Abstract
Many research-to-practice approaches exist, yet fundamental differences in interpretation and uneven application of research translation approaches create a real challenge in generating research-driven solutions for development challenges. In this review, we aim to clarify the plethora of approaches to increase research uptake into international development practice and policy. Through a qualitative analysis of 93 articles, we identified four interrelated factors that differentiate research translation approaches. Building off the variation expressed across these factors, we propose a conceptual framework called the Research Translation Continuum. This framework illustrates the range of ways to produce useful evidence and the range of ways that academics and practitioners can collaborate. Our study makes a unique contribution by providing a conceptual framework for guiding research translation efforts in international development. We conclude that critical reflexive engagement is needed to situate, use, and produce evidence for development policy and practice.
Keywords
I. Introduction
Scholars and donors in international development call for research to have an impact beyond academia and to connect more directly with policy and practice (Fowle et al., 2021; Fransman et al., 2021; Gunn and Mintrom, 2021; Toukan, 2021). This trend is paired with increasing funding opportunities for academic research to improve development outcomes (Newman et al., 2012; Mach et al., 2020; Toukan, 2021) in many parts of the world. For example, in the United Kingdom, the Global Challenges Research fund allocated £1.5 billion to development research between 2016 and 2021 to support research that addressed the challenges faced by developing countries and identify ‘pathways to impact’ to ensure the developing countries benefit from the research (McKay, 2021). In the United States, the Higher Education Solutions Network (HESN 1.0) program allocated $140.5 million to leverage science, technology, and innovation to solve development challenges from 2012 to 2017 (Lebrón et al., 2014; Lebrón et al., 2014), recognizing that ‘higher education institutions play a critical role by spurring innovations to solve global development problems’ (Moore et al., 2021). Furthermore, there is also a growing instrumentalization of research for development, including funders being more involved in defining research questions and aims, as well as an increased focus on applied research. Despite this momentum in many areas of the world, there is uncertainty on how to increase the uptake of research in international development.
Many research-to-practice approaches exist, including knowledge exchange, implementation science, knowledge mobilization, research-to-action, research uptake, and transdisciplinary research (Graham et al., 2006; Islam et al., 2020; Nguyen et al., 2020; Hakkarainen et al., 2021). Yet, these approaches come from different disciplines, ontologies, and epistemologies. There are fundamental differences in interpretation and uneven application of research translation approaches which can lead to implementation failures (Oliver et al., 2019; Hakkarainen et al., 2021). For example, implementation science is the ‘scientific study of methods to promote the systematic uptake of research findings and other evidence-based practices into routine practice, and, hence, to improve the quality and effectiveness of health services’ (Eccles and Mittman, 2006: 1). There is a tendency in the field of implementation science to assume that strategies to overcome barriers to research translation are like tailoring clinical treatments in health problems (Nilsen et al., 2013: 8). Therefore, implementation science has a more positivist orientation as academics aim for generalizable predictions related to the strong influence from the medical sciences (Boulton et al., 2020). In contrast, development studies literature focuses on the interplay between social, economic, ecological, cultural, and gendered aspects of social change at all scales, from the local to the global (Baud et al., 2019). There is a plethora of debates within development studies, including whether the discipline is primarily focused on academic research or centred on policy or practice (Kothari, 2019). Moreover, critical development theorists, like Escobar (1995), and participatory scholars, like Chambers (1997) have challenged how academics produce knowledge that reinforces existing power hierarchies and excludes disadvantaged groups (Mitlin et al., 2020: 549). Therefore, development studies has a more interpretivist or constructivist paradigm that emphasizes the subjective nature of knowledge, reflexivity, and iterative theorizing (Steketee et al., 2020: 2). These epistemological and ontological positions shape how research translation is conducted and, together with the profusion of these approaches, create a real challenge for generating research driven solutions for development challenges.
The aim of this study is to clarify the plethora of approaches to increase research uptake into international development practice and policy. Our review makes a unique and novel contribution to analyse the intersection of multiple disciplines and approaches relevant to international development and offers a conceptual framework for navigating the various avenues for evidence production and use. Given the complexity in addressing global challenges, including cross-disciplinarity, the necessity of a diverse range of stakeholders, and an overwhelming range of translation approaches, there is a great need for a conceptual framework to guide the application of evidence into development practice and policy.
