Untitled Document

Introduction: Identifying the Qualities of Research–Policy Partnerships in International Development – A New Analytical Framework*

James Georgalakis1 and Pauline Rose2

Abstract
This article sets out a framework for analysing research–policy partnerships for societal change in international development settings. It defines types of change associated with engaging research evidence with policy and practice and draws on existing literature to explore how partnerships between researchers, intermediaries, and evidence users may be better understood. The proposed framework sets out three interconnected qualities of effective partnerships: (1) bounded mutuality, (2) sustained interactivity, and (3) policy adaptability. We apply this framework to the articles included in this IDS Bulletin describing ESRC-DFID-funded research projects in a variety of international development scenarios.

Keywords: research partnership, policy engagement, impact, international development, evidence-informed policy, research uptake.

1 Introduction
This analysis of research–policy partnerships builds on our understanding of the potential impact of knowledge on development processes. These concepts are based on previous work undertaken by the ESRC-DFID-funded Impact Initiative for International Development Research, for which the authors of this article are its Director (James Georgalakis, Institute of Development Studies) and its lead for education (Pauline Rose, REAL Centre, University of Cambridge). Over the past four years, we have worked with over 200 social science projects in 79 countries seeking to maximise their impact. One of the most persistent messages to emerge from this work has been around the perceived benefits of partnerships between academics, communities, and policy actors. However, there appear to be many meanings of partnership in the context of collaborations around research for policy. We became increasingly curious about the possibility of identifying the key characteristics of partnerships capable of influencing how evidence is produced, accessed, and used to reduce poverty and inequality.

In order to develop an analytical framework with which to assess the learning arising from partnerships designed to have an impact beyond academia, we look at three areas of literature and associated practice. Firstly, there is the wide-ranging literature exploring the relationship between research and impact on policy and practice. In particular, this identifies interactive models of research-to-policy processes that are directly relevant to our focus on partnerships. Secondly, there is theoretical and practice-based work exploring partnerships in international development more broadly (rather than specifically between researchers and non-research users). A number of analytical frameworks for assessing the efficacy of such partnerships, particularly between Northern and Southern institutions, have some relevance to our inquiry. Finally, we consider the smaller, but growing, literature on the societal impact of partnerships between researchers and non-research users or knowledge intermediaries to which we hope this article will make a useful contribution. What becomes clear from a review of this literature is that, despite a variety of approaches to try and understand evidence use for policy formulation, implementation, and practice, the dynamics of partnerships between researchers and non-research users need to be better understood. It is this aspect that forms the focus of our article.

2 Definitions of research impact
The Impact Initiative has aimed to increase the uptake and impact of research from two major research programmes jointly funded by the UK’s Economic and Social Research Council (ESRC) and the Department for International Development (DFID): the Joint Fund for Poverty Alleviation Research and the Raising Learning Outcomes in Education Systems Research Programme. The starting point is that the ESRC-DFID strategic partnership is funding good-quality, policy-relevant research for which identifying ‘pathways to impact’ is integral.3 The Initiative developed a wheel of impact derived from the definitions of research impact adopted by the ESRC: capacity building, conceptual, and instrumental impacts. From the outset, the Impact Initiative’s strategy added to this ‘networks and connectivity’, as another anticipated outcome of the programme (Figure 1). This framework recognises the different modes of research impact as interrelated.

Although we regard all four modes of impact as desirable outcomes from research–policy collaborations, the focus on partnership inevitably recommends that we pay particular attention to the segment of the wheel that relates to building and strengthening networks. This gives rise to some of the reoccurring questions raised by the study of partnerships in development: whether new networks and strengthened relationships are an end in themselves, and/or a means to the end of research uptake and evidence use; whether they are best accomplished informally or also through more formal contractual relations; and how to address potential power dynamics within them.

3 Understanding research and policy: From linear to interactive concepts
The considerable interest in translating research into policy and practice is not new. It is apparent, however, that it has become a greater focus of attention in recent years. A recent systematic review of literature related to how academics can increase their impact on policy finds that, of 86 academic and non-academic publications dating back to the 1950s, 34 were published in the last two years (Oliver and Cairney 2019). During the 1970s and 1980s, there was growing recognition in the academic literature that there is no simple linear relationship between research knowledge and policy change. Alternative models ranged from theories suggesting that policy was just as likely to determine research agendas as the other way around, that social science gradually percolates into public consciousness (Weiss 1979), to a more complex interdependency between science and society (Jasanoff 2004). Despite these advancements, commitment to linear instrumentalist models of research use remained dominant in the public domain, particularly in the UK and Canada whose governments were at the forefront of an evidence-based policy movement in the 1990s that had originated in clinical practice (Gwyther 2014).

In the international development sector, increasing attention is being paid to the attribution of policy and practice change to specific research studies as a means of demonstrating academic excellence and value for money. This is apparent in DFID’s approach to investing in research (ICAI 2018). In this field, conceptualisations of a gap between research communities on the supply side and policy networks on the demand side has tended to result in recommended practices to bridge this gap that are largely technical in nature. In particular, communication tools and training for researchers and policy actors, as well as development of digital initiatives to increase the accessibility and availability of research has become commonplace (Georgalakis et al. 2017). However, these more technical approaches to achieving research uptake have come under increasing critique from those who argue that policy processes tend to be messy, and require ongoing engagement. This challenges the potential of bridging an evidence–policy gap with better communication or training (Cairney 2016; Parkhurst 2017).

