University policy engagement bodies in the UK and the variable meanings of and approaches to impact

Over the last decade, there has been a proliferation of policy engagement bodies set up within universities worldwide. The present study focuses on the British experience of this phenomenon but with relevance to other contexts. Multiple factors are at play to explain this growth, from the Research Excellence Framework impact agenda (which assesses and ranks the quality of research in UK universities and has been echoed in other countries) to universities’ renewed focus on their civic mission, and a growing demand within policy and practice circles for more research evidence. Based on interviews with senior staff and analysis of their websites and documentary outputs, this article offers a comprehensive catalogue of university policy engagement bodies across the UK, and classiﬁes them into types based on their activities, outputs, impact, and staff. We enrich this categorization by examining the stories these bodies tell to explain how and why they have emerged, and the variable approaches they adopt as they seek to inﬂuence both academic and policymaking practices. In doing so, we develop a typology of university policy engagement bodies, and demonstrate how they seek to partake in changing the roles and identities of universities, and their relationship with policymaking.


Introduction
There has been a noticeable growth in the number of bodies being set up by universities in the UK, and elsewhere, to inform and impact national and local policymaking and practice, by transferring knowledge.Unlike traditional research support offices and discipline-specific research centres, these university policy engagement bodies-which have names such as policy institute, policy hub, or Policy@[university's name]-focus on brokering research outputs from their respective institutions into policy and practice.They are typically a centralized function, spanning the research portfolio of the university.Our research identified 46 such UK university bodies, and their practices vary widely (see Supplementary Appendix S1 for a list of these bodies and their characteristics).Some were established or have evolved to focus on developing bespoke opportunities for in-depth engagement between policymakers and academic researchers, for example through policy fellowship and secondment programmes.Others have added specific policy engagement functions to more traditional research support offices that assist individual academics to think about policy engagement and impact pathways in their work.
While the practice of policy engagement is not new-with individual academics having done so for decades, often cultivating their own networks with local, national, and international decision-makers-these dedicated policy engagement bodies are relatively recent and are under-accounted for in the literature on Higher Education (HE) reform and the role of evidence in policymaking.This article responds to two research questions.How and why have these bodies emerged in different universities, and which activities and strategies do they employ to support policy engagement?In doing so, it makes three contributions.First, it provides a comprehensive catalogue of UK HE policy engagement bodies.Second, it categorizes these bodies to create a typology which can be applied and refined to explain variation in approach to policy impact.Third, it addresses a gap in our understanding about the rationale and motivation for these bodies, as well as their strategies and activities, that provides insight on what is meant by policy engagement in practice in the different 'types' of policy engagement bodies.
To answer these questions, we draw on literature examining the changing role of universities and of knowledge, as well as those discussing and critically analysing the concept of evidence-based policymaking (EBPM).A review of both highlights the relative absence of research on these policy engagement bodies, although the themes and perspectives they develop are useful to frame our contributions.Specifically, their value lies in the alternative explanations they offer for the emergence and activities of these bodies; either as a mechanism for enabling evidence uptake in policymaking processes, or as an example of the commodification of knowledge.Our analysis discusses the emergence of university policy engagement bodies, and constructs four types, based on the core dimensions of activity that they work across.It finds that both literatures go some way to explaining the purpose and approach of these bodies.The degree to which interviewees' accounts reflect each of these literatures depends on their organization's objectives, strategies, resources, and opportunities; their occupational background; and the meaning they give to policy engagement.However, the accounts they give suggest that these are not two competing perspectives to explain these bodies but, rather, interviewees incorporated elements of both literatures in the stories they tell.Indeed, the emergence and activities of these bodies indicates a changing role for UK universities-shaped by opportunity, context, and what meaning policy engagement has for them in policy-research interaction, which neither literature fully accounts for.Our categorization helps map and account for the variation in bodies and practices that currently exists.We also begin to critically analyse how these bodies make sense of the research-policy relationship, and internal university practices and structures.
The article begins by reviewing the two explanatory bodies of literature before outlining our research methods.Next, we present our findings which are structured to respond to our research question on the emergence of university policy engagement bodies, and the core dimensions of activity that these bodies are focused on.Finally, we conclude and reflect on our findings, and what future research in this area could investigate.