II. Methods
Our literature review used a phased methodological approach to explore the concept of research translation through a scoping review, a rapid review, and two rounds of qualitative analysis. A transdisciplinary research team consisting of two technical advisors at Catholic Relief Services (LR and AT) and one post-doctoral researcher at Purdue University (PB) conducted the study. Each author has a doctorate in a different academic discipline: human geography (LR), ethnobotany (AT), and business (PB). Although a single term is unlikely to adequately encompass all approaches to research uptake and knowledge exchange (Fazey et al., 2013: 20); we choose the term ‘research translation’ to generally refer to any approach to producing academic knowledge for uptake into policy or practice. ‘Research Translation’ is a term used in the health care disciplines to refer to ‘the process of applying laboratory research to human studies and enhancing the adoption of evidence-based practices in real-world settings to reach broad populations’ (Zoellner et al., 2015: 1057). We recognize that the term can be problematic (Estabrooks et al., 2006); however, since research translation does not have a singular theory (Krieger et al., 2021), we chose to use it as an umbrella term for our review but did not include it as a search term.
In conducting the review, we expected to find some translation approaches would be more linear, focusing on evidence use after the completion of research (Newman et al., 2012; Banks et al., 2015; Waddington et al., 2018), while other approaches would be more iterative, emphasizing the co‑production of evidence and knowledge (Lemos et al., 2018; Mach et al., 2020: 30; Norström et al., 2020). Although the terms ‘evidence’ and ‘knowledge’ are debated, contested, and utilized differently (Jennings Jr. and Hall, 2012; Golden et al., 2015), for the purpose of this study, we use the term ‘evidence’ to refer to knowledge produced through any scientific method.
First, we conducted a scoping review (Pham et al., 2014; Waddington et al., 2012) to examine the depth and breadth of literature published on the topic of research translation in multiple sectors relevant to international development. Drawing on 119 documents collected in a snowball approach, we identified the distribution of keywords and phrases used in the document. Through a selective sample method (Xiao and Watson, 2017), we examined the title, abstract, keywords, and discipline of the 119 articles. Due to the variety of terms found, a ‘pearl harvesting’ method (Sandieson, 2006; Waddington et al., 2012) was used to finalize the search terms and disciplines for the rapid literature review. We selected the four most common disciplinary bodies of literature that appeared in our scoping review: agriculture, political science, development studies, and health care science and services. We then selected the primary research translation approaches that corresponded with each of the disciplines as our search terms: evidence-based policymaking (EBPM), technology transfer (TT), participatory action research (PAR), knowledge translation (KT), and integrated knowledge translation (IKT). Because of the over-representation of translation approaches in the health sciences, we chose one approach that we perceived to be more linear and one that was more iterative.
Second, we conducted a rapid review (Khangura et al., 2012; Tricco et al., 2015; Tricco et al., 2017) using the Web of Science database (see Figure 1). The rapid review is a ‘type of knowledge synthesis in which systematic review processes are accelerated and methods are streamlined to complete the review more quickly than is the case for typical systematic reviews’ (Tricco et al., 2017: 3). We adopted the rapid literature review method of Khangura et al. (2012) because it follows the same principles as a full systematic review but is less time and resource‑intensive, which aligned with the context and scope of carrying out this study in a limited term. The translation approaches included were purposefully selective to keep the scope of the study manageable within our time constraints. Despite these limitations, we are confident that our analysis of the literature is accurate, and that our conclusions are relevant for the use of evidence in the international development community.
PRISMA Diagram of Process of Identification, Screening, Eligibility and Inclusion of Articles for Rapid Literature Review.