Of all the approaches to research use to emerge in reaction to both these linear theories on the one hand and interpretivist accounts of socially constructed knowledge on the other, interactive models are perhaps one of the most useful in navigating complex policy and practice environments (Nutley, Walter and Davies 2007). Rather than focusing on autonomous streams of problems, policy, and politics which collide at key moments (Kingdon 1984), these models set out what Huberman calls ‘sustained interactivity’. This approach recognises the blurring of boundaries between research producers, intermediaries, and users (Huberman 1994). This social and interactive process is also supported by well-established theories such as the role of policy and epistemic communities (Haas 1992) and the power dynamics of knowledge (Lukes 2003; Gaventa 2006).

4 Developing mutual agendas through research–policy partnerships
Whilst understanding how research and policy interact is fundamental to understanding the role of partnerships in leveraging societal impact, it is also informative to explore the nature of partnerships themselves in international development contexts, including how partners come together around mutual agendas. Even the meaning of the term ‘partnership’ is highly contested. In international development, there are concerns around terms like partnership becoming an empty buzzword. Whilst the language of partnership may be deployed, the reality is that politics and power dynamics are still at play (Cornwall and Brock 2005). As the imbalance of power may itself affect the organisational identity of those dependent on external funds, some have chosen to avoid the term partnership for a more neutral term of ‘relationship’ (Batley and Rose 2011).

Given the North to South power dynamics of development aid, it is perhaps to be expected that the most common approach to exploring international development partnerships is an analysis of power between donors and national actors, such as national governments and non-governmental organisations (NGOs), intermediary delivery agencies, the private sector; and local civil society groups (Brinkerhoff 2002; Lister 2000). The rationale for such partnerships in international development settings referred to in the literature are diverse, including mutual learning, the leveraging of expertise and local knowledge, building local capacity, and achieving value for money (Morse and McNamara 2006; DFID 2011).

However, until recently, this literature rarely went beyond an assessment of relatively binary donor–national relationships. In research initiatives designed to influence policy or practice, there is not necessarily an international donor playing an active role, and partnerships may take a whole variety of forms, including South to South governmental and NGO collaborations, and partnerships between Northern and Southern research institutions. Power dynamics affect relationships between researchers themselves and between them and policy actors, such as those in national governments and multilateral agencies. The asymmetry in power between different actors, and the influence this has on achieving desired outcomes, does not reside solely on traditional development paradigms. As Henning Melber points out, power inequalities penetrate every social context and will always affect how knowledge is generated and used (Melber 2019).

Brinkerhoff’s evaluation tool (in the context of partnerships between government and non-government actors) helpfully provides a framework that acknowledges both normative and pragmatic approaches to defining the key dimensions of partnership for sustainable development (Brinkerhoff 2002). Mutuality, she argues, or mutual interdependence, needs to be present to maximise the benefits for each party. This means mutual commitment to the objectives of the collaboration and a strategy that is compatible with each actor’s mission, values, and goals. However, Brinkerhoff is interested in more than the function and insists that development partnerships should also be assessed on their performance. Did they achieve what they set out to achieve?

For partnerships that are focused on bringing together researchers with NGOs, community-based organisations, or local researchers, a great deal of the reflection focuses on the principles of engagement. They seek equity, respect, mutual agendas, and trust (Baker et al. 1999). A movement has emerged supporting a vision of equitable research partnerships between Northern researchers and Southern researchers or practitioners. Leading this movement has been an innovative network of social scientists and international non-governmental organisations (INGOs) who have set out eight principles of fair and equitable research partnerships (Newman, Bharadwaj and Fransman, this IDS Bulletin, and Fransman and Newman 2019). Although their work was initially focused on achieving equity and fairness (Fransman, Newman and Cornish 2017), in this issue they also argue that if research is to have an impact, it must build directly on the knowledge and experience of those working at the coal face to challenge poverty and contribute to social justice. They optimistically argue that although a fair and equitable mutual agenda might not always be apparent at the outset of such partnerships, with time and patience they can develop. They propose that research excellence is more commonly understood from the perspective of Northern-dominated definitions which privilege those from relatively well-resourced universities with better access to high-impact journals. They argue for an embedded systems model of research impact that recognises practical experience, such as the experiential learning of development practitioners, as an important source of knowledge.

Whilst recognising that equitable partnerships are desirable and morally imperative, they may not always be a necessary condition for innovative research and societal relevance. Newman et al.’s definition (this IDS Bulletin) of equitable appears to require redressing knowledge hierarchies between different actors, engaging critically and transparently with the politics of knowledge, and valuing all voices equally. However, under some circumstances, in real-world interaction between science and policy, the tensions, trade-offs, and compromises experienced when the research and policy come together may still lead to progressive change (Cairney 2016). However, in many policy contexts this may look more like temporary convergence around a common policy agenda than equitable partnership.