The changing roles of universities in the knowledge-policy nexus
A review of literature on the knowledge-policy nexus, and the changing role of universities within it, highlights that university policy engagement bodies have rarely been the subject of academic publications.We draw on two major and related research fields-EBPM and critical discussions of the changing relationship between knowledge and policy-to begin making sense of this phenomenon.These literatures, while not dealing explicitly with these bodies as part of the HE policy engagement architecture, offer alternative explanations for their emergence, activities, and evolution-as either a response to the EBPM movement and the calls for academic knowledge to have greater social, economic, and environmental impact, or as an example of the commodification of knowledge, and as a challenge to academic freedom.
The first body of research relevant to explaining the emergence of policy engagement bodies within UK universities addresses the concept of EBPM.Proponents of EBPM mobilize a generalist and positivist understanding of the knowledge-policy relationship.It starts from the twin assumptions that policy-making proceeds on the basis of the best available evidence of 'what works', and that evidence can and does exist to inform value-free policy decisions (Sanderson 2006;Boaz et al. 2019).Rather than being interested in the contextualized politics of knowledge production, translation and use, this literature tends to perpetuate Caplan's two-communities' thesis to explain barriers to the application of evidence to policy.According to Caplan (1979), research and policy occupy two separate worlds with different values, norms, practices, deadlines, and communications practices.To bridge the divide between these worlds, advocates of EBPM have developed knowledge translation, knowledge exchange, and knowledge mobilization tools and 'how to' guides (Cairney and Oliver 2020), focusing on the processes and practices of knowledge transfer.
This understanding of the knowledge-policy relationship has led to a number of government initiatives since the 1990s aiming to 'close the gap' between these communities (Boswell 2016(Boswell , 2018)).In recent years, this scholarship has increasingly focused on the 'in-between' world of knowledge brokering.Knowledge brokers are seen as an effective approach to improving policy engagement with evidence.Research into these actors and their activities is focused on identifying the approaches and tools they employ-e.g.how they facilitate policy/research interactions, foster mutual understanding of policy relevant research, generate positive attitudes towards evidence for policy, and influence policymaking structures (Best and Holmes 2010;Breckon and Dodson 2016)-and the structural, behavioural, and relational enablers and barriers to success (MacKillop et al. 2019).
Taking an EBPM lens, university policy engagement bodies could be understood as types of knowledge brokers.They respond to imagined government demand for research evidence produced according to robust academic standards, and work at the boundary between research and policy to overcome barriers between the worlds.When seen in this way, we would expect the approaches and tools from knowledge brokering to be part of the everyday practices of these bodies.
Some empirical scholarship within the EBPM literature describes bodies set up by/within universities (predominantly outside the UK) to improve knowledge translation and transfer to policy.Because we anticipate that the Research Excellence Framework (REF) is a fundamental and very specific catalyst for and determinant of the form and function of these bodies, our concern is HE policy engagement bodies in the UK.However, it is important to note that this is not just a UK phenomenon (Phipps and Shapson 2009;Gaze and Stevens 2011), and these studies reflect that.Given their somewhat depoliticized and transactional EBPM orientation, this scholarship focuses on process and outcome evaluation of university policy engagement bodies, and reports quantitative metrics of impact, and barriers and enablers to knowledge transfer.For instance, Bakker et al. (2019) describe how the University of Minnesota (USA) set up a Policy and News Media Impact Service to demonstrate the research impact of its health research, which was limited to tracking citations of its research by different levels of government.Bennett et al. (2012) conducted a study examining the influence of 'independent health policy institutes' in six low-and-middle-income countries, two of which were university-owned.All the bodies provided policy advice, usually demand-led, in the form of policy reports and verbal briefings.The criteria for successful engagement with policy identified by these studies included a level of independence in governance and financing, and trust from policymakers.
These studies are of relevance for the present research because they provide some examples of how and why universities go about creating policy engagement bodies, and what they do.However, what we observed in our fieldwork goes beyond this existing scholarship, notably identifying the different types of bodies being created across the UK, and the varieties of strategies, and activities mobilized by them.In doing so, our study allows us to examine in a comparative way how different UK universities are approaching the question of policy engagement, and how that relates to concepts of evidence-based policy.
In contrast to EBPM, other studies have adopted a critical approach, often influenced by a constructivist ontology, which interprets the changing role of universities as the result of the neo-liberalization and commodification of education and knowledge (Olssen 2011;Ball 2012).This research argues that universities are increasingly required to demonstrate their productive and monetizable impact on society and the economy, as illustrated by the introduction of performance measurements, league tables, and competition between universities for funding and student quotas (Holmwood and Marcuello Servo ´s 2019).The UK REF is viewed as an example of a performance regime within which what counts as productive knowledge and research impact is shaped in particular ways, and in articulation with particular social, economic, and environmental interests (Smith et al. 2020;Thomas et al. 2020).
The commodification of knowledge is part of a wider neoliberal movement towards a global knowledge economy, and exemplified by the introduction of internationally performative metrics, such as global university rankings (O'Connell 2013; Upton, Vallance and Goddard 2014).For critical scholarship, universities and research are increasingly and problematically seen as economic rather than public investments by governments, and must thus demonstrate their utility, notably by producing and transferring 'useful' knowledge (Mowery and Sampat 2005;Gaze and Stevens 2011).Feldman and Sandoval (2018) write about the relationship between neoliberalism, metrics, and knowledge, drawing on Beer's (2016) 'metric power' concept.They examine how neoliberal agendas effect the structure of universities and their knowledge production practices, for instance challenging existing roles and introducing new professions (Whitchurch 2007).In this view, the emergence and multiplication of policy engagement bodies focusing on transferring knowledge and impacting policy could be seen as another demonstration of this metric culture in practice.Indeed, many universities and academics have been involved in knowledge transfer and building relationships with policymaking actors for decades.Yet, it is the rush towards impact measurement, or the demonstration of impact, that appears to have been one of the factors fuelling the rise of these new bodies.Policy engagement bodies could also be framed as resources targeted at improving rankings, ratings, and other performance indicators in the HE audit culture (Gill 2014).Within this culture, the academic becomes a 'technopreneur', contributing knowledge that must have market value (in the broadest sense) (Thornton 2013).Some types of policy engagement bodies may function as institutional homes for these technopreneurs, amplifying the visibility and gravitas of the policy-relevant knowledge they produce.
Within this critical strand, studies have argued that successive government policies have led to, or sought, a narrowing of what impact means, to the detriment of real impact (Pardoe 2014;Smith and Stewart 2017).In contrast, these researchers argue that multiple understandings and evaluations of impact may exist (Gunn and Mintrom 2017), and that demonstrating impact has become more important than actual impact (Wilkinson 2019).In a university sector now subject to performance-based research funding, the disparity in research power between UK universities is increasingly important (Papatsiba and Cohen 2020;Thomas et al. 2020), with policy engagement bodies playing a key role in this contest by growing the impact and brand of their respective university.In addition, critical studies into the changing nature and use of knowledge being produced by universities and mobilized in policy-making are also of interest.Bandola-Gill (2019), for instance, identifies different conceptualizations of knowledge as either socially-impactful-i.e.relevant-versus academically-impactful.A potential explanation that we explore in this article may be that policy engagement bodies participate in this diversification of the types of knowledge produced by universities (Budtz Pedersen, Følsgaard Grønvad and Hvidtfeldt 2020).However, the closer universities become to policymakers, the more difficult it may become to deal with conflicts over the types of knowledge produced, and its mobilization in policy (Warin and Moore 2021).
These studies all help to reflect on the various factors and conditions influencing practices and reforms in universities.However, the literature seems to suffer from a dichotomous approach of either advocating for EBPM or critiquing the neoliberalization of universities and the commodification of knowledge.Both may play a part in explaining university policy engagement bodies and their activities, which appear missing from both camps of research-maybe a result of their recent appearance and emergent public-facing presence.This article aims to plug that empirical gap, and reflect on whether and how the existing literature can help us understand the phenomenon and experience of university policy engagement bodies.