We conducted a Boolean search in the Web of Science by inserting the research translation approaches (e.g., technology transfer) into the ‘topic’ field, and the selected disciplines (e.g., agriculture) into the ‘research area’. We did this process separately for each approach and discipline pair and following our rapid review methodology, limited the selection of only 25 articles for each pair based on highest ‘usage count’ 1 . Thus, we identified a total of 125 articles relying on the largest ‘usage count’ from the five terms as shown in Table 1. Usage count is a measure of the level of interest in a specific item on the Web of Science platform. No grey literature or book chapters were included in this review; only articles found in peer‑reviewed journals published between 2000 and 2020. Our inclusion criteria included: (a) material self-identified as research translation or knowledge co-production or describing a process similar to it in the title or abstract, (b) material written in English, (c) peer reviewed journal articles with full text available, (d) any type of article (case studies, empirical papers, theoretical articles, literature reviews, or other types), (e) material that would contribute evidence or tools to answer the research. Our exclusion criteria included: (a) material for which titles or abstracts did not explain a linear or iterative knowledge translation process, (b) duplicates, (c) items with no full text available, (d) chapters from books, and (e) grey literature. We did not restrict based on geographic areas.
Next, we screened 125 titles and abstracts from the search and excluded 31 articles that did not pertain to our research (Table 1). The screening criteria for the articles included if the abstract mentioned a type of translation, the inclusion of practitioners and researchers, or research utilization. The final breakdown of type of articles from each discipline varied. For example, the articles for IKT and PAR were overwhelmingly empirical, while those for KT were overwhelmingly theoretical or were reviews. The geographic focus of the empirical articles also varied, for example, in IKT was exclusively in Canada, whereas PAR was predominantly in the Global South.
Results of Literature Search and Screening Processes.
Our analysis of 93 peer reviewed articles was a three-step process: first, we used inductive data analysis to produce the narrative summaries of each research translation approach (Schutt and Chambliss, 2013); second, we used comparative analysis to compare the approaches (Patton, 2002); third, we used an interpretive analysis to create typologies of research translation approaches (Patton, 2002: 459–460).
III. Results
Our results elucidate the similarities and differences between each of the five research translation approaches to discover concepts not accounted for in our original formulations (Patton 2002: 493–494). Our analysis and comparison of each of the five approaches found that research translation is highly contextual and can be conducted in a variety of ways. Then, we found four intertwined factors that differentiated the research translation approaches. Lastly, we propose that research translation incorporates a continuum of approaches between proactive to post-facto translation and that evidence use and partner engagement are essential dimensions within the continuum. Each of these findings are further detailed below.
Comparative Analysis of Approaches
We focused on nine areas in our comparative analysis: (a) history, (b) purpose, (c) core characteristics, (d) initiating actors, (e) level of collaboration, (f) power distribution, (g) timing of translation and (h) scale of translation, and (i) common challenges (Table 2).
Summary of Comparative Analysis Across Five Research Translation Approaches.
We began our analysis with comparing the histories of each of the approaches. The origin of each approach helps to identify the depth of the knowledge base and its intended use. The five approaches varied in their origin, starting with TT dating back to the early 20th century to more recently, IKT starting in the early 2000s. Next, we sought to compare the aims and purpose of each of the approaches. TT aims to improve agricultural output and food security, and thus livelihoods. PAR aims to transform power inequalities in research through promoting the co-design of research and the co-production of knowledge with communities, practitioners, and other relevant stakeholders for social change. While the purpose of EBPM is rational decision and policymaking, it has historically been geared toward improving society through measurable outcomes and providing rationality for cost effectiveness. The motivation for KT and IKT is to improve health outcomes and health care practices, address evidence gaps, and implement evidence.
We then compared the core characteristics of each approach, which helped to differentiate how each approach is implemented. Innovation is a core characteristic of TT. PAR is characterized by inclusion, action, and reflection to collectively improve the circumstances of its participants and their communities. EBPM focuses on evidence use to produce better public program outcomes, yet it also has been characterized as an opaque process involving various sources of evidence. KT can be implemented in a variety of ways and cultivates relationships in multidisciplinary teams. IKT aims to address the limitations of KT and it emphasizes the partnership between researchers and knowledge users throughout the research process.
