literature review we propose three particularly promising areas for further diversifying future research on socio-technical transitions towards sustainability: theo- retical bridging, novel methods, and multi-scalarity across multiple sectors.


Introduction
Research on socio-technical transitions towards sustainability has experienced an enormous increase in publication output during the last two decades (see Fig. 1). Looking at just the past four years (2016)(2017)(2018)(2019), the number of Scopus listed journal articles on sustainability transitions has doubled since 2016, reaching an annual output of almost 450 articles in 2019. This growth of the transitions literature is accompanied by an increasing engagement of scholars from different research disciplines (Köhler et al., 2019). However, using the 'transitions research onion' framework to delineate different methodological dimensions, Zolfagharian et al. (2019) found that, for the core of the transitions literature published before 2016, the methodological underpinnings of the field were rather underdeveloped and narrow.
Given the recent dynamics and disciplinary expansion of the field we investigate if research on sustainability transitions is also diversifying methodologically, as Köhler et al. (2019) postulated in the current transitions research agenda. For this assessment we compare the state of the art of methodologies applied in transition studies between 2016 and 2019 with the assessment of the methodological diversity of the field before 2016 done by Zolfagharian et al. (2019). We close with reflections on the analytical scope of the 'transition research onion', apparent methodological shortcomings in the field and thus promising areas for future research.

Recent developments in methodological diversity
By adopting the search string of Zolfagharian et al. (2019), which draws on the review article by Markard et al. (2012) (see Appendix A), we identified 260 articles in Scopus published on sustainability transitions between 2016-2019 in the five main journals in the field (Environmental Innovation and Societal Transitions, Research Policy, Technological Forecasting and Social Change, Energy Policy and Technology Analysis & Strategic Management). Of these, we excluded 43 publications (17%) from the methodological analysis, as they are conceptual in nature. Compared with the period until 2015, the share of empirical studies increased by 20 percent, indicating that the research field is steadily maturing (Köhler et al., 2019). Accordingly, the subset of 217 empirical papers was examined using key methodological dimensions of Zolfagharian et al.'s 'transition research onion' framework.
We find that while the transitions literature before 2016 predominately focused on the explanation of (past) transitions (above 50%), current research questions have given greater attention to the role of actors (up from about 20% to almost 48%) as well as transition policies and pathways (up from over 30% to nearly 50%, see Fig. 2). With actors and their agency being key for socio-  technical systems change (Kern and Rogge, 2016;Wittmayer et al., 2017) and policies playing a key role in redirecting and speeding up socio-technical change (Kanger et al., 2020;Kern et al., 2019), a stronger emphasis on these topics is welcomed.
When it comes to transition theories our analysis finds that the four transition-focused theoretical frameworks (MLP, TIS, SNM, TM) together cover an unchanged share of approximately 30 percent of empirical studies (see Fig. 3). Unpacking this a bit further confirms the dominance of the Multi-Level Perspective (MLP) which remains unchallenged and was even applied more widely (up from roughly 15%-20% of studies). At the same time, papers draw less frequently than before on the Technological Innovation System (TIS), Strategic Niche Management (SNM) and Transition Management (TM) frameworks, which are also considered "major conceptual developments in the field" (Markard et al., 2012: 957). The percentage of transition studies using at least one of these three frameworks decreased from just under 15 to nearly 9 percent.
Nowadays, almost half of the published work includes borrowed/not transition-focused frameworks originally developed in other disciplines (e.g. innovation ecosystem approach) compared to a quarter of studies before 2016. This trend underlines the increasing interdisciplinary character of transitions research. Without neglecting the crucial role of the main theoretical frameworks, the use of alternative concepts to study transition processesi.e. theoretical pluralisationmight enrich the field substantially (Hopkins et al., 2020).
Indeed, our analysis shows that approximately one fourth of the studies develop new frameworks by combining concepts and theories across disciplines (e.g. integration of MLP and discourse theory); however, the development of such new frameworks decreased by 12 percentage points, implying that less emphasis is being placed on the further refinement of existing frameworks by integrating concepts from other fields. While these developments could be interpreted as maturation of the field, they could also indicate a limited willingness of scholars newly entering the field from other disciplines to engage with and build upon established socio-technical transition frameworks. Consequently, the transitions community is becoming more diverse theoretically, but we argue that harnessing such increased diversity would require increased efforts in bridging and integrating old and new theoretical frameworks. Despite the identified shifts concerning research questions and diversification of theoretical frameworks, neither research methods nor the data sources used to empirically study transition problems have changed significantly since 2016 (see Fig. 4). With a share of more than 80%, qualitative research designs continue to dominate the field over quantitative and mixed-method approaches. Simultaneously, the usage of documents (70%) and interviews (approx. 54%) has remained at almost the same high level, followed by workshops, observations and surveys with a share of about 10 percent each. Given this background, it is not surprising that qualitative data analysis techniques, especially grounded theory and qualitative content analysis (e.g. Becker et al., 2018;Weisenfeld and Hauerwaas, 2018), still prevail. This may indicate that the transitions research field continues to struggle to connect with disciplines leaning more on quantitative methods (Alkemade, 2019) and to generate generic insights beyond single cases (Köhler et al., 2019). However, quantitative and mixed-methods research, although less commonly used, offer promising prospects for sustainability transition studies as they already cover a variety of analytical approaches ranging from social network analysis, foresight and scenario modelling to Q methodology and (statistical) analysis (e.g. Cantner et al., 2016;Falcone et al., 2018;Gruszka, 2017;Papachristos and Adamides, 2016).
Finally, it should be noted that between 2016 and 2019 the application of cross-sectional research designs increased (from representing just 12%-30%), corresponding with a drop of longitudinal studies to 70 percent. This also corresponds with our earlier observations of a decrease of research questions which are concerned with (whole or part of) past transitions. Beyond the time framing, methodological diversity is likely to be influenced by spatial and sectoral dimensions; however, these are not yet considered in the 'transition research onion'. Attention to spatial and sectoral coverages reveal that between 2016 and 2019 more than three quarters of all empirical publications dealt with energy transitions and the vast majority (80%) have a national focus. This is in line with the results of the review by Markard et al. (2012), indicating a certain continuity over the years.