Identifying, cataloguing, and categorizing UK university policy engagement bodies
Our research process was inductive, beginning by listing and cataloguing all publicly identifiable policy engagement bodies in UK universities, looking for similarities, differences, and patterns, and then categorizing these bodies according to a set of key features.We adopted a two-step approach to identifying and cataloguing bodies.First, we listed UK universities alphabetically and, using a common search engine, combined each university's name with the keywords 'policy', 'engagement', 'institute', 'research office', and 'research support'.We only used the first page of results as a rule.Second, we identified members of the Universities Policy Engagement Network (Universities Policy Engagement Network 2022) and cross referenced the UK Cabinet Office Open Innovation team's crowdsourced list of policy engagement bodies. 1 Duplicates were identified and removed.
Desk-based analysis of the identified bodies' websites led us to observe variations, based on key features, such as the policy areas covered, types of activities and outputs, stated impacts, and staffing.Initial categorization based on these features was undertaken in April 2020 and identified five types of policy engagement body (n ¼ 80).These were labelled as follows: research support office; policy impact support office; policy knowledge brokers; policy responsive research producers, and fully demand-led bodies.This categorization was updated prior to the interview phase of the research and the first type of body (research support office) was excluded.This was because the research support provided was not specific to policy engagement, but more broadly oriented to support any form of stakeholder engagement and research pathway to impact.Other exclusion criteria were traditional academic departments undertaking research and teaching on policy; traditional disciplinary research centres with a dedicated knowledge exchange function; universities that did not provide online information on research/policy engagement support; and consortia of universities.The categorization was checked again in December 2021 to ensure bodies were correctly classified according to updated types (see below explanation of how types were enhanced following interviews).The final table (see Supplementary Appendix S1) includes 46 university policy engagement bodies, representing four types.
We recognize that our catalogue may not represent a complete list of all university policy engagement bodies, nor a complete representation of their features and activities.Our search, while systematic, may have overlooked some bodies.For instance, not all bodies are easily identified by keyword searching.Furthermore, as the development and refinement of university policy engagement functions is highly dynamic, our list may miss newer bodies (or bodies with a newer online presence), or the evolution or extinction of existing bodies.We also recognize that our categorization is based on data and information made public, and available online by these bodies.Certain information that would be useful in categorization is often not reported-e.g.start date (which would provide an indication of how responsive they are to the experience of REF-2014); size and background of staff; evolution of form, function, and funding.What is reported may also amount to aspirational claims, rather than actual practice.Nonetheless, we believe this catalogue (see Supplementary Appendix S1) represents the most comprehensive assessment of university policy engagement bodies todate.Furthermore, it is the first to differentiate between types of bodies; categorize them on the basis of activities, outputs, impact, and staff; and locating actually-existing bodies (at the time of research and writing) within or between types.
To explore the variance between bodies further, we conducted 15 semi-structured interviews with senior staff (including directors and deputy directors of policy engagement institutes and centres) from 13 of the identified policy engagement bodies.The interviews explored themes such as institutional origins (emergence and history), cultures of policy engagement (how they work with internal academic and professional services staff, and with external policy audiences and stakeholders), evolution of activities and impact, and included broader questions relating to their experience of the evidence-policy relationship, and the changing nature of universities and policymaking.
We selected interviewees to ensure we adequately represented each of the types of bodies identified by our desk-based categorization.All interviews were digitally recorded, transcribed, anonymized, and coded manually.Initial coding was undertaken separately by both authors and discussed during meetings to agree codes and ensure consistency.Codes were modified and refined as more data were collected to identify the background, drivers, stories, and motivations of these bodies and their members.Codes were also designed to capture roles, activities and mechanisms employed to seek impact on policy, as well as perspectives on wider phenomena, such as HE reform, the changing nature of knowledge, and the professionalization of policymaking.

Analytical approach
Our desk-based categorization of four types of university policy engagement bodies was enhanced by an analytical approach to making sense of our interview data, informed by interpretive methods that seek to go 'beyond text and talk', to examine the underlying meanings behind what people and organizations do and say (Yanow 2007).Although we focus solely on the voices of the university policy engagement bodies in this article, those voices provided significant insight into the complex muddling together of multiple discourses of origin and practice linked to both EBPM, and the commodification of knowledge.
Our research focuses on understanding what these organizations and their staff do and how they make sense of their own identities, roles, challenges, and successes.This is referred to as the practice of meaning-making, where actors aim to reconcile contradictions, in this case associated with working towards REF, building links with policy-makers to improve policy, and changing academic practice (Hajer and Wagenaar 2003; see also Van Leeuwen 2008).Our literature review highlighted that most research on the reason and practice of university policy engagement tends to be normative, either advocating for EBPM or critiquing the commodification of knowledge.In contrast, our analysis was interested in teasing out the multiple and often contrasting stories told by respondents about why these bodies exist, the activities they undertake, and with what effect (Yanow 2007;Wagenaar 2011).These stories do not just echo an abstract reality-e.g.playing an instrumental role in EBPM or performing the neoliberalization of the university-but, by trying to make sense of multiple and complex realities, and responding to their environment, mould understanding and perception of it (Fischer and Forester 1993;Van Leeuwen 2008).We used themes highlighted by the literature review above to inform our interview questions.In turn, we analysed and coded the interviews, according to the same themes.Rather than fact-checking what interviewees had told us, this process helped understand the different narratives and practices that can be created around a same event.For instance, we examined how the introduction of the REF and its different rounds were being interpreted-e.g. as helping to focus on impact versus a hindrance on important research-and how this led to different practices and strategies in different universities.

University policy engagement bodies: Their emergence and types
From the desk-based categorization and interviews, we established the origins and histories of these policy engagement bodies, and identified and refined key 'dimensions' of activity that they work across.These dimensions of activity include the role that policy engagement bodies play in supporting REF (from drafting impact case studies, to supporting policy relevant research by making connections with local, regional, or national policymakers, to producing research that directly contributes to REF); the approach they take to internal and external capacity-building to improve the supply and demand for policy evidence; and the ways in which they define policy engagement and seek relationships and networks beyond the university.
Our analysis reveals that, generally, bodies focus on different dimensions of activity and conduct these activities in slightly different ways, placing emphasis and meaning on their purpose and potential for impact, depending on their internal structures, staffing, relationships, and opportunities.We develop an empirically-driven taxonomy that reflects the range of what these bodies do and how they cluster into four broad types, with some bodies sitting between two types.These types are: the policy impact support office, the knowledge brokers, the policy evidence producers, and the demandled relationship builders.Table 1 below outlines how many participants were interviewed by type of policy engagement body.