Next, we compared the initiating actors in each approach; the literature refers to the question of research initiation as ‘push’ or ‘pull’ efforts (Reardon et al., 2006; Tetroe et al., 2008). If researchers initiate the research and seek recipients for the research findings, it is considered a push effort. If policymakers or practitioners seek out research, it is a pull effort since they have identified a need and seek evidence to make decisions (Reardon et al., 2006). For TT, most often there are two approaches; either (a) a government agency asks for research or (b) researchers ‘push’ innovations to agricultural extension organizations, who then ‘push’ these to farmer communities. Similarly, in KT, either researchers or practitioners could initiate research depending on the context. In PAR there are relational linkages between researchers and practitioners; we found that the request for research could originate from community members, non-profit organizations, government agencies, or researchers. EBPM is primarily a ‘pull’ approach in which policymakers use existing published research evidence in their policymaking. Lastly, IKT originated as a donor-driven approach, in which the funding agencies ‘push’ the approach onto researchers to develop partnerships with practitioners.
We then explored the differences in level of collaboration between researchers, practitioners, and other stakeholders. The level of involvement of practitioners varied across the five approaches. We found that PAR and IKT are the most collaborative. In each, practitioners are included in the research process from the start. Some knowledge translation models were described as more collaborative than others, but not all models described in the reviewed literature assumed collaboration between researchers and practitioners. TT typically was less collaborative but has become more cooperative and participatory in recent years, where stakeholders such as farmers are being included as research participants. The EBPM approach does not tend to include any collaboration between researchers and policymakers.
Next, we compared power distribution between actors in the five approaches. PAR projects often result in action taken by participants, thus empowering populations typically left out of the research and decision-making process. Therefore, the PAR methodology can shift power relations so that research participants are able to challenge social inequalities. IKT also has shared power distribution as research products are developed with both researcher and practitioner knowledge, thus democratizing the knowledge-creation process. In contrast, power distribution in TT, KT, and EBPM depends on the context. In some cases, researchers have no contact with policymakers or end users, and this provides power to each to make decisions independently. In other cases, in EBPM, policymakers are seen as having more power as they collect and produce evidence through negotiation over relevance and appropriateness (Mele et al., 2013: 6).
We continued our analysis by comparing the timing and scale of translation across the approaches. More traditional TT and KT approaches tend to initiate translation at the end of the research lifecycle, while more participatory models include practitioners in the data collection phase, or sometimes as early as problem identification. Translation in an EBPM approach could begin after research is completed or if there is an evidence gap in published literature. In PAR and IKT, research translation begins in the research planning process, as early as problem identification. Furthermore, the articles we reviewed discussed research translation occurring at all scales. The articles generally did not address outcomes or impact, thus the ‘scale of translation’ refers to the level at which the research was intended to be used. The TT articles discussed agricultural innovation at all levels- the national policy level, agricultural extension level, community and individual farmer level. KT and IKT articles also described translation at the macrolevel (health care policy), mesolevel (hospital or health centre, and community) and microlevel (individual health care provider and patient). The EBPM articles discussed translation at the national, state, municipal, and classroom level. The PAR articles discussed translation more at the microlevel, for specific organizations and communities.
Lastly, we analysed the most common challenges across the approaches. We found that of the constraints cited repeatedly, the importance of understanding context and the time and resource-intensive nature of translation were the most salient. We found multiple aspects of context to take into account including (a) organizational structures (May et al., 2016), (b) stakeholder engagement (Ortiz et al., 2013), (c) geography, history, and colonial histories (Fowler and Rockstrom, 2001), (d) culture (Jull et al., 2019), and (e) environment (Lybbert and Sumner, 2012; Friedlander et al., 2013). We also found that the success of an intervention in one context did not guarantee adoption or impact in another context (Fowler and Rockstrom, 2001; Ward et al., 2009; Preyde et al., 2013). Furthermore, we found that academia frequently underestimates the amount of labour and time involved in maintaining partnerships and undervalues the outcomes of more participatory approaches. Moreover, researchers, their partners, and academic institutions have different expectations of the timeliness of outputs (Lyons et al., 2014). While all approaches generally highlighted the need for an investment of time and resources that was beyond the scope of what was available to them, the more collaborative approaches such as PAR and IKT, required a greater expenditure of time, money, and labour.
Four Interrelated Factors Shaping Research Translation
After our in-depth analysis across these nine components, we identified four interrelated factors that differentiated each of the approaches to research translation (Figure 2). The first factor, intention of evidence use, emerged from common themes observed across the components of history, purpose, characteristics, timing, and scale. The second factor, commitment to partnership, emerged from the themes of initiating actors, level of collaboration, and power distribution. From our analysis of barriers we identified the two additional factors- understanding of context and investment of time and resources. Together, these four factors exert influence on the approaches; the configuration of the factors shaped the research translation approaches’ design, implementation, and impact. Each factor is further detailed below.