Mark Swilling, inspired by the work of Hajer, provides a compelling case for researchers to exercise some ‘reflexive caution’ when seeking to partner with policy actors to influence positive change (Swilling 2014; Hajer 2005). He describes researchers’ mobilisation around particular public policy issues in what may be a combination of advisory services and policy advocacy as a stage-managed process. This performance is controlled by the dominant partners. Therefore, although convergence of agendas may have occurred, institutional conditions and the broader political and social context, in which each partner is governed by a separate mandate, places limitations around mutuality. Therefore, rather like Herbert Simon’s ‘bounded’ rationality of decision makers, in which their understanding and use of evidence is shaped by political, social, and cultural factors, even in the best case scenario, mutuality of agendas in a research–policy partnership also appears bounded (Simon 1972).

5 Partnerships’ engagement with policy processes
Forms of engagement between researchers and evidence users with the aim of achieving impact might differ depending on the disciplinary and associated methodological approaches of the research. For some researchers in development studies, co-production of research and meaningful engagement with partners is viewed as an end in itself. From this perspective, research is seen as development, not for development. Partnership is seen as a democratic tool that aims to promote equity and inclusivity. This approach has been central to concepts of participatory development that primarily belong to civil society advocates and scholars working on participation and empowerment (Mohan 2008). Similarly to Newman et al. (this IDS Bulletin), many participatory researchers argue that the failure to co-produce evidence with those most affected by the issues can have adverse consequences beyond the failure to achieve the ethical principle of cognitive justice. The negative effects of scientific-based agricultural reform in East Africa in the 1970s that played to the interests of a political elite and Western-dominated science over pastoralists’ local knowledge is one such example (Scott 1998). The power of participation and local partnership has become a dominant normative perspective in development studies, although the challenges of influencing policy at scale through participatory approaches remain enormous (Chambers 2017).

Our focus here, however, is on research–policy partnerships in all their forms within the remit of ESRC-DFID-funded research, not just those employing a participatory or community-based approach. We therefore recognise that approaches to research methodology that promote co-production are not the only way that researchers aim to achieve impact. Other approaches might envision a clearer division of labour between different partners in the pathways to impact process. For example, approaches to knowledge mobilisation in the early part of the new millennium were increasingly focused on the role of intermediaries, conceptualised by some as ‘boundary partners’ (Cash et al. 2003) and by others as brokers (Datta 2012). These partners are presumed to have key abilities and motivations around the translation and exchange of knowledge with policy and professional actors. This is brokerage built on attempts to move beyond a unidirectional model of knowledge transfer. Common strategies in sectors such as global health attempt to overcome ‘impermeable barriers’ between researchers and policymakers through fundamentally relational processes such as building multidisciplinary teams of academics, practitioners, and government officials (Sen et al. 2017).

This brings us to consider network analysis which focuses on the interactions themselves and related changes in individual relationships. Network analysis aims to reveal deeply rooted individual and networked relationships that may have indirect impact on evidence-informed policy and practice, regardless of the research approach being used (Jessani et al. 2018). It identifies that deliberative and ongoing interactivity is an essential part of engaging in non-linear, multifaceted policy formulation and implementation processes (Datta 2012). This brings us back to our earlier points on interactive theories around evidence use, and recommends that an assessment of partnerships for impact looks beyond the mutuality of agenda. Sustained interactivity that strengthens networks and results in changes in relationships appear equally important to promoting evidence use.

However, the emergent dominance of interactive theories of research to policy and practice dynamics have not gone unchallenged. For pressure groups and advocacy organisations, the existence of connections between research producers and users, and productive relationships between key individuals and institutions is important but on its own inadequate (Mayne et al. 2018). They argue that, in addition, advocates of evidenced-informed policy need to be capable of framing information so that it meets the demands of policymakers, often operating beyond the specialist policy community that partnerships are regularly engaging with. Paul Cairney describes how policymakers operate in an environment full of uncertainty where they must make decisions based on ambiguous information. This requires the simplification of complex problems and, as mentioned above, bounded mutuality. In this political context, longstanding partnerships with policy actors who are often mid-level civil servants and policy professionals, does not make it any less important to construct compelling policy-friendly narratives and identify key influencing opportunities in the political sphere (Cairney 2016).

Whether operating as a policy outsider, insider, or both at once, the capability to adapt for policy requires good timing, policy-relevant research, and the ability to contextualise research evidence for live policy issues (Oliver and Cairney 2019). It also relies on having individuals positioned appropriately as members of the collaboration (Carden 2009). These qualities do not automatically emerge from mutual agendas and interactivity, and so should be given special attention. Partners in the policy sphere, such as policy professionals in government ministries or multilateral agencies, can provide access to closed policy spaces and may be privy to forthcoming announcements or new initiatives. Above all, they understand how policy issues are perceived and what opportunities might exist.