Origins and histories
Interviewees offered a rich historical overview of the drivers leading to the creation of their policy engagement body.Some have existed for over a decade (e.g.D3-B and D3-C), whereas others have only emerged in the last 2 years (e.g.D1-A).The REF has played a significant role in establishing policy engagement with research as a priority for the sector.Several respondents referred to an institutional reckoning-post REF-2014-that led to a renewed focus and investment in engaging with policy as an audience for research, and a site for potential high-quality (four-star rated) impact.
Probably not unlike other centres, it was a response to the REF.I see impact as something that should be pursued for impact's sake rather than as an assessment exercise.(D1-A) Other drivers and motivations include: meeting the needs of individual academics in engaging with policy; discharging the civic responsibility of the University to ensure research realizes its social and economic value; to enable researchers to do more engaged, responsive research; to meet the external demands of increasingly professionalized policymakers, or for the university to incubate new external-facing initiatives.Many of the more established policy engagement institutes and centres (e.g.D3-A, D4-C, D3-C) describe themselves as having effectively resisted getting overly involved in the technical aspects of preparing REF submissions (especially writing impact case studies) and distinguish themselves from other university bodies and services that they perceive as having this more limited administrative function.D3-C for instance stressed that their institute was not 'sorting out impact for academics [. ..] that is someone else's role', adding that they were not 'an impact service but an impact partnership', and clearly cleaving out a distinct policy engagement profession.
The location of REF alongside other factors motivating the establishment of these bodies reflects their attempts to position policy engagement within the everyday academic mission of the University and its researchers, and to dismiss perceptions that policy engagement activity can ebb and flow with REF cycles.Indeed, the interviews raised a 'tension' between responding to the REF agenda and doing 'good' policy engagement.
I think to really make the most of what the REF framework offers, there's a need to make this an ongoing part of how the university approaches its work rather than every seven years, it becoming the agenda and then disappearing off again.I think the structure of REF can, sometimes, create that incentive.(D2-B) You need to be able to create environments in which [policy engagement] can happen.That's not an easy or quick thing to do when you are driven by the need for [. ..] positive REF results [. . .].Sometimes those activities aren't as obvious in terms of the reward because they're, kind of, slow-burning things that take a long time to come to fruition.(D1-C) Policy engagement with research is described as something that has, to some extent, always been done (e.g.something that individual academics or research centres have done) and has often had some level of institutional support (e.g.part of the function of research services to support pathways to impact).However, the policy engagement bodies we identified appear to have a level of central strategic backing and institutional embeddedness beyond that, with many of their websites and publications carrying messages of endorsement from the vicechancellor and other members of the University executive team.When we asked interview participants about the of their respective bodies, almost all referred to the role of the vice-chancellor or senior management in its inception.For instance, one policy institute deputy director referred to the 'recognition at senior level on the need for this external facing role' when the institute was created (D3-C).Equally, another recognized the role of the new vice-chancellor in establishing their institute as fitting with their commitment to the civic role of the University (D2-A).
Interviews describe a process of formalizing their function in recent years, as a way of establishing what they offer and creating a clear brand identity.
What we're now describing is the [organisation name], which I guess is a bit of a sort of brand that we've put across a policy engagement function that existed within the university for the last four or five years now.(D1-C) For the oldest bodies, their name was described as 'an established brand' (D3-B).While brand identity is seen as important externally, considerable importance is also placed on having a formal function to raise awareness of policy engagement, and the support that is available internally.For example, respondents talked about creating a 'name' (e.g.D1-C, D2-D, D2-C) and describing an 'offer' that makes it clear to colleagues within the University what they can expect and how they can get involved.
The Institute is an effort to really do three things . . .Externally, it's to showcase what we do and to facilitate those links between our academics and various public audiences and policymakers.Internally, it's about, firstly, making sure that we have more of an institution-wide handle on what we're doing.Secondly, it's about capacity-building.(D2-B)

Types of policy engagement bodies and how policy engagement is done
We identify four broad types of policy engagement bodies across the sector: • Type 1: The policy impact support office • Type 2: The knowledge brokers • Type 3: The policy evidence producers • Type 4: The demand-led relationship builders.Our categorization of policy engagement bodies based on deskresearch (Supplementary Appendix S1) finds there are 17 bodies in category Type 1, 11 in Type 2, 4 in Type 3, and 3 in Type 4; 11 bodies fall between two categories; 7 are Type 1-2, 3 are Type 2-3 and 1 is Type 3-4.Based on our interviews with a sub-set of bodies broadly representing each type, we find that bodies within the types work across common dimensions of activity, have similar staffing profiles, and employ similar policy engagement strategies and tools (see Table 2).However, there is a degree of flexibility within and porosity between types.For example, a body may, at different times, mobilize strategies and tools from different dimensions of activity based on the opportunities it has available within a particular policy engagement context.Figure 1 below illustrates this flexibility and porosity, with some bodies evolving from one type to the next, and some universities hosting several bodies simultaneously.

Type 1: The policy impact support office
The first type describes a group of bodies (n ¼ 24 Type 1 or Type 1-2 in our desk-based categorization of all UK HE policy engagement bodies) that are almost entirely internally focused.This is the most commonly occurring type of body.They are linked to (e.g. they have evolved or have been spun out from) more traditional research support offices.They tend to employ professional services staff on research support career pathways and provide dedicated policy impact support.Their origins are most directly and explicitly in response to the REF, which is an ever-present focus in their everyday activities.They describe spending large amounts of the time 'digging around for evidence' to maximize the quality of potential REF impact case studies (D1-A).
However, respondents also describe efforts to create balance between the demands of preparing REF submissions and having time and resource to dedicate to upstream activities, such as helping academics identify and engage policy audiences earlier in research activity (preferably at the point of drafting funding bids), and to plan points of research interaction with policymaking.They particularly describe the importance of supporting PhD students and early career researchers to develop as policy impact leaders of the future.
For me, what I find most rewarding personally is where you can invest over the longer-term with somebody who is perhaps completely willing and enthusiastic but doesn't quite know where to take the research or where to go with it.(D1-C) Across this type, there is a prominent focus on raising awareness of policy engagement opportunities, e.g. government consultations and fellowship.They provide regular mailouts signposting these opportunities and providing training and development to build capacity among research staff.This can involve organizing events and workshops with external speakers or working one-toone with academics.
Typically, respondents describe a move towards providing more responsive and bespoke packages of one-to-one coaching and support for academics who are willing to engage policy with their research.They describe efforts to shift usual academic communication and publication priorities.Their role in 'handholding' (D1-B; D1-A), to enable academics navigate unfamiliar policy processes outside of their comfort zone, was a common theme.However, they recognized that even where there is enthusiasm among research staff, this process of adjusting traditional academic practice to target different audiences for research takes time (D1-C).
For most respondents in this type, there was a notable absence of capacity within their role and team to proactively interpret policy agendas.Their process of policy engagement starts with understanding the research and supporting researchers to refine processes for disseminating that research.They do not tend to have, or seek to develop, direct relationships with policymakers, nor do they see that as part of the offer they provide to academics.Some describe using political monitoring tools (such as DeHavilland or Dods Information) to help broadly target research communications, but in general their role is to help plan policy engagement strategies, and map potential policy contacts for academic staff to pursue.
they have to make the contact, it's not something I can do for them.They have to develop their personal relationships because it's their area of expertise.A lot of the time, even if I do find somebody, 'This person seems to be the right person that you need to contact.' It's like, 'You need to contact them, I'm not doing it for you'.(D1-B).
However, there were some exceptions that highlighted an evolution from policy impact support towards more of a knowledge broker role.In these cases, respondents spoke about a shift in their thinking about how policy engagement is done, and a reorientation of their role away internal support and towards external intelligence gathering, regarding the evidence demands of policymakers.They spoke about the need to invest more of their time and effort in developing contacts and relationships with policymakers nationally, regionally, and locally; and determining areas of overlap between their evidence needs and the research strengths of the University (D1-C).

Type 2: The knowledge brokers mobilizers
The second type describes a set of approaches to policy engagement that places considerable emphasis on developing relationships, internally with researchers and externally with policymakers, and brokering relevant research evidence to meet specific policy demand.These bodies' strategies include identifying and prioritizing points of intersection for university research with policy (locally, nationally, and/or internationally), intensive, bespoke support for policy engagement, and building capacity among key research teams through intensive, collaborative research-policy engagement ventures and mentoring.We identified 14 Type 2 or Type 2-3 bodies in our deskbased categorization of all UK HE policy engagement bodies.These bodies tend to employ former senior civil servants or policy advisors, as well as people with policy communications or media backgrounds.Like Type 1 bodies, their origins are REF responsive.However, they are more likely to have developed a narrative about policy engagement as a good thing in its own right, because it supports individual and organizational motivations of (good) academic researchers and civic universities to affect positive social change.They present the process of influencing policy with research as taking time and effort and prioritize spending time nurturing existing networks or developing new relationships with policymakers (politicians and officials) and advisers.They use these relationships to determine demand for research and opportunities to input on the policy process (D2-A).They then actively scan the University for those aspects of current research activity that speak to live and pertinent policy (and sometimes public) debates and identify outputs that can speak to policy agendas and will interest policy audiences.Unlike the policy impact support office, they do not see themselves as providing a general service for researchers to help push research out to policy, especially when they lack a clear sense of where and how that research will land and its application to current policy priorities (D2-D).As such, these bodies are much more selective in 'picking winners'.
So, we're not an open opportunity for all academics who are interested in doing some research to talk to us about their pathways to impact and fancify their bids.But where it fits, where there's relevance, we come in.(D2-E) You want to be nice to people and get them on side and make them your allies, but sometimes, you just have to say, 'Do you know what?That's not going to interest anyone'.So, there's that constant challenge of bad news. . .

. (D2-A)
Where they see opportunities for research to pique policy interest, active support is provided to package that research for policy.They place considerable focus on capacity-building for effective policy engagement.However, this is less about generic training and more about one-to-one, tailored support to raise the profile of that research among highly specific policy audiences.They use their networks to get their researchers in front of the right policymakers and ensure they are seen as credible experts on the subject.They have a corporate objective driving their endeavours.For the specific areas of research that they support, they talk about working to 'make ourselves one of the regular faces' (D2-D) and 'place us as a university in that conversation' (D2-C).
Respondents describe identifying mechanisms for getting research to policy.These may involve, for example, supporting applications for government secondment opportunities or writing good evidence for select committee calls.Informally, they may use their relationships to reinforce and champion these select committee evidence submissions, arrange and coach academics for media opportunities, and host high profile events with prestigious external speakers and key decision-makers-again, recruited from their policy networks.
We've done a bit of arm-twisting with committee clerks to be able to say, 'You've got a written submission from Professor X or Y.We think that they're really good because of this, this and this.
Have you decided about who you're going to have for oral evidence?This person would be really good'.(D2-D) We've provided a platform to create events on campus where we've been able to use [. ..] colleagues' contacts with policymakers to come in and have that conversation with us.So, we brought in MPs from all parties, councillors from all parties . ...

(D2-C)
For the knowledge brokers, policy engagement is about having the networks that allow them to know demand, selectively support supply, and broker between the two.As one respondent described, they seek to 'provide the oil on the wheels just to make those introductions' (D2-C).