Four Intertwined Factors that Shape Research Translation Strategy.
Intention of Evidence Use
‘Intention of evidence use’ refers to the degree of forethought and planning for research utilization. It includes the motivation behind the evidence collection and the plan for evidence use. We found that research projects that invested more forethought into research utilization employed a vastly different translation approach than those projects which invested less planning. This forethought can include discussions of the goal, methodology, audience, or scale for evidence use. Four out of five of the literatures reviewed—TT, PAR, KT, and IKT- approached planning for evidence uptake at the outset of the research design, but in varying degrees. In some PAR and IKT approaches, the purpose of research and planning for research use occurred simultaneously when academics and non-academics jointly defined the problem, established research questions, designed problem focused research, and decided how to apply research (Beh et al., 2013; Golden et al., 2015; Gutberlet, 2015; Graham et al., 2018; El-Masri et al., 2019). These intensive collaboration efforts led to evidence production that was intricately bound with a specific context, audience, scale, and use. In contrast in EBPM, researchers could produce evidence without the intention for it to be used in a specific context or at a specific scale (Jennings Jr. and Hall, 2012). Or research could be conducted for very specific contexts and scales like improving the deliberative governance of parent councils in Swiss schools (Gundelach et al., 2017) or international human trafficking policymaking (Weitzer, 2014). Furthermore, some scholars intended to collect scientific evidence, indigenous knowledge or experiential knowledge from inception because they could complement each other and produce a better understanding of on-the ground realities (Graham et al., 2006; Straus et al., 2011; Mele et al., 2013). The forethought to include knowledge bases typically left out of research show a deeper level of investment and planning for evidence utilization.
Commitment to Partnership
‘Commitment to partnership’ refers to the level of investment between academics and practitioners in engaging as partners in a research translation project. In our review, researchers were typically from academic institutions, while practitioners were those that used the evidence produced including hospital administration, clinicians, policymakers, non-profit organizations, government agencies, or civil society organizations. We found that scholars in PAR and IKT were highly committed to a collaborative partnership (Fabricius and Pereira, 2015; Graham et al., 2018) as their approaches were characterized by concepts such as inclusion, joint action, and collaboration, while scholars in TT and KT engaged with practitioners based on the objectives of each project (Zwarenstein and Reeves, 2006; Sewell et al., 2014). In EBPM, generally there was no collaboration between policymakers and academics, which was considered valuable to politically sensitive research (Jennings Jr. and Hall, 2012; Mele et al., 2013). Thus, PAR and IKT were the most collaborative in nature, followed by TT and KT, which partnered in different ways depending on the context, and lastly, EBPM which involved no direct engagement between academics and policymakers.
Understanding of Context
‘Understanding of context’ refers to the consideration of environmental, social, economic, and political circumstances that shape the capacity for translation and potential for adoption. The PAR and IKT approaches advocated for a high level of understanding of local context, in tandem with their emphasis on collaboration and inclusion. In PAR, for example, the academics and participants collaboratively chose the research methods and determined the intended action for their context (Gutberlet, 2015; Smith and Jenkins, 2015). Studies from the TT, KT, and EBPM approaches were mixed in their expression of this factor. One TT study indicated that ‘upstream research’ was applicable and fairly easily transferred across geographic regions, unlike ‘downstream research’, which required addressing local conditions (Lybbert and Sumner, 2012). Since policymakers translate research in EBPM, evidence is subject to multiple interpretations depending on the subjective experiences and perceptions of the participants engaging with the research to form policy (Jennings Jr. and Hall, 2012). This factor can be understood as a limiting factor as unexpected external factors may influence research impact.
Investment of Time and Resources
‘Investment of time and resources’ refers to the financial and human resource support needed for research translation to have an impact. Across all five approaches, time and resources were consistently highlighted as necessary investments. The more collaborative approaches such as IKT and PAR generally required more time and resources given that they regularly interacted with a broader group of stakeholders. We found multiple examples illustrating the time-intensive nature of research translation: time to form and maintain collaborations (Grandin, 2003; Moodie et al., 2011), tight research timelines (Corson, 2012), and academic partners’ teaching duties (Preyde et al., 2013). The collaborative approaches to research translation, like TT, IKT, and PAR, were considered ‘slow science’ and required time to build relationships which extended the duration of the research and required extra financial resources for labour and materials (Sewell et al., 2014: 72). For example, some of the PAR projects reviewed took 7 to 15 years of work to document the full extent of their research (Humphries et al., 2012; Gutberlet, 2015). Furthermore, we found that institutions of higher education typically undervalued research translation in tenure processes (Lyons et al., 2014; Wathen and MacMillan, 2018) and practitioner organizations frequently lacked resources for research (Preyde et al., 2013). In contrast, EBPM required policymakers to integrate information that changed over time and to sometimes make rapid decisions (Jennings Jr. and Hall, 2012: 265), therefore there was often a lack of time to find the evidentiary data to make quick decisions (Mele et al., 2013).
Research Translation as a Continuum
To capture the complexity and variation of research translation expressed across the four inter-related factors, we propose a conceptual framework called the Research Translation Continuum. This framework (Figure 3) illustrates the range of ways to produce useful evidence, and the range of ways that academics and practitioners engage in research translation. On one end of the continuum is proactive research translation, or intentionally planning for evidence use before research begins and ensuring continuous collaboration between academics and practitioners until evidence is utilized to change policy or practice. On the other end of the continuum is post-facto translation or applying academic research to policy or practice after the research has been conducted and published, with minimal collaboration between academics and practitioners. Between proactive and post-facto research translation, there are numerous variations of research translation approaches.
Two internally influential factors, intention of evidence use and commitment to partnership, can be further explored as theoretical reference points on a spectrum within the Research Translation Continuum (Figure 3). We pulled out the evidence utilization and partner engagement spectrums of the continuum as foundational factors influencing the variability of research translation. Like Fransman and Newman (2019: 525 italics in original) who concluded that ‘understanding the relationship between evidence and participation is paramount for understanding and improving partnerships’, we focused on these two realms within the Continuum since these are the factors that researchers and practitioners have the most control over in terms of situating and recognizing their epistemological, and ontological positions and the subsequent impact of those decisions. However, as the four interrelated factors demonstrate, the additional factors of context and time/resources influence which approach is feasible for any given setting.
The Research Translation Continuum with Evidence-utilization Spectrum and Partner Engagement Spectrum.
The evidence-utilization spectrum illustrates the incremental change in forethought for evidence utilization along the Continuum 2 . We represent the gradual changes in evidence use through theoretical reference points along the spectrum: evidence impact evaluation, evidence uptake, evidence application, evidence dissemination, and evidence diffusion. Reference points on the left indicate substantial forethought to evidence utilization and substantial intention for research evidence to be readily accessible to practitioners. In contrast, reference points on the right indicate minimal forethought to create evidence for use and minimal efforts to make evidence readily accessible to practitioners. We termed the most proactive approach to evidence utilization as evidence impact evaluation, as it goes beyond planning for and applying evidence to also incorporate an evaluation of evidence uptake.
For the partner engagement spectrum, we designate five reference points to represent the incremental variation in engagement among research translation approaches: co-productive, participatory, consultative, transactional, and minimal. The reference points on the left indicate a deep and ongoing engagement between academics and practitioners, while the reference points on the right indicate the least amount of engagement and collaboration between academics and practitioners.
Our final analysis entailed attempting to map the five approaches we reviewed onto the Research Translation Continuum. Even among these approaches, however, there can be wide variations in application based on the specific context so we found that it was not possible to locate each approach directly onto the Continuum. Furthermore, not all research translation approaches are situated in parallel on the two spectrums. For example, in conventional TT there is high intention for evidence utilization, so TT falls on the left side of the evidence-utilization spectrum. At the same time, conventional TT approaches have minimal to transactional partner engagement and therefore, TT is situated on the right side of the continuum. However, in general, IKT and PAR are closest to the proactive end with EBPM closest to the post-facto end. KT and TT vary considerably and can be mapped anywhere on the continuum depending on the context.