Outlook
Based on this update of methodological patterns in socio-technical transition studies we offer some reflections on promising areas for future research for the STRN community.
First, future transition research should utilise a diversity of frameworks to analyse sustainability transitions (Sovacool and Hess, 2017). This, however, will require a critical engagement with established, new and borrowed frameworks to study socio-technical transition processes, and dedicated efforts in bridging diverse theories, concepts and approaches. A case in point are attempts to improve transition theorizing by harnessing organization theories (van Mossel et al., 2018) or policy process theories (Kern and Rogge, 2018).
Second, data collection and analytical methods ought to be diversified to better reflect the complexity of sustainability transitions and the inter-and transdisciplinary nature of the field. Although Zolfagharian et al. (2019) highlighted action research as a fruitful method to be applied in transitions research, none of the reviewed articles published between 2016-2019 explicitly use this approach. Therefore, we urge transitions scholars to reflect on their role in research processes, and to consider the stronger use of action research methods. At the same time, however, some studies are already applying promising alternative methods, such as modelling (Li and Fig. 4. Data collection methods and sources then (before 2016) and now .
Third, we argue for a greater awareness of both the sectoral and spatial coverage of empirical studies, which so far have been neglected in the 'transition research onion'. Yet, different sectoral structures with their diverse and increasingly interconnecting sociotechnical systems may necessitate adapted research designs (Andersen et al., 2020;Rosenbloom, 2020). Similarly, with global industry and transition dynamics being locally and regionally embedded, the multi-scalarity of sustainability transitions deserves more attention (Binz et al., 2020;Coenen et al., 2012;Truffer et al., 2015). Therefore, we agree with others that more research is needed that goes beyond the national level and embraces the multi-scalarity of transition dynamics, which could build, for example, on evolving theorizing about global regimes (Fuenfschilling and Binz, 2018), regional lead markets for environmental innovation (Losacker and Liefner, 2020) or multi-layered innovation systems (Binz and Truffer, 2017).
In conclusion, we argue that the methodological diversity of transition studies has increased significantly over the period 2016-2019. Arguably, as the field advances, further diversification, but also consolidation, should be guided by pressing research questions of our time on how to accelerate transitions towards more sustainable systems of production and consumption.

Declaration of Competing Interest
The authors report no declarations of interest.

Acknowledgements
The research underpinning this commentary was funded by the in-house SYSTRA project at Fraunhofer ISI. Its key findings were presented at the virtual IST conference in Vienna during Karoline Rogge's keynote on August 21, 2020.

Appendix A
Based on the review articles by Zolfagharian et al. (2019) and Markard et al. (2012), we identified the relevant literature by using the following search string in Elsevier's Scopus database: