Australian Social Work Research: An Empirical Study of Engagement and Impact

Abstract

Internationally, non-academic research impact is assessed by governments as part of evaluating the quality of publicly funded research. A case study method was used to investigate the non-academic impact of Australian social work research. Interviews were conducted with fifteen leading researchers about outputs (research products, such as publications and reports), engagement (interaction between researchers and end-users outside academia to transfer knowledge, methods or resources) and impact (social or economic contributions of research). Twelve case studies were prepared using a standardised template. Content analysis highlighted examples of impact, and theoretical and in vivo coding uncovered processes of engagement and impact. Different types of engagements with research end-users influenced impact in three areas: legislation and policy; practices and service delivery; and quality of life of community members. Engagement and impact were intertwined as research altered policy discourses and illuminated hidden social issues, preparing ground for subsequent, more direct impact. Likewise, academic and non-academic impacts were intertwined as research rigour and academic credibility were perceived to leverage influence. There was no evidence of achieving impact simply through the trickle-down effect of scholarly publication. The findings broaden understandings of how research influences policy and practice and iterative and indirect relationships between engagement and impact.

research assessment, research impact, research partnerships, research utilisation, translation

Social work research is concerned with improving practices, policies and services to reduce inequalities and improve the well-being of marginalized and diverse individuals, families and groups (e.g. Shaw, 2007). From this perspective, achieving the purpose of social work research requires research that makes a positive contribution to the community, beyond an input to formal knowledge. Internationally, research funders have begun assessing the non-academic impact of research alongside traditional markers of research quality like peer review and publication citations. The concept of research impact, or the wider value of research for society, has been on national and European Union policy agendas for several years (Donovan and Gulbrandsen, 2018). Unsurprisingly, a considerable literature has emerged exploring the concept of research impact and the complexities of its measurement within widely diffuse research fields and contexts. Concerns have been expressed about the potential effects of the government impetus to measure impact on scientific research agendas (Kenyon, 2014; Smith and Stewart, 2017), as well as suspicion of its bureaucratic origins (Alla et al., 2017; Williams and Grant, 2018). Poorly implemented impact assessment risks changing knowledge production behaviours, resulting in unintended (and unpredictable) consequences such as the inappropriate valuing of certain types of knowledge or research over others (Kenyon, 2014; Alla et al., 2017; Smith and Stewart, 2017). The attention to non-academic impact has been described as a fundamental shift from academic-initiated and curiosity-driven research, to knowledge production driven by a multitude of factors concentrated more distinctly on research applications (Dobrow et al., 2017). Another debate surrounds the financial impost and efficiency of the assessment process itself, and whether the process simply measures what is easiest, rather than what counts most (Banzi et al., 2011). However, as Brewer (2011) contends, perhaps the impact agenda looks more problematic than it really is, and it presents an opportunity for social work as an applied discipline to showcase its merits.

Regardless of the potential benefits and limitations, in Australia, engaging with research impact is non-negotiable for the profession and academic social workers. National assessment of engagement and impact occurs in parallel with the Excellence in Research for Australia (ERA) framework that assesses research excellence in universities. The introduction of ERA was preceded by lengthy political debate, with concerns expressed about costs, methodology and propensity to skew research agendas to government priorities (Gunn and Mintrom, 2018). The stated goals range from accountability for public expenditure on research to encouraging the translation of research into tangible public benefit (Australian Research Council, 2019). Although Australia differs from other international assessment models in that research impact measures are not yet directly used to determine public funding, there are reputational implications as universities tend to support high-performing disciplines (Tilbury et al., 2017a; Williams and Grant, 2018).

Although research funding schemes have long required applicants to consider potential public benefit (Kuruvilla et al., 2006), a wide gap continues to exist between research and its translation into practice and policy in health and social services (Cohen et al., 2015; Gray et al., 2015). A greater emphasis on achieving research impact may encourage academia to formulate new approaches to the way researchers engage, co-produce and partner with other researchers and disciplines, research end-users and the wider public. Reinforcing this trend is the growing attention to research translation, implementation science, evidence-based policy and practice and intermediary organisations aiming to connect research with practice. All these developments are concerned with strengthening the capacity of research to increase the effectiveness and efficiency of human services.

The Australian Research Council first assessed engagement and impact in 2018. Engaging with non-academic stakeholders to use knowledge in practice was recognised as integral to achieving impact. Engagement was defined as a process or pathway to knowledge mobilisation, the ‘interaction between researchers and research end-users outside of academia, for the mutually beneficial transfer of knowledge, technologies, methods or resources’ (Australian Research Council, 2017, p. 5). Impact was defined as the ‘contribution that research makes to the economy, society, environment or culture, beyond the contribution to academic research’ (Australian Research Council, 2017, p. 5). Definitions of this nature are based on a linear logic connecting the process [engagement] to the result [impact]. In contrast to Australia, the UK Research Excellence Framework accords less value to engagement. Whilst seen as a precursor to impact, engaging with non-academic audiences to disseminate research findings was excluded from assessment because, on its own, it may not lead to impact (Research England, 2019). It is important and timely to understand research engagement and impact, and the relationship between them. Insights about how research impact is generated will assist researchers to cultivate strategies for maximising the use of research. The aim of this study was to examine the influence of social work research on social work and human services policy and practice in Australia, and how this was achieved.

Study design and method

Design

A case study design was used to examine selected programmes of social work research in each of three fields (aged care, child protection and disability). These fields were selected based on their size and diversity. The research questions guiding the study were as follows:

1. what strategies were used by social work researchers to engage with policy-makers and practitioners?

2. what was the impact of the research on social work and human services policy and practice?

Participants

The sample was derived from scoping reviews that identified the four most productive Australian social work researchers in each of the three fields of practice during the past decade (Tilbury et al., 2017b; Bigby et al., 2018; Hughes et al., 2018). This approach to sampling ensured that the participants were engaged in a sustained programme of research. The three Australian co-authors met the criteria and were included in the sample. Researchers were contacted by email to ascertain their interest in the study, and they all agreed to participate in an interview. Three requested the inclusion of a colleague, who they worked closely with, making a total of fifteen researchers interviewed. All were senior researchers, and three were male, approximating the gender composition in social work.

Procedure

In-depth interviews were conducted with the researchers about outputs (direct products of research, such as publications and reports), engagement and impact (as defined by the Australian Research Council, 2017). Researchers were requested to focus on a particular project or a defined programme of research. The interviews were supplemented by curriculum vitae supplied by participants and internet searches for demonstrable evidence of impact such as media coverage and citations in government reports to produce twelve case studies of social work research using a standardised template. The case studies are listed in Table 1 and are available from https://www.griffith.edu.au/criminology-institute/our-research/our-projects/social-work-research.

Table 1

Case studies—impact of social work research

Name of researcher	Title of case study
Simon Biggs and Dina Bowman	Understanding and preventing workforce vulnerabilities in midlife and beyond
Briony Dow and Betty Haralambous	Improving the early detection of dementia, depression, and anxiety for people from Culturally and Linguistically Diverse (CALD) backgrounds
Mark Hughes	LGBTI ageing
Jill Wilson and Cheryl Tilse	The assets, ageing, and inter-generational transfers research programme
Elizabeth Fernandez	The well-being of children in out-of-home care
Karen Healy	Family participation in decision-making about child protection
Cathy Humphreys	Integrating child protection and domestic violence interventions to provide whole family responses when families experience domestic and family violence
Clare Tilbury	School to work transition for young people in care
Christine Bigby	Understanding and improving the quality of group home services for people with intellectual disabilities
Lesley Chenoweth	Impact of Local Area Coordination Programs in rural and regional Australia
Michele Foster	The operation and impacts of health and rehabilitation policy, systems and services for people who have chronic and life-long health conditions or disability
Grahame Simpson	Strength to Strength: Understanding and building resilience amongst families supporting relatives with traumatic injury

Data analysis

The case studies described specific examples of engagement and impact and formed the basis for the next steps of more conceptual data analysis. This involved theoretical coding of interview transcripts using concepts from the literature and a priori and in vivo coding to understand how participants described processes of engagement and impact. The coding facilitated comparison across cases. The approach, derived from Charmaz (2014), Corbin and Strauss (2008) and Silverman (2013), is set out in Table 2. Quotations from interviews have been chosen to exemplify themes, de-identified using a participant code. To calculate programme length and time to impact, researchers were asked about the date of programme initiation and first impact. Usually, the year a grant was awarded was given as the start date and first impact involved selecting the most significant event from a series of engagement processes. Calculations were based on whole years because participants reported the year in which an event occurred, rather than a specific date. The interview date (2018) was the cut-off for length of programme although in many cases the research continues to the time of writing.

Table 2

Approach to data analysis

Process	Attention to:
Preparation of the case studies	Descriptive account of programme of research, engagement and impact based on content analysis of interviews and other data
Theoretical sensitising using literature and notes	Avoiding prejudgement
Open coding	A priori codes (types of impact; strategies to engage)   In vivo codes   Code notes
Examining codes for categories	Constant comparison of incidents   Finding dimensions of categories through code and theory notes
Axial coding	Relationships between subcategories and categories, evidenced by original-coded data.   What caused the phenomenon?   What contextual issues were important?
Finding core categories and the story line

Ethics

Ethics approval was obtained from Griffith University human research ethics committee (Reference 2017/012). Confidentiality was not provided, as the case study listed publications and other research outputs. Participants had an opportunity to amend or correct their case study and consented to its publication.

Findings

The initial content analysis highlighted examples of impact that were categorised into three areas, based on those most frequently named by the researchers. These were impacts on social policy, service delivery and the quality of life of community members or clients. The 12 case studies represented 172 years’ work; programme length ranged from 8 to 23 years with an average length of time to initial impact of 6.9 years.

Changes to social policy

Research was perceived to contribute to the ways that human services policy issues were understood and defined. For example, research reporting perspectives from clients about their lived experiences or from practitioners about the difficulties of implementing certain policies or programmes, highlighted policy deficiencies that in turn had contributed new ideas about how to improve services. Sometimes naming and describing a problem—for example, family members are not always good custodians when managing the financial assets of their parents; or explaining why and how children are harmed by domestic violence between their parents—had reframed debate and opened up new directions for policy. But, as observed by one participant, naming a problem and describing its adverse effects on people did not always lead to change. Further examples of policy impact are listed in Table 3. In these examples, impact was mainly conceptual, in that it was the general messages from the research (rather than specific findings) that were influential.

Table 3

Examples of policy impact

Influenced policy debates: Research-based submissions were made to a public inquiry about age discrimination in the workforce and subsequently highlighted in media reports. Research-based briefings for parent advocacy groups and the social work professional association about parent participation in family group conferences influenced advocacy messages and were subsequently reflected in legislative changes.
Reframed or altered the understanding of a social problem:   Research raised awareness about the effects of domestic violence on children’s well-being and the mother–child relationship, leading to policy changes.   Research pointed out that improved life outcomes for care leavers requires casework attention to education and employment whilst in care, leading to policy changes.
Raised community awareness by publicly articulating new aspects of a problem:   Research gave attention to minority groups in aged care (culturally and linguistically diverse groups and LGBTI populations) and was subsequently reported in blogs, newsletters and websites targeting the general public.   Research exposed the long-term effects of institutional abuse of children in care, providing support for the compensation claims of survivors.
Changed public policy leading to new programmes, changes to legislation, budget allocations or regulations:
Changes were made to application forms for enduring power of attorney and advanced care directives so they could be more easily understood by people subject to them.
The experiences of older job-seekers informed the development of policies to enhance employment services for mature age job-seekers.
Research highlighting service shortfalls in rural areas informed government planning for the state-wide introduction of a major disability initiative.
Data collection fields were amended in a child protection database to improve outcome measurement.

Changes in service delivery and human service organisations

The analysis identified examples of specific findings from research that resulted in changes such as a new or improved service delivery practice or organisational process. In this category, research tended to produce implementable findings, meaning that cause (research findings) and effect (social change) were more discernable. For example, research found foster-carers lacked confidence and knowledge in talking with young people in care about future jobs, which led to the development of training and practice resources; a programme of intervention research led to a manualised intervention for families who have a family member with brain injury. But changes had also been indirect. For example, observational research on family meetings in child protection highlighted a lack of parental participation, which had led to a recommendation from a public inquiry for independent mediators to be used. Examples of changes to service delivery are listed in Table 4.

Table 4

Examples of changes to service delivery

Practice approaches, tools or resources were developed based on research
Research produced findings about family misuse of older family members’ assets, which informed service providers about how to improve decision-making arrangements.
A research-based career development booklet and interactive website for young people in care was developed and distributed, then used by foster-carers and practitioners in government and non-government agencies.
A research-based publication about how to improve the quality of group homes for people with intellectual disabilities was translated into a guide and check-list that was used by regulators when conducting visits for quality assurance purposes.
Assessment tools, guidelines and tip sheets were developed to enhance the assessment of older people with dementia and depression from culturally and linguistically diverse backgrounds.
Social work interventions developed and tested through research
The Strength to Strength programme, a manualised resilience-focused family intervention for rehabilitation from brain injury, was implemented in front line social work practice in multiple jurisdictions.

Improvements to the quality of life of community members or clients

Changes at the client or consumer level had not necessarily been direct and were frequently not measured. Except for one programme of research that demonstrated empirically that an intervention led to improved family outcomes, researchers had difficulty nominating examples of measurable impact on clients. There was research demonstrating a chain effect between an intervention and better practice. For example, early research had demonstrated the link between use of a practice known as Active Support in group homes for people with intellectual disabilities and their increased levels of engagement, which in turn are an indicator of quality of life. Further research demonstrated links between staff training in Active Support and their greater use of this practice. However, client outcomes were not directly measured, rather the level of Active Support practice in a group home was a proxy indicator of improved client participation. Similarly, researchers had used direct contact with community members to disseminate their findings and try to change knowledge and behaviours. For example, researchers had provided information sessions for family members of culturally and linguistically diverse populations using aged care services, aiming to improve community knowledge and social inclusion, although the impact of these had not been measured. Messages from research about how to improve service access for lesbian, gay, bisexual, transgender and intersex (LGBTI) groups to aged care services were taken up in the operations and management of some agencies, but improved access was not measured. Table 5 lists examples of potential impacts on clients.

Table 5

Examples of potential impacts on clients and community members

Information sessions were provided for families from culturally and linguistically diverse backgrounds about aged care services, aimed at improving knowledge about, and access to, services. Family outcomes were improved as a result of the resilience-based Strength to Strength intervention. Research demonstrated to decision-makers the methods for, and benefits of, consulting with patients, service users, carers, and family members about the design and delivery of health and disability services.

Dimensions of impact

In addition to identifying concrete examples of impact from the case studies, as set out above, data analysis also examined at a more conceptual level how participants described the process of achieving impact.

No straight line

Although the researchers perceived their research had led to positive changes to practices and policies, analysis showed they were hesitant to claim specific impact. One noted that social workers were possibly more realistic about change because of their orientation: ‘I think we probably know the complexity of it more because we’re in the practice world. So we actually know it’s not easy to show that something has changed, much less that it’s had any impact on issues whereas [other disciplines] naively think they do’ (participant 9C). A significant theme was that achieving impact was difficult to track explicitly because it was such a roundabout process. Participants explained that when many other people or groups had been involved in efforts to change policy and practice, it was difficult to know how research was influential when these other groups were saying similar things. They recognised that even though research may not have produced implementable findings, their research was part of providing a different framework for thinking about an issue. One participant noted that findings from research should be seen within a context of wider social change. That is, messages from the research had been more likely to be picked up in policy if they were consistent with the direction of social change already underway.

A voice in the chorus

Related to the lack of a direct line between research and research impact, analysis showed that whilst participants perceived that they had made a contribution to change, they considered their research was not decisive: ‘We are a voice in a chorus towards something’ (Participant 3A). They recognised they had contributed to framing a debate, although they did not position themselves as the main player and were more inclined to say they had empowered others. They indicated that pushing a change agenda requires resources that academics may not have; therefore, they needed to get other stakeholders to run with the key messages from their research: ‘we provided the evidence but we haven’t got the resources to personally push a lot of these agendas. So we really do rely on putting it out there and then … getting other people to run with it’ (Participant 2A). Noting social work was a female-dominated profession, a participant observed that men seemed to be more willing to claim influence, whereas women were more likely to claim membership of a team that had influenced change.

Invisibility

The lack of visibility of social work research in policy discourse in many fields of practice was noted by participants. Most researchers had operated in multidisciplinary research teams in fields of practice in which social work had been one of multiple occupational groups. This was a plus because the multidisciplinary team had been valuable in accessing and influencing different audiences. However, social work’s contribution had often not been named or identified, with one participant ascribing this to social work not having a national framework within which to develop and promote a knowledge base, like an academy of social work. The perception was that frequently, social work research had not been badged as ‘social work research’ or coded with the social work Field of Research code, which would have promoted its visibility, at least in the academy. There may be many reasons for this; for example, one participant noted that their research was not directed at improving social work practice; rather, it showed the need for broader social change through reform of legislation: ‘although it came out of social work, to be honest it never occurred to me that it was important to say we were social workers’ (Participant 14A).

Strategies to achieve impact

The analysis explored the strategies researchers had used to achieve impact and position themselves to influence policy and practice. Three main interlinked strategies were identified as engagement with research end-users; establishing authority and capacity building.

Engagement

Engagement with research end-users was focused on the specific research undertaken and efforts to mobilise that knowledge in the field. Engagement had included a wide range of activities and the production of various types of non-academic outputs. These were considered essential because of difficulties experienced by human services organisations in accessing scholarly literature. The interaction with research end-users was undertaken to communicate their research findings to the field. The processes of advising, informing, collaborating, disseminating and advocating were integral to the research agendas of participants. For many, their research had been funded by ‘industry’ grants (from government or human service organisations) or the Australian Research Council’s Linkage Program, which required having an industry partner. In turn, this provided a platform for them to link with research end-users. The researchers involved with the twelve case studies were productive academics as well as being strongly engaged in policy and practice realms. They had been clear about the purpose of their research—to bring about change in a policy field—but had unfortunately infrequently recorded engagement activities. There had been multiple audiences for their engagement activities and outputs, ranging from defined and specific to numerous and dispersed. The audiences included the following:

• policy-makers and government officials with power to bring about change

• practitioners

• advocacy organisations

• other individuals with influence or power

All participants had been engaged with non-academic audiences although some more purposefully than others in seeking to influence practice. Some noted that this had changed over the course of their career, as their confidence to exert influence had grown alongside their experience: ‘I didn’t think consciously about impact initially, I was more concerned about building a body of knowledge to provide evidence to make decisions .. Later I began to think more strategically about where I published and what I did to get findings out’ (Participant 7C). Some researchers had a clear plan and took an active, hands-on approach that involved personal contact with the targets for change, or formation of alliances with others advocating for change. The engagement activities of others were more indirect, opportunistic and ad hoc, and had involved for example, making research outputs available for intermediaries to use or empowering others to take on a knowledge dissemination role. They also recognised that impact could have occurred without their knowledge. Table 6 lists examples of the activities and products used to engage with research end-users.

Table 6

Examples of engagement activities and products

Engagement activities
Offering professional development and training (knowledge and skills development)—presentations, seminars, lectures and workshops for human service organisations   Presenting findings from research at field-specific conferences for academics, service providers and policy makers   Undertaking media and public awareness activities   Establishing industry partnerships and repeat contracts with industry partners   Taking up industry appointments, secondment arrangements between industry and universities and co-funded positions   Participating in policy and professional networks   Developing partnerships and collaborations with consumers
Products (non-academic outputs)
Research reports and research summaries for industry—often posted on industry partner websites   Submissions and media releases on public policy issues   Media stories, social media posts and articles on public websites   Practice resources, guides, fact sheets, training materials

Establishing authority

Establishing authority was a broader process of gaining credibility with a range of non-academic stakeholders, so they were receptive to messages from research. This had involved aligning with platforms or umbrella groups aiming to change policy and practice as well as conducting research projects in partnership with human service industry partners. Undertaking research, publishing and achieving academic status were perceived to have conferred authority and expertise in policy and practice debates to some researchers: ‘you create your authorisation to be able to contribute to this area through actually having a body of work’ (Participant 11C). Longevity in a research field, creation of a programme of work, development of a critical mass of researchers and being able to draw on a practice background were all seen to be associated with increased authority. Cultivating individuals was often seen as a powerful strategy, although the investment this required had often been limited to the local level given the distances and time involved in travel between Australian cities. Proximity was perceived to aid relationship-building. Having become known to government or other political leaders or having achieved recognition as an individual (e.g. through awards) was seen to have conferred additional, personal authority (i.e. the authority primarily related to the individual rather than to the research).