IV. Discussion
While we hypothesized that some research translation approaches would be linear in nature and focus on using evidence, and others would be more iterative and focus on the co‑production of knowledge, we found that each of the five approaches defines and uses evidence differently and advocates differently for researcher–practitioner partnerships. There were also variations in both the evidence use and partnership aspects within the research translation approaches we reviewed. These findings led to the understanding that research translation is not a monolithic process; it is a much more complex and variable concept than one or two models can convey.
The Research Translation Continuum showcases the incremental variations in evidence utilization and engagement that we observed across the five approaches, but it does not claim any one approach as the most effective or impactful. For most scenarios, increased, meaningful participation with partners will improve the potential to produce relevant research and adoption of the outcomes (Beh et al., 2013; Kothari and Wathen, 2013; Golden et al., 2015; Gutberlet, 2015; Gagliardi and Dobrow, 2016; Gundelach et al., 2017; Graham et al., 2018; Kreindler, 2018; Walton et al., 2018; Roberge-Dao et al., 2019). However, some situations call for the post-facto approach to translation. Characterized by evidence diffusion and minimal partner engagement, this approach can be appropriate in circumstances where policy and science are more impactful when separated. For example, post-facto approaches may be more valid in situations when academics are pulled to legitimize political positions (Oliver et al., 2019: 5), or when Indigenous knowledge must be centred instead of integrated with western knowledge to support Indigenous sovereignty and leadership (Latulippe and Klenk, 2020).
While our results resonated with sustainability-oriented research that identified distinct ‘research modes’ that fall on a ‘spectrum of transdisciplinary research modes’ (Hakkarainen et al., 2021; Jahn et al., 2021), what is unique about our findings is that we examine approaches that are relevant to international development. We offer the Research Translation Continuum as a conceptual framework to navigate the complexity of the concept of research translation in international development and create a common understanding and terminology for practitioners and researchers from across different disciplines.
Regardless whether the translation approach fell on the post-facto or proactive end of the Research Translation Continuum, we found that the motivation of the actor who initiated the research matters. There are power differences based on if research translation is requested by end users (Lavis, 2006; Reardon et al., 2006; Jennings Jr. and Hall, 2012; Mele et al., 2013; Lyons et al., 2014; Sewell et al., 2014; El-Masri et al., 2019). For TT, at least for the agricultural extension model, it is the government agencies that request for innovations from researchers. PAR has relational linkages between researchers and participants, and we found that the request for research could originate from either stakeholder, depending on the context. In contrast, with the EBPM approach, policymakers use existing published research evidence in their policymaking. Lastly, IKT originated as a donor‑driven approach, in which the funding agency requested academics and practitioners to form research partnerships. Furthermore, these different ways to initiate projects are exacerbated by external factors like tenure review, grant funding timelines, or the political environment (Jennings Jr. and Hall, 2012; Lybbert and Sumner, 2012; Buadi et al., 2013; Preyde et al., 2013; Lyons et al., 2014; El-Masri et al., 2019). These different motivations and purposes can create power imbalances that shape project outcomes.
Furthermore, power imbalances between academics and practitioners can hinder research translation. We found that academics frequently have more power and this inequality hinders the formation of meaningful partnerships (Lyons et al., 2014; Sewell et al., 2014; Smith and Jenkins, 2015; Jull et al., 2017). We found that inequalities can even occur in approaches with high commitment to participation and collaboration. In IKT, academics decide how to use funds which creates skewed power distribution and limits the impact of research utilization (El-Masri et al., 2019). In contrast, PAR scholars actively promote a shift in power dynamics to change hierarchical relationships to horizontal relationships by valuing practical knowledge (Lyons et al., 2014), shared learning (Gutberlet, 2015), and reflexive inquiry (Beh et al., 2013). In the PAR research reviewed, humility is an integral aspect of research collaborations and involves reflection on one’s positionality. Some scholarship emphasized the necessity to shift from ‘power-over’—in which there are unequal relationships to - to ‘power with’—in which all partners contribute (Turnhout et al., 2020). The ‘power with’ is revealed in the power to develop capacity in oneself and others to transform hierarchies of power around knowledge production (Toukan, 2021). However, hierarchies of knowledge, institutional structures, and geopolitics shape the formation of research partnerships (Fransman and Newman, 2019). If the power imbalances are addressed in research collaborations, then there is great prospect to achieve effective and sustainable solutions to societal problems (Stokols, 2006). Similarly, we found that some academics reflected and addressed the power inequalities in research translation while others did not. Hakkarainen et al. (2021: 12) made a call for academics to employ tools to ‘assist them in self-positioning on a continuum of ontological, epistemological, and philosophical structures’ to be able to ‘determine whether the assumptions and values are in conflict’ with the goals for research impact. Our Research Translation Continuum is a visual conversation starter for academics and practitioners to reflect on assumptions and power differences in research collaborations. Research uptake for global development is not a linear process from the Global North to the Global South, but a series of power-laden processes, connections, and relationships manifested in many ways.