Type 3: The policy evidence producers
We identified five Type 3 or Type 3-4 bodies in our desk-based categorization of all UK HE policy engagement bodies.These bodies are well established and institutionally embedded, with visible University executive endorsement.They perceive themselves to have developed an identifiable internal and external brand identity, and as critical to the REF success of the University without being a response to it.They tend to employ a combination of former senior civil servants/policy advisors and senior academics with a track record of policy impact with research.Like the knowledge brokers, they stress the value of their policy networks but place additional emphasis on their ability to understand the nuanced interplay between research evidence demand and supply within an evidence-policy system.Specifically, they recognize that efforts to mitigate the disjoint between demand and supply of research evidence for policy is subject to the complex vagaries of policymaking processes.Respondents acknowledge the multiple considerations that inform policy decisions-including political will and governmental competency-and the role of research and evidence as, at best, one input among many (D2-A; D2-C; D2-D; as well as D3-A; D3-B).
Policy evidence producers describe a need to be agile to take advantage of multiple means and opportunities for communication and influence, and tend to work across a spectrum of activities.They undertake broader and less targeted research disseminationseeking to raise general awareness of high-quality research of public and policy interest-while also responding to current policy need and engaging in proactive 'thought leadership' on emerging issues (D3-B).They often describe high profile blogging on issues of the day and on the horizon, as part of their early and ongoing strategies to establish themselves as relevant to policymakers.
There are lots of ways in which you can influence public policy.And some of that is just a climate of ideas, it's feeding things into public discussion by writing newspaper articles and blogs and talking to politicians.(D3-A) While retaining many of the functions that support their academic base to engage better with policy, they also place considerable focus on the need to produce research for policy that directly addresses their evidence demand, rather than looking for loose overlaps research and policy interest.Many bodies that closely fit this type tend to have an independent research programme (D3-A; D3-C) or provide added value by bringing different research agendas across the university together, to generate new insights relevant to policymakers (D3-B).They describe their research programme as critical to how their reputation is established-internally as well as externally.Externally, they describe a need to be seen as credible expert knowledge producers, not just brokers.Internally, they are pulled towards the core academic functions of winning grants, publishing high-quality papers, and contributing REF impact case studies from their own research activities (D3-A; D3-C).Navigating this back-and-forth between university expectations and the policy world is illustrated by a deputy director of one of these bodies who explained how to adapt the characteristics of traditional research to influence and impact on policy: It's about involving stakeholders early in your work to create the environment for transfer, involve them to shape your policy research question, but also about how to communicate research, for example by turning the academic paper on its head.(D3-C) Bodies in this type use intelligence from their policy networks to work upstream of policy agendas, prioritizing research that is directly related to current and future policy priorities, is timely, highly visible, and has a compelling narrative.In this way, they operate like academic versions of think tanks (D3-A; D3-C; D2-C).
Some of it though is just thinking and acting like think tanks and others do.Get good press coverage for your work; make sure you've got good networks with politicians; feed your things into decision making processes at the right time; create the narratives for them, so that they can fit.(D3-A) While stressing the attributes of think tanks, these bodies are also keen to describe distinctions in their approach and practice.They talk about the importance of maintaining academic rigour and independence from government(s).To this end, they recognize the value of developing partnerships within the wider research-policy ecosystem, and the importance of working with different bodies that play different roles.In particular, one respondent describes the strategic value of developing networks with other evidence producers and brokers at different degrees of distance from policymaking to amplify their activities and gain greater credibility, visibility, and influence for their research (D3-A).
As well as producing evidence for policy, bodies in this type describe structural approaches to maintaining policy relationships and closing the gap between research and policy thinking on contemporary and emerging issues (D3-B).In some cases, they offer 'masterclasses' and formal policy fellowship opportunities for policy-makers, as well as postgraduate level policy profession training and development opportunities, including Masters programmes and professional doctorates (D3-A; D3-C).They describe these functions as part of an approach to deepen networks with policymakers, and develop mechanisms to formalize and embed feedback loops between research evidence and policy decision-making (D3-A).For example, they describe circumstances in which policymakers in their networks set live policy questions for Public Policy Masters students to respond to, and offer placements in government departments (D3-A).In this way, these bodies undertake policy engagement by providing a site for the exchange of research and policy questions and ideas between research-oriented policymakers and policy-oriented researchers.

Type 4:
The demand-led relationship builders This is the least commonly occurring type of UK HE policy engagement body.Our desk-based categorization identified three Type 4 bodies.What sets this final type of bodies apart is that they have developed very close relationships with policymakers (national or local governments) that facilitate relatively open dialogue between them about current policy challenges, and the contribution that research evidence and expertise can make.They have developed credibility, trust, and processes that allow them to talk regularly to Ministers, local politicians, or officials (D4-B; D4-C; D4-D).
We have a very mature relationship with [policy body] and so can have quite involved conversations with them about the decisions that they face.(D4-B) So, we have three key research themes [. ..].These have been chosen very much because they play to the interests within [policy body] and the policy community.(D4-D) They operate a range of mechanisms to organize these relationships and maintain dialogue.Some are predominantly based on intensive programmes of meetings and events, others have internal research capacity and undertake more traditional commissions, albeit with elements of co-producing the policy-research questions, and there are often combinations of both.Respondents describe a process of evolving and adapting the offer-over time and/or in response to different policy concerns (D4-A; D4-D)-and the sets of mechanisms for enabling closer working (D4-C).Unlike the other types, these demand-led bodies are often partly funded by government, be it via a core grant (D4-A) or individual projects that are commissioned by policymakers (D4-C and D4-D).
They tend to employ a mix of senior academics, former senior civil servants/policy advisors, former think tank staff, and researchers with both academic and consultancy backgrounds.However, similar to the policy evidence producers, these bodies describe themselves as 'different from think tanks and [. ..] different from academic centres' (D4-A).They describe themselves as having developed a hybrid of policy and academic research practice (D4-C).Despite their close relationships with policymakers and direct funding streams, respondents place considerable value on the perception of their independence, rigour, and commitment to the integrity of the evidence they produce or broker (D4-A; D4-D).For the director of one of these bodies, the academic background of the staff, combined with a prestigious advisory board comprising senior policymakers and a peer-review system for their outputs, brought credibility when dealing with government (D4-A; also, D4-D).
The motivation for seeking and establishing this particular type of close relationship with policy is described as coming from a desire to influence 'agenda setting or upstream' of policy decisions (D4-C), and provide good evidence for policy decisions (D4-A; D4-B; D4-D).Notably, these bodies make very little mention of the REF in their origin story.
We work with [policymakers] to help them to identify where they and we believe that there is evidence available which would be useful to them in helping to understand the nature of a policy challenge that they're seeking to address, and to identify other challenges which they haven't thought of that they should be addressing.D4-A) Some of these institutes, because of their establishment in the policy community, feel able to choose the projects that they undertake and have a more open conversation with policymakers about problems and solutions (D4-B).They reject projects they feel they are being steered away from their role of conducting research or interpreting evidence, and driven towards making judgements between sets of political values (D4-A).While they often conduct their own research, they also seek to engage with experts beyond their team and beyond their university, as they deem relevant to addressing the policy questions (D4-A; D4-B; D4-C).They are highly selective about who they work with and introduce to policy audiences-selecting researchers from the pool of experts that will interact well with policy officials.They invest considerable time and resource in getting the match between academics and policy challenges right (D4-C; D4-D).Being able to interact well is defined by these bodies as having a set of personal attributes, such as being able to communicate clearly and succinctly, speak about a broad body of evidence where relevant and being reliable and punctual-as well wanting to inform and support policy decision-making, and not being overly critical of policymakers and policy decisions (D4-A; D4-C).Respondents were concerned to stress that, for them, these attributes do not compromise rigour or commitment to the evidence.Managing critique was described as 'being politically aware enough to know the right ways to deploy that evidence and the wrong ways to deploy that evidence' (D4-A).
These bodies emphasize how knowledge produced by universities takes increasingly multiple forms, in line with the specific policy questions, including experiential knowledge, or knowledge made in the exchange of ideas when policymakers and academics meet around a table to discuss an issue from their various perspectives (D4-A; D4-D).Again, they invest considerable time and resource in bringing hand-selected groups of people together (often behind closed doors) in formal and informal ways.They describe organizing meetings and hosting roundtables and dinners (D4-A; D4-C; D4-D), as well as operating as a locus (the centre of loose and fluid networks) for conversations between researchers, experts, evidence brokers, policymakers, and public services (D4-A).
Our founding director, had a great belief in the value of networks but also, sort of, slightly serendipitous meetings or, sort of, enabling serendipitous meetings.(D4-C).
Respondents describe their ability to enact informal contact and interaction between trusted academics and policymakers as providing a safe environment for policy problems and evidence solutions to be discussed in depth.