Capacity building

The efforts of researchers to create the broader conditions that would promote social work research to diverse end-user groups were categorised as capacity building. Researchers had used various strategies to build their general influence more widely, beyond specific research programmes, to encourage research-informed policy and practice. These included teaching, research supervision and ensuring professional associations adopted research-informed positions on issues: ‘my influence has been through a range of strategies – some through research, some through advocacy, but the other piece is teaching students, that’s huge’ (Participant 4D). Strategies researchers had used to build capacity and increase sector-wide receptivity to research are listed in Table 7.

Table 7

Strategies used for capacity-building

Linking to policy or professional networks—for example, training other professionals or contributing to position papers of professional associations Promoting and disseminating research via events such as lectures, workshops and roundtables Undertaking public advocacy—for example, media releases, social media commentary Participating in partnerships and collaborations with consumers or umbrella organisations Undertaking research supervision, arranging field placements, teaching and curriculum development to influence future practitioners and to encourage research-informed practice Educating professionals about the importance of research-informed practice Encouraging researchers to include minority voices in research Holding leadership roles in professional associations or editing practice-based journals

Discussion

The analysis points to similarities between the impact strategies of social work research and those of other applied disciplines (Cherney et al., 2012; Cohen et al., 2015). There were many examples of social work research influencing social policy, human service practices and organisations, and (to a lesser extent) the quality of life of service users. However, for social work research the situation is complicated by its relatively marginalised status in policy debate, the lack of visibility of social work as an occupational group in many agencies and some ambivalence within social work research about claiming impact.

Nonlinear relationship between engagement and impact

Engagement and impact were intertwined, often because changing policy discourses and raising hidden social issues prepared the ground for subsequent, more direct impact. Likewise, academic and non-academic impacts were intertwined. Consistent with Cherney et al. (2012), the researchers in our study believed that research rigour, trustworthiness and academic credibility leveraged influence. They were all externally oriented to some degree and sought to have an impact, but they were not necessarily purposeful, planned and direct in furthering this ambition. Some were less direct in their engagements, whereby research had been made available for intermediaries to use, empowering others to act. This was an informational role, disseminating findings to a broad non-academic audience who may find it useful. Whether employing direct or indirect approaches, usually social work researchers had been a voice in the chorus, and recognising social change and policy reform was made by many stakeholders. This suggests that social work researchers either work with other people who are trying to get the same change, or indirectly assist intermediaries as they work for change. We found no evidence of the trickle down approach being effective (i.e. publish the research in a scholarly journal and it will be picked up if it is worthy). Similarly, Bastow et al. (2014) found that researchers with impact generally had an external orientation as well as academic credibility, contrary to the narrative that excellent research will get noticed by policy-makers without any need for active engagement.

The theme, a voice in the chorus, reflects the efforts of social work researchers to create consensus about an issue, an ethos of promoting teamwork rather than claiming individual pre-eminence and the practical difficulties of accurately tracking influence. The social work researchers in our study aimed to engage with policy-makers, professionals and advocates, and in doing so aimed to inject new thinking into policy debates. The result is that any achieved influence is often ascribed to the group and precise direct lines of influence (that x study resulted in y policy or practice change) become indistinct. Gender may also influence this, with a view that men more often promote their personal influence, whilst women pointed to their membership of a group or team achieving influence. There was a sense from some participants that social work may be its own worst enemy: poor at linking with the big picture and lacking resources (such as a national academy or strong professional association that values research) to push a research-led agenda for change and to claim impact.

The no straight line theme pointed to the scientific reality that achieving impact requires several strands of engagement by multiple groups, with the result that it would be inaccurate to attribute impact to any single study or individual. There are multiple stakeholders and knowledge claims operating alongside research in the policy process. This picture was complicated by the issues captured in the theme invisibility. Social work research is producing evidence or knowledge that seeks to have influence beyond the profession including, for example legislators, funders, managers, other professionals, paraprofessionals and consumers in a field of practice. It is not straightforward to translate messages from research into practice directly to social workers, as in Australia at least, there are many generic roles such as case manager or child protection practitioner that are not reserved for social workers, meaning there is no distinct social work workforce. More broadly, if social work is not an influential force in national debate, then social work research may not be recognised or categorised as belonging to the discipline of social work. Policy based on social work research may not directly cite it, with the result that social work influence is obscured. These findings suggest that academic status can position the social work researcher to influence policy and practice, but the core scientific problem remains to establish a direct line of influence between research and policy and practice change. Again, this problem is shared with many applied disciplines. However, for social work research, it calls attention to the tension between restricting the scope of the research to specific and possibly narrow goals in order to achieve impact and the broader aims of societal change that often characterise social work research.