Research–policy partnerships can also be relatively technical or niche, partners may not be politically influential within their own institutions, or the issues may be low on the agenda of senior decision makers. Policy adaptability requires compromise, negotiation, and often an appetite for risk. Key policy spaces or moments must be prioritised despite research institutions working to very different timescales to policy actors and practitioners (Mayne et al. 2018). These interactions rarely appear consistent with the linear approach of researchers simply aiming to disseminate their research to decision makers. Donors and policy partners shape research agendas and funding opportunities and, as in the case recently in Tanzania (Dausen 2018), may even legislate around which types of evidence are politically acceptable. Policy partners can act as gatekeepers, filtering out politically awkward evidence or prioritising their own agendas.

The policy engagement capacity of partnerships appears closely related to the roles and responsibilities of their members. Partners may be constrained by their perceived roles as politically neutral, as in the case of government officials, or as neutral brokers, as in the case of academics. Some partnerships deliberately exploit these differences in the partners’ mandates to their advantage, letting campaigning organisations lobby senior decision makers, and researchers provide expert advice to government officials whilst informing advocacy objectives (Pittore et al. 2016). You cannot predict the future and prepare for unforeseeable policy opportunities. However, you can create a partnership with the resources, members, networks, and knowledge to adapt fast to changing circumstances.

6 Proposed framework
Building on this review of literature related to an understanding of the nature of research–policy partnerships in international development settings, and the interactions between research and policy and practice, we propose that any assessment of learning arising from such partnerships should critically investigate: (1) bounded mutuality, (2) sustained interactivity, and (3) policy adaptability. Each of these qualities appear capable of contributing to processes which may result in changes in: (i) capacity to produce and use evidence, (ii) relationships that connect up the creation, interpretation, and use of evidence, (iii) knowledge and awareness of the implications of research, and, finally (iv) evidence uptake. It could be argued that high-quality evidence is a pre-requisite for research to improve policy processes, although this itself requires defining, and in a development context will almost certainly be subject to different views on what quality is (Moore et al. 2017).

Each of the three qualities of effective research–policy partnerships identified in our article can occur independently of each other. However, they are likely to be reciprocally reinforcing, and their combination offers the greatest opportunities for achieving the desired change. And, as many of the articles in this IDS Bulletin highlight, cutting across the three components of the framework is the importance of building and sustaining trust. As Hinton, Bronwin and Savage (this IDS Bulletin) indicate, for policy uptake to occur, ‘the process of research matters alongside the findings’. We therefore suggest that relative weakness in one partnership quality compared to another does not automatically render partnerships obsolete.

Strength in a mutual agenda, for example, may be more crucial to a specific initiative than well-established networks. Likewise, partnerships built on partially aligned agendas may still successfully leverage awareness of a body of evidence at a critical political moment. However, success is most likely where the three characteristics converge. We therefore need to look at all three qualities together, within the broader political and social context in which they occur, and assess how they may accomplish the intended (or unintended) outcomes of the partnership. What is of central importance to the framework presented here is that the intended impacts, whether conceptual, capacity related, relational or instrumental, are understood as being interrelated to the core research partnership qualities themselves. The partnership is as much a product of, as a contributor to, the external environment.

It is also worth noting that investing in building partnerships may not always be a win-win. The time required will be at the expense of other activities in which researchers, policy actors, and practitioners each engage. For example, in academia, there is a healthy debate about whether such engagement may be at the expense of time spent developing and publishing high-quality research which is, itself, a pre-requisite for expanding theoretical and practical knowledge. There are also reputational risks where academics heavily engaged in informing policy are sometimes criticised for being an academic ‘lightweight’ (Oliver and Cairney 2019). Importantly also, where the evidence contradicts the direction of policy driven by political imperatives, the position of researchers and policy actors within the networks may be jeopardised. This raises a question of the costs and benefits of researchers being insiders within policy processes, often in an advisory role, or whether they are likely to be more influential by remaining outside. There is no clear answer to this.

7 Applying the proposed qualities of research–policy partnerships to ESRC-DFID projects
In developing a framework for effective research–policy partnerships in international development settings, we take as given the case for equity and fairness. The importance of such equitable partnerships is apparent in all the articles in this IDS Bulletin. Nonetheless, whilst power dynamics within partnerships cannot and should not be ignored, our framework is primarily focused on the elements of successful partnerships for contributing to positive change. As many of the articles acknowledge, we also recognise that research–policy partnerships require resources and capacity, the lack of which is a major barrier to success.

7.1 Bounded mutuality
We have identified bounded mutuality as a key component of successful partnerships. A starting point for this is identifying the extent to which there is a common understanding of the policy problem and set of values underpinning the collaboration, even if partners are mandated differently. In research–policy partnerships, this can occur where demand for and supply of evidence converge. This could be, for example, around a shared agenda of improving the quality of education or health, where policy actors or practitioners are keen to draw on evidence that will inform their design of programmes.

Mutual agenda can develop where researchers and policy actors or practitioners are involved in the co-production of research. This was the case for an international partnership focused on the development of a regional monitoring ‘toolkit’ for pro-poor health policy for the Southern African Development Community (SADC) (Yeates et al., this IDS Bulletin). Yeates et al. identify the importance of the partnership built on a mutual agenda, namely a common recognition of the problem of the high social and economic cost of the disease burden amongst countries within the SADC. This shared recognition, together with a demand for the research by the SADC Secretariat presented a ‘live opportunity’.

Johnson et al. (this IDS Bulletin) further highlight the bounded nature of mutuality. In their case, a shared vision of the political and social justice issues of promoting youth rights through listening to their voices, needed to recognise that the partners involved had different personal, organisational, social, and political agendas. For their project, such partners included NGOs, researchers in Ethiopia, Nepal, and the UK, as well as young people themselves. As the authors identify, a mutual understanding of the value of research amongst different partners cannot necessarily be taken for granted. In their case, the NGOs involved in the project supported the research to varying degrees, with some needing to be ‘convinced of the value of the research as opposed to, for example, longstanding intervention strategies or participatory action processes’.

7.2 Sustained interactivity
Sustained interactivity between the partners in the research–policy process itself and the wider environment that they are focused on is a second important condition for effective partnerships. ‘Sustained’ means building such engagement from the very start of the research process and beyond. The most successful research partnerships continue after the projects have ended: the partnerships see value in working with each other and so look for other opportunities as part of a longer-term, iterative process, rather than merely seeing engagement beyond academia as a function of the dissemination of results. Like bounded mutuality, sustained interactivity may be built on transdisciplinary co-production within the research process itself. For Johnson et al. and Yeates et al. (this IDS Bulletin), interaction between researchers and policy actors and practitioners existed before the development of their ESRC-DFID research proposals. In both examples, sustaining this interactivity via a variety of platforms to ensure smooth communications throughout the project and beyond was essential for building trust and combined ownership of the work.

For Yeates et al., relationships in the UK, in Africa, and with the SADC Secretariat itself in developing a toolkit were based on participatory principles. They recognise, however, that there were some limits to the participatory approach inasmuch as the broad goals were pre-defined by the terms of the funding grant. An important aspect of this interaction was the mutual learning that occurred, such as in the capacity of partners to use evidence and evidence-use behaviours in policy development. Interaction also occurred in the process of developing publications, which was both seen as beneficial for developing trust as well as for self-reflection of government officials on the efficacy of their health programmes. As the authors note, such sustained interactivity needs to not only begin before a project starts, but also continue beyond the lifetime of a particular project.

Processes for building and maintaining diverse individual and networked connections are an important consideration in planning for impact. Advisory groups, committees, learning events, and regular meetings are one approach, which ideally need to directly inform how the research and related policy issues are understood and framed, and how the findings are developing on an ongoing basis. Such structures were found to be important for the success of Johnson et al.’s work in Ethiopia and Nepal (this IDS Bulletin). They note that meetings to share approaches, together with platform and spaces for dialogue, to agree findings and discuss potential impact, has been important, both through face-to-face and remote engagement.

Beyond these more formal structures and approaches, relationships may be fluid and informal, with knowledge and understanding passing both ways. Hinton et al. (this IDS Bulletin) compare two examples of ESRC-DFID-funded research as part of the Raising Learning Outcomes (RLO) programme: one where there were pre-existing informal relations between NGO partners of the research with both DFID and the government, and one without prior experience in the country. They identify that the former was able to open up opportunities for its research to inform DFID and the government’s work much more effectively.

In some instances, attempts of sustained interactivity may be politically charged. However, it remains important in such situations. For Mulugeta et al. (this IDS Bulletin), an iterative, interactive process was needed for gaining pastoralist perspectives and using this to engage and inform district, regional, and national government officials in Ethiopia of their findings. The authors note that the Ethiopian government has a strong demand for policy-relevant research, even if this has been seen as contributing to government-led development initiatives rather than critiquing it.4 In the context of their work, a fundamental difference in understanding of the purpose of research between the authors and government officials meant that the relationship was not straightforward. The starting point of their research, on a politically sensitive topic of understanding the causes of conflict in pastoralist areas, was based on a fundamental difference in understanding of the causes of conflict: for the researchers, it related to a ‘rational response to environmental change and state-led dispossessions, while to government, it is an expression of “backwardness” and “irrationality” ’.

Indeed, partnership to achieve a conceptual shift in understanding was at the heart of their research. The iterative and evolving approach to partnership was essential for building trust. This, in turn, opened the door for dialogue around differing perspectives with the intention that the evidence could help to shift understandings amongst government partners of how pastoralists understand conflict in their communities. It would seem that the potential for shifting understandings was very much facilitated by the fact that two of the Ethiopian researchers were part of the Omo-Turkana Research Network,5 which came about as a result of their ongoing informal engagement at a local level of government within the Southern National, Nationalities, and People’s regional state of Ethiopia.

For Kett et al., Johnson et al., and Mulugeta et al. (this IDS Bulletin), interactivity involves co-production through the direct engagement of marginalised populations (in these cases, people with disabilities, marginalised young people, and pastoralists, respectively). For all of them, this is a means to build trust amongst the partners, and credibility of the evidence. Kett et al. (this IDS Bulletin) identify how their research on disability in Liberia, Kenya, Uganda, Sierra Leone, and Zambia has benefited from active collaboration and co-construction with (rather than more passively including) advocates, practitioners, and policymakers. As the authors point out, it is not necessarily that people with disabilities (or others who are the focus of the research) will be gathering and analysing data, but they can play a key role in informing the design and understanding the implications of the results of the research, for example. They also highlight how engagement of people with disabilities throughout the process is a way to build capacity and the confidence of other partners in understanding their perspectives. This requires ongoing, genuine interaction and not tokenistic involvement (such as through data collection) which is likely to lead to resentment rather than the building of trust.