V. Conclusion
In this review, we analysed five research translation approaches relevant to international development (technology transfer, participatory action research, evidence-based policy making, knowledge translation, and integrated knowledge translation) and harnessed the collective lessons learned from each of the approaches. We identified four interrelated factors that influence the research translation approaches: (a) intention of evidence use (b) commitment to partnership, (c) understanding of context and (d) investment of time and resources. We also found that research translation incorporates a range of approaches from proactive to post-facto translation within a Research Translation Continuum.
Our review fills a gap in bringing together literature on implementation science and development studies. Given the complexity in addressing global challenges, there is a great need for a conceptual framework to guide the application of evidence into development practice and policy. The Research Translation Continuum, together with the four interrelated factors, enable academics and practitioners to recognize the ontological and epistemological choices in all research translation approaches and reflect on external factors that may inhibit research uptake and utilization. Scholars increasingly recognize that partnerships between knowledge users and producers is often insufficient to ensure that research is used to inform decision making (Meehan et al., 2018; Oliver et al., 2019; Mach et al., 2020). Additionally, partners may hold different epistemic and ontological views which shape the purposes and design of collaborative research projects. Therefore, co-productive research collaborations may need to ask uncomfortable questions about participation, legitimacy, and knowledge reconciliation (Mach et al., 2020: 32). We find that there is great potential in the collective knowledge of academics and practitioners to improve development practice and policy; however, it is necessary to critically reflect on our own and others’ practices (Haraway, 1988: 579). Each actor in research translation has a different position, understanding, and course of action, and therefore it is not possible to separate the production and application of knowledge (Ludwig et al., 2021).
As development institutions and funders increase calls for evidence-based practices and research impact, there needs to be robust discussions and empirically grounded research on the research translation processes between academics and society (Williams, 2013). Therefore, we call upon future scholarship to continue to examine theories and impacts of research translation approaches because knowledge production processes shape what we know and who benefits (Toukan, 2021). Research collaborations have the potential to impact products, policies, and services and also change the lives of people, and the health of the environment (Phipps et al., 2016: 32); however, there are very few systematic studies that ‘reveal what co-production actually produces for either its participants or other intended beneficiaries’ (Oliver et al., 2019; Jagannathan et al., 2020). There is very limited scholarship on collaborative processes (Klenk and Wyatt, 2015), theories to sustain evidence-based practice utilization (Birken et al., 2020), or the long-term effectiveness of research partnerships (Lemos et al., 2018).
In sum, future research needs to systematically examine theories, processes, and long-term outcomes of research translation, particularly as they relate to international development. In the meantime, our review fills a gap in bringing together literature on implementation science and development studies and proposes the Research Translation Continuum as a tool for critical self-reflection and awareness as a first step in advancing research for development.
Footnotes
Acknowledgements
This study draws on the experiences, knowledge, and insights of colleagues from the LASER PULSE consortium including Purdue University, Catholic Relief Services, and the University of Notre Dame. Funding for the study came from USAID through the LASER PULSE project. This publication was made possible through support provided by the Innovation, Technology and Research Hub of the U.S. Agency for International Development, through the LASER PULSE Program under the terms of Cooperative Agreement No. 7200AA18CA00009. The opinions expressed herein are those of the author(s) and do not necessarily reflect the views of the U.S. Agency for International Development.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
The Innovation, Technology and Research Hub of the U.S. Agency for International Development provided the support through the LASER PULSE Program under the terms of Cooperative Agreement No. 7200AA18CA00009.