Discussion and conclusions
Our research demonstrates how and why policy engagement bodies have emerged in different universities, what form they take, and which strategies they mobilize to support policy engagement.We provide a comprehensive catalogue of UK university policy engagement bodies and categorize them within or between types, according to differences in key features (such as policy area, activities, outputs, stated impact, and staff).From this, we develop a typology to explain divergence, which we enrich through in-depth interviews to establish the core dimensions of activity common to each type.
Our findings show that the REF has been a common and significant driver in the formation and evolution of most university policy engagement bodies in the UK (Middlehurst 2014).At times, respondents-particularly from internally facing, Types 1 and 2 bodies-harness REF narratives and incentives to validate their existence.Simultaneously, to associate their activities and the purpose of policy engagement with something 'more than' REF, they also distance the work of their teams from the narrow and instrumental confines of REF submission cycles and align it with the academic and civic mission of the university.They articulate their role as creating the internal conditions, external visibility, and access points for policy engagement and knowledge transfer to happen.Yet, practice across these bodies is varied, as illustrated by the four types that we identify in this article.We also observe different narratives being combined and mobilized by staff to make sense of their activities in relation to policy engagement-thereby defining what engagement means and how it is done, and carve out their specific institutional role.
Our analysis shows that these bodies work across different core dimensions of activity (see Table 2).Differences emerge in the strategies and tools employed to work internally and externally, what form knowledge for policy takes, and the relationships and networks they build.As illustrated by Figure 1, bodies that look more like policy impact support offices or knowledge brokers tend to focus internally within the university and increase researchers' capacity to engage with policy.Albeit with different emphasis on and investment in the ability to target and mediate policy engagement opportunities.Bodies that look more like policy evidence producers and demand-led relationship builders tend to focus externally, developing close relationships with policymakers, and adopting upstream and responsive research processes and outputs.They employ staff with a broader mix of backgrounds, and are selective about who they work with and how.
These differences between types of policy engagement bodies reflect different interpretations of what constitutes policy engagement and how it is practiced, as well as different opportunity structures, resources, and expertise.We do not present these differences to suggest that some are better or worse than others in terms of either their strategies and tools, or the impact of their activity.Rather, we seek to highlight the extent to which these different meanings of policy engagement are a product of, and produce or constrain, structures and functions that seek, in different ways, to change either/both academic and policy practice, in the production and utilization of knowledge.In doing so, we emphasize the benefits of critical interpretative methods to unpick the coexisting, multiple and potentially contrasting, discourses of policy engagement bodies.Significantly, the link between the two research questions of origins and activities did not emerge as a strong or central contribution.Neither deskbased research nor interviewee accounts suggested a clear link between origin and activity.On this basis, we are concerned not to suggest that bodies evolve in a linear way from 1 to 4, as we have little evidence for this, and indeed we have some evidence to the contrary.Similarly, we do not suggest that a Type 1 body formed as a 1 would necessarily follow a Type 1 body trajectory.A number of factors come into play in determining the origins, as well as activities of these bodies, with the context and different circumstances (Vice-Chancellor support, staff background, local demand, pre-existing links with policymakers at different levels) often playing a role.frame this study by drawing on literatures examining the changing role of universities and the commodification of knowledge, and those discussing and critically analysing the concept of EBPM.For us, their value lay in the alternative explanations they offer for the emergence and activities of these bodies.These go some way to explaining these bodies-what they do, how, and why-depending on the organization's objectives and activities.However, we observed that bodies mobilize the explanations offered by these literatures together, rather than as competing accounts, and neither literature fully explains the changing role HE plays in policy-research interaction, which is worthy of further research and international comparison.Specifically, we find that respondents often combine narratives drawn from polarized positions on EBPM and the commodification of knowledge to make sense of their body's existence, how they work, and their personal experiences, thus emphasizing the complex dynamics at play in the HE and evidence-policy contexts.Thus, rather than being a story of (1) EBPM-informed knowledge brokering, or (2) the neoliberal commodification universities and their knowledge, the interviews explored the importance of everyday practice and practical judgement to navigate policy-research interaction, and respond to their environment and the various demands on their service.
The bodies that we studied align themselves with EBPM narratives on the importance of evidence for policy and concern themselves with the task of brokering between the worlds (Caplan 1979).Our findings illustrate these bodies repackaging traditional research into more cessible forms of briefings, evidence review, rapid research, and research-policy dialogue relevant to current and emerging policy problems, as well as investing in improving communications through capacity-building initiatives, secondments in both directions, and the active curation of networks with policy.These examples demonstrate the brokering skills and tools being developed by these different bodies, adding substance to EBPM studies of knowledge brokering, by illustrating how this is attempted in a specific context.However, in contrast to that literature, we underline how the brokering strategies available to these bodies are dependent on perspectives of policy engagement, circumstances, and opportunity, as well as the capacity, skills, and expertise of the bodies themselves, thus highlighting the importance of ideas, resources, politics, and serendipity in policy engagement (Cairney and Oliver 2020).
Equally, we find that in practice these bodies do not straightforwardly take part in the commodification of knowledge.Closer relationships between research and policy indeed have pitfalls (e.g.threatening independent and critical thinking) but also benefits, such as influencing and informing policy, and therefore wider society, economy, and environment, with potentially relevant and useful research.Policy engagement bodies present nuanced understandings of the multifaceted and opaque nature of policy impact.Supporting knowledge producers realize their aspirations for impact speaks to broader interpretations of what constitutes impact, albeit with the constant backdrop of the REF (Pardoe 2014;Smith and Stewart 2017).The degree to which policy engagement involves working upstream of policy agendas, curating networks to influence policy demand for evidence, and develop close critical thinking feedback loops and knowledge co-production processes, varies by type, demonstrating the extent to which perspectives and discourses of policy engagement relate to and play out through the structures, resources, and opportunities of these bodies to produce activities that seek-in different ways-to influence the practices of both academics and policymakers.