However, co-production of research is not always necessarily the approach used for achieving sustained interactivity. Chowdhury (this IDS Bulletin) provides an example of how sustained interaction between BRAC’s Research and Evidence Division and the Oral Therapy Extension Programme resulted in a breakthrough in the effective use of oral rehydration therapy to address diarrhoea that was a major killer in Bangladesh in the 1980s. In this example, independent research was seen as important ‘not just for research’ sake but to solve a problem or to improve delivery of interventions’ (Chowdhury, this IDS Bulletin). He notes that sustained, ongoing engagement of researchers from the pilot of the programme through to its scale-up enabled ‘mid-course corrections’.

Similarly, an ESRC-DFID RLO project focusing on improving literacy in northern Uganda has been built on a partnership between Mango Tree – an NGO established in 2009 – and researchers at the University of Illinois who undertook a randomised controlled trial to identify the impact of the interventions (see Hinton et al., this IDS Bulletin). Whilst the research was undertaken independently, the design was informed by engagement with the NGO who in turn has used the results both to inform their own practice as well as for wider engagement with DFID and government to inform policy change. Interactivity between NGOs and researchers in such independent research may not, however, always be straightforward, as this example indicates. NGOs are likely to regularly change their interventions for a variety of reasons, which can inevitably create challenges where researchers are aiming to identify differences between those receiving an intervention and those not.

7.3 Policy adaptability
Adaptability refers to the capability of partnerships to identify key influencing spaces and re-frame evidence for specific policy audiences. It also enables partners to adapt when the policy environment changes, such as in the light of shifts in the political or social contexts. Such adaptability means that partnerships are in a strong position to capture policy windows of opportunity swiftly as they arise. It might also involve collaboration with boundary partners (outside the core partnership) such as policy advocates, or other brokers such as the media who can incorporate the evidence into their own campaigns and priorities. For policy impacts, this is perhaps best understood as the ability of research partnerships focused on policy areas to provide responses to perceived policy dilemmas. This can relate to longer-term agenda setting as well as more instrumental impact on policy deliberations.

Hinton et al. (this IDS Bulletin), writing from the perspective of DFID advisers, note the importance of taking account of the differing incentives of partners – with researchers primarily recognised on the basis of high-quality publications, while governments (with short-term political cycles) having a more immediate need to identify ‘what works’ from a value-for-money perspective. The responsiveness of researchers, potentially with the involvement of brokers, can help to bridge this gap. Hinton et al. argue that donors themselves can play the role of knowledge brokers and translators (or ‘super communicators’ as they call them) between researchers and governments. In Uganda, for example, even though DFID did not identify itself as a knowledge broker, it clearly had a key role to play in engaging the research it had funded within national policy debates. Having the evidence on the importance of early literacy at its disposal enabled DFID colleagues to act, as key opportunities arose in design phases and strategic plans in relevant areas.

In India, the DFID adviser noted a potential mismatch between the types of publications researchers prepare and the need for research to be articulated in a way that policymakers, who are not experts, can engage with. This can be as much about framing for policy and timing as simplifying. Bridging the gap is a role that he, as a DFID adviser, saw himself as playing, including, for example, recognising the best way to communicate with government officials where sensitive issues emerging from research might arise. However, as both the cases identify, donor advisers in-country are themselves time-constrained, incentives are not necessarily aligned with them actively engaging with research, and they might not possess the skill set needed to engage effectively with research.

Super communicators can, therefore, take other forms. In the India case, researcher engagement with the media as well as ongoing engagement by locally based academics and NGOs, has been fruitful in promoting policy uptake by government. Hinton et al. (this IDS Bulletin) note the importance of agile ways of working, as windows of opportunity, such as due to new political appointments, can be fleeting. An example of this comes from Johnson et al.’s work in Ethiopia (this IDS Bulletin). The formation of a new government presented the research–NGO partnership with a unique possibility to ensure that marginalised youth voices were heard in the development of the new youth policy.

Yeates et al. (this IDS Bulletin) recognises that despite a clear mutual agenda with the SADC Secretariat and the relatively good timing of their project, there was far more limited engagement with national-level policy actors than specialists working at the regional level. This restricted the partnership’s scope for engaging in policy processes at a national level. In order to be able to mobilise research evidence for policy, the target decision makers are much more likely to be receptive when they have some sense of ownership of the knowledge. In the end, the toolkit they were working on came too late to directly impact on the current policy process that had helped frame their research. Although, two years later, it still proved relevant to the revision of a regional monitoring and evaluation initiative subsequently led by the SADC Secretariat. This demonstrates that a partnership’s ability to operate effectively in the policy sphere is as much about connections with critical boundary partners as the extent to which projects are synchronised with policy formulation timescales. The first of these is all part of adapting to policy, but the second is often beyond the control of specific research initiatives.

8 Conclusions
This article has set out to review partnerships between researchers and evidence users aimed at achieving positive change in international development settings. Drawing on the existing literature, an analytical framework based on three interrelated partnership qualities of bounded mutuality, sustained interactivity, and policy adaptability emerged. These characteristics are shaped by an understanding of evidence-into- policy processes as fundamentally social and interactive, and underpinned by political context, social norms, and power. When applied to case studies of partnership from ESRC-DFID-funded research, all three partnership qualities have been found to exist to varying degrees and in different ways. Although there is evidence that these conditions helped bring about desired changes in terms of evidence use, capacity, knowledge, and relationships, the comparative strength of these qualities in specific partnerships also suggests that even more could have been achieved if they had been more deeply rooted. And, in many cases, the principles emerge as part of the research process, rather than being planned from the outset.

We therefore conclude by proposing that the use of this analytical framework at the design stage of a research process could increase the viability of a partnership, by taking into account the importance of mutuality, interactivity, and policy adaptability from the outset. We hope others will seek to validate this concept with existing methodologies and literature, and apply variations of it to their own work.

Notes

* This issue grew out of the Impact Initiative for International Development Research which seeks to maximise impact and learning from ESRC-DFID’s Joint Fund for Poverty Alleviation Research and their Raising Learning Outcomes in Education Systems Research Programme.

1 Institute of Development Studies, UK.

2 Research for Equitable Access and Learning (REAL) Centre, University of Cambridge, UK.

3 ESRC funding is based on three criteria: (1) quality – all ESRC research awards are made in open competition, subject to transparent peer assessment at the outset and evaluation on completion; (2) impact – the research is intended to make a difference; and (3) independence – independence and impartiality of ESRC-funded research is viewed as key. For further information, see: https://esrc.ukri.org/about-us/what-we-do/.

4 There has been a change in government in Ethiopia since this research took place.

5 Omo-Turkana Research Network: an international consortium of social and environmental scientists researching the impacts of hydrological, agricultural, and social change on the people and ecosystems surrounding the Lower Omo Valley and Lake Turkana. See: www.canr.msu.edu/oturn/.

References
Baker, E.A.; Homan, S.; Schonhoff, R. and Kreuter, M. (1999) ‘Principles of Practice for Academic/Practice/Community Research Partnerships’, American Journal of Preventive Medicine 16.3.suppl: 86–93

Batley, R. and Rose, P. (2011) ‘Analysing Collaboration between Non-Governmental Service Providers and Governments’, Public Administration and Development 31.4: 230–39

Brinkerhoff, J.M. (2002) ‘Government–Nonprofit Partnership: A Defining Framework’, Public Administration and Development 22.1: 19–30

Cairney, P. (2016) The Politics of Evidence-Based Policy Making, London: Palgrave Macmillan

Carden, F. (2009) Knowledge to Policy: Making the Most of Development Research, Ottawa ONT: International Development Research Centre (IDRC)

Cash, D.W. et al. (2003) ‘Knowledge Systems for Sustainable Development’, Proceedings of the National Academy of Sciences 100.14: 8086–91

Chambers, R. (2017) Can We Know Better?: Reflections for Development, Rugby: Practical Action Publishing

Cornwall, A. and Brock, K. (2005) ‘What Do Buzzwords Do for Development Policy? A Critical Look at “Participation”, “Empowerment” and “Poverty Reduction” ’, Third World Quarterly 26.7: 1043–60

Datta, A. (2012) ‘Deliberation, Dialogue and Debate: Why Researchers Need to Engage with Others to Address Complex Issues’, IDS Bulletin 43.5: 9–16, https://bulletin.ids.ac.uk/idsbo/article/view/276 (accessed 15 April 2019)

Dausen, N. (2018) ‘Tanzania Law Punishing Critics of Statistics “Deeply Concerning”: World Bank’, Reuters, 3 October, www.reuters.com/article/us-tanzania-worldbank/tanzania-law-punishing-critics-of-statistics-deeply-concerning-world-bank-idUSKCN1MD17P (accessed 15 April 2019)

DFID (2011) DFID’s Approach to Value for Money (VfM), https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/49551/DFID-approach-value-money.pdf (accessed 15 April 2019)

Fransman, J. and Newman, K. (2019) ‘Rethinking Research Partnerships: Evidence and the Politics of Participation in Academic–INGO Research Partnerships for International Development’, Journal of International Development (advance online publication)

Fransman, J.; Newman, K. and Cornish, H. (2017) ‘Rethinking Research Partnerships: Discussion Guide and Toolkit’, Christian Aid, www.christianaid.org.uk/resources/about-us/rethinking-research-partnerships (accessed 23 April 2019)

Gaventa, J. (2006) ‘Finding the Spaces for Change: A Power Analysis’, IDS Bulletin 37.6: 23–33, https://bulletin.ids.ac.uk/idsbo/article/view/898 (accessed 15 April 2019)

Georgalakis, J.; Jessani, N.; Oronje, R. and Ramalingam, B. (2017) The Social Realities of Knowledge for Development: Sharing Lessons of Improving Development Processes with Evidence, Brighton: IDS, www.ids.ac.uk/publications/the-social-realities-of-knowledge-for-development-sharing-lessons-of-improving-development-processes-with-evidence/ (accessed 23 April 2019)