The policy engagement bodies we document are taking part in renegotiating what role universities could or should play in policy deliberation and decision-making, and society and social outcomes more broadly, as well as the type of knowledge they produce.To the extent that engaging with these bodies confers potential impact opportunities-and therefore status and reward-on academic staff, these bodies are reorienting, or seeking to reorient, academic practice in different ways across the different types; promoting the value of policy relevant communication; changing timelines and research questions; and influencing the deployment of criticality.At times, our interviewees defended the academy-its rigour, freedom, and independence for instance-and at others, they attempted to adjust it and move towards greater usefulness for policy.Furthermore, beyond producing (or repackaging) different types of knowledge to that traditionally associated with universities, these bodies engage in the production of 'interactive' or 'experiential' knowledge produced when academics and policymakers meet and discuss the framing of, and potential solutions to policy problems.
While, these bodies speak to the EBPM and critical HE reform literatures, they also speak to wider themes of the role of knowledge in society and policy, and the relationship between academia and policy.These bodies are involved in a number of 'agendas': e.g. research quality assessment (REF), building links with society and policy at different levels, growing brand recognition, and the reform of universities.These bodies also have a purpose internally, often symbolizing an attempt to elevate impact and policy engagement.A simplistic or unidimensional understanding of these bodies as simply REF-focused and playing into the performative metrics culture falls short of explaining the multiple stories of emergence, activity, and the pursuit of impact being mobilized by these bodies, in attempting to establish themselves as 'more than REF'.In doing so, our empirical analysis has highlighted, via the categorization and dimensions that we put forward, how the story of policy engagement by universities is not a simple and unified one.Instead, each body strives to develop different activities, strategies, and structures, juggling requirements such as REF, with other agendas and ideals about policy engagement.
This study has a number of limitations.We are aware that we are only developing our findings based on the voices of those working within those institutes, which provides a rather one-sided, view.There would be value in gathering data from stakeholders outside of these organizations, within and beyond the universities.Future research enquiry could focus on individuals and organizations targeted by these institutes, such as in central and local government, as well as other evidence producers, evidence intermediaries and academic experts in the field of evidence and policy-to discuss the purpose of these bodies, how they function, and what impact they are having on policy, society, and knowledge.The value of this next step would be in recognizing that policy engagement is increasingly a multifaceted, non-linear, complex process that involves multiple partners, e.g.beyond researchers and policy makers, and including public, private, and third sector evidence producers, users, advocates, and commentators.Furthermore, there is a need to better understand and compare how other countries are experiencing the changing relationship between academia and policy.Based on what is being published, the UK seems to be in the vanguard for these types of bodies--however, other countries have adopted different processes in achieving research impact, and contributing to policy and practice-e.g.Canada and its vast knowledge mobilization effort between universities and society (e.g.Phipps and Shapson 2009).
In conclusion, there is a need a more comprehensive and critical examination of universities' knowledge transfer initiatives, and this article focuses on a fairly recent one: the emergence of dedicated policy engagement bodies.Our research focused on the emergence and activities of these bodies in the UK, how they are organized, and the strategies and practices they adopt.In addressing these questions, we develop a comprehensive typology of bodies and determine the key dimensions of activity that they work across to support the transfer of knowledge into policy.This study provides wider reflections on the changing functions of universities, as well as the type of knowledge and relationships that are being fostered and/ or expected of universities in the 21st Century.In doing so, this study contributes to research into HE, knowledge brokering, and the evidence-policy relationship.The first step of understanding divergence in the form and functions of university policy engagement has been the object of this article, and is a necessary step to begin critical discussion on how these organizations are impacting policy and academic practice.Further work to validate the typology and examine links between organizational origin and dimensions of activity would be useful.In addition, starting with the categorization of bodies in the present article, it should be possible to further understand whether certain approaches or activities are more 'effective' in influencing policy, and what the implications for policy might be of inequalities in effective policy influence.It should also be useful as a starting point to understand whether these bodies are changing how universities are organized, how research is developed and communicated, and how academics work and are rewarded.All these themes are present in the literatures that we have discussed but have not been examined from the premise of formalized university policy engagement support, which should constitute a worthy endeavour for future research.
So, the 2014 REF saw that impact was important.It's going to be even more important in 2021.(D2-A) I think a massive driver is the REF.[. ..].If you can do it [policy engagement] properly, it can be really prestigious.It is really worth investing in a policy engagement post or a team because if it's done properly, you can be really influential.(D1-A) Nonetheless, while all talked about REF, they equally sought to place REF in the context of a range of drivers and motivations, and locate it downstream of what they do, how they do it, and why.I think the more interesting driver is that people are seeing it [policy engagement] as a good thing to do, and that, actually, the fact that the REF [. ..] is creating an incentive to be an engaged researcher [. ..].Actually, for a lot of people, that's what they came into research to do.So, it kind of speaks to a lot of universities talking about their civic purpose.It's a much more engaged approach.(D2-A)

Figure 1 .
Figure 1.Institutional blend of dimensions of policy engagement.
If it's a normative question, should I do this or should I do that, based purely on political values, that's not something which we have anything to speak to.(D4-A) [We're] very wary of being seen as political or being pushed by a politician to do something in a certain way.(D4-D)

Table 1 .
Interview schedule by type of policy engagement body

Table 2 .
Types and description of dimensions of policy engagement functions Type 4: The demand-led relationship builders Establish close relationships with policymakers to directly respond to demand Multiple mechanisms for ongoing interaction with policy Engagement with wider evidence community Former civil servants or policy advisors Former think tank/consultancy staff Senior academics Demand-led research programme responding to an articulated demand for evidence Open dialogue events, meetings, roundtables, etc. often behindclosed-doors.Mobilize evidence and knowledge in context.