Gwyther, A. (2014) Supporting Better Research Impact Evaluation, Research to Action blog, 30 June, www.researchtoaction.org/2014/06/alex-gwther-supporting-better-research-impact-evaluation/ (accessed 15 April 2019)

Haas, P.M. (1992) ‘Introduction: Epistemic Communities and International Policy Coordination’, International Organization 46.1: 1–35

Hajer, M.A. (2005) ‘Setting the Stage: A Dramaturgy of Policy Deliberation’, Administration & Society 36.6: 624–47

Huberman, M. (1994) ‘Research Utilization: The State of the Art’, Knowledge and Policy 7.4: 13–33

ICAI (2018) DFID’s Approach to Value for Money in Programme and Portfolio Management, London: Independent Commission for Aid Impact,
https://icai.independent.gov.uk/wp-content/uploads/DFIDs-approach-to-value-for-money-ICAI-review.pdf (accessed 15 April 2019)

Jasanoff, S. (2004) States of Knowledge: The Co-Production of Science and the Social Order, London: Routledge

Jessani, N.S. et al. (2018) ‘Relationships between Public Health Faculty and Decision Makers at Four Governmental Levels: A Social Network Analysis’, Evidence & Policy: A Journal of Research, Debate and Practice 14.3: 499–522

Kingdon, J.W. (1984) Agendas, Alternatives and Public Policies, Boston MA: Little, Brown

Lister, S. (2000) ‘Power in Partnership? An Analysis of an NGO’s Relationships with its Partners’, Journal of International Development 12.2: 227–39

Lukes, S. (2003) Power: A Radical View, Basingstoke: Palgrave

Mayne, R. et al. (2018) ‘Using Evidence to Influence Policy: Oxfam’s Experience’, Palgrave Communications 4.1: 122, https://doi.org/10.1057/s41599-018-0176-7 (accessed 15 April 2019)

Melber, H. (2019) ‘Knowledge Production, Ownership and the Power of Definition: Perspectives on and from Sub-Saharan Africa’, in I.S.A. Baud, E. Basile, T. Kontinen and S. von Itter (eds), Building Development Studies for the New Millennium, Cham: Palgrave Macmillan

Mohan, G. (2008) ‘Participatory Development’, in V. Desai and R.B. Potter (eds), The Companion to Development Studies, 2nd ed., London: Hodder Education

Moore, S.; Neylon, C.; Eve, M.P.; O’Donnell, D.P. and Pattinson, D. (2017) ‘ “Excellence R Us”: University Research and the Fetishisation of Excellence’, Palgrave Communications 3

Morse, S. and McNamara, N. (2006) ‘Analysing Institutional Partnerships in Development: A Contract between Equals or a Loaded Process?’, Progress in Development Studies 6.4: 321–36

Nutley, S.M.; Walter, I. and Davies, H.T.O (2007) Using Evidence: How Research can Inform Public Services, Bristol: Policy Press

Oliver, K. and Cairney, P. (2019) ‘The Dos and Don’ts of Influencing Policy: A Systematic Review of Advice to Academics’, Palgrave Communications 5.1

Parkhurst, J. (2017) The Politics of Evidence: From Evidence-Based Policy to the Good Governance of Evidence, Abingdon: Routledge

Pittore, K.; te Lintelo, D.J.H.; Georgalakis, J. and Mikindo, T. (2016) ‘Choosing between Research Rigour or Support for Advocacy Movements, a False Dichotomy?’, IDS Bulletin 47.6: 101–18, http://dx.doi.org/10.19088/1968-2016.202 (accessed 22 May 2019)

Scott, J.C. (1998) Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed, Newhaven CT: Yale University Press

Sen, G.; Virani, A.; Iyer, A.; Reddy, B. and Selvakumar, S. (2017) ‘Translating Health Research to Policy: Breaking through the Impermeability Barrier’, in J. Georgalakis, N. Jessani, R. Oronje and B. Ramalingam (eds), The Social Realities of Knowledge for Development: Sharing Lessons of Improving Development Processes with Evidence, Brighton: IDS

Simon, H.A. (1972) ‘Theories of Bounded Rationality’, in C.B. McGuire and R. Radner (eds), Decision and Organization: A Volume in Honor of Jacob Marschak, Amsterdam: North-Holland Pub. Company

Swilling, M. (2014) ‘Rethinking the Science-Policy Interface in South Africa: Experiments in Knowledge Co-Production’, South African Journal of Science 110.5–6: 1–7

Weiss, C.H. (1979) ‘The Many Meanings of Research Utilization’, Public Administration Review 39.5: 426–31

© 2019 The Authors. IDS Bulletin © Institute of Development Studies | DOI: 10.19088/1968-2019.103

This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 International licence (CC BY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are credited and any modifications or adaptations are indicated. http://creativecommons.org/licenses/by/4.0/legalcode

The IDS Bulletin is published by Institute of Development Studies, Library Road, Brighton BN1 9RE, UK This article is part of IDS Bulletin Vol. 50 No. 1 May 2019 ‘Exploring Research–Policy Partnerships in International Development’.