Engagement is relationship-based

Proximity to the site of change affected capacity to impact. Researchers were more likely to identify impact when they had a direct and personal relationship with a service delivery agency. So geographical proximity and projects undertaken with agencies fostered impact. It is a challenge to reach beyond regional boundaries. To an extent, this is understandable because social work is oriented to state and national contexts that frame human service provision through legislation and policy. At an operational level, effective implementation takes account of context, and large service delivery systems are not uniform because they face different challenges of demography and geography. Nevertheless, there are global developments that affect social work practice worldwide and an international trade in social interventions. It is also a challenge to reach beyond immediate research partners to other human service organisations. This highlights the relational nature of engagement: it is enhanced by the strength of personal relationships that people can form in undertaking research. Relationships and roles can be advisory, such as board and committee appointments; contracting or collaborative projects; strategically building a research agenda together; dissemination and promotion; or joining with others in public advocacy or activism. However, a downside of relationship-based nature of engagement is that it is temporal—people change positions.

Many types of research can be influential

Research methods did not appear to influence impact; both qualitative and quantitative approaches could be influential. Qualitative studies may enhance understanding of a problem, reveal service user experiences and lead to different thinking about how to respond, whereas quantitative studies may demonstrate the scale of problems or whether a programme made a difference to outcomes. However, as the findings showed, it was always difficult to attribute changes to specific findings from specific research projects. Research often leads to evidence about the type of knowledge and skills that staff need to improve client outcomes, and researchers may develop training based on this evidence. However, further translation, evaluation and implementation research is necessary to identify whether the training has any benefits for clients. It was clear that impact was accumulated through programmes of research; single studies were rarely decisive. Targets of change might be specific (e.g. social work in hospital rehabilitation settings) or otherwise very amorphous (e.g. allied health practitioners in community settings). Whilst intervention research and human services industry research partnerships can be more direct and deliberate in aiming to impact on practice, other types of research can also be influential. Research may have impact by holding government to account or subjecting a policy to analysis and critique. This may have the effect of a planned change not taking place or plans being modified. For example, research may find that a government programme is not delivering its objectives, which leads to the approach being questioned or amended. However, it is difficult to measure the impact of research in preventing poor policy or practice.

Implications for social work

The findings broaden understandings of how social work research is used, the iterative relationship between engagement and impact and the types of research that are valuable. The strategies used by social work researchers are probably shared across applied disciplines seeking to influence policy and practice. The difference is that the background of invisibility, no straight line and a voice in a chorus may make these strategies more difficult to achieve and require additional engagement efforts. The implication for social work research is that additional resources may be needed to achieve impact compared with those required in other disciplines. Whilst social work is well-positioned to achieve impact given its applied nature, both instrumental and less tangible conceptual impacts should be counted. The research highlights the differences between discovery and engagement models of research impact. Linear research payback or research transfer models whereby the researcher produces knowledge for human service organisations to apply are not suited to social work, as evidence is generally being produced alongside practice (Bannister and Hardill, 2013; Gray et al., 2015).

The study of research impact should not be bound by bureaucratic definitions of impact designed to suit government policy objectives; rather, it needs to expand and critique such definitions (Alla et al., 2017). Our research supports a model whereby engagement and impact are intertwined processes, with productive exchanges of knowledge, feedback and advocacy amongst stakeholders, rather than there being a direct, identifiable process to achieve research outcomes (Spaapen and van Drooge, 2011; Greenhalgh and Fahy, 2015). It is necessary to be realistic about the place of research in social change and policy-making. Policy involves interactive networks of stakeholders all asserting interests and making knowledge claims. Research is one element only. Research can highlight issues of concern to practitioners and policy-makers and contribute to new analysis of existing issues—it may not answer particular questions, but it can change the way people think about what the questions are (Bastow et al., 2014).

Knowing what social work research impact looks like, and how it happens, is central to guiding its practice and measurement. Generating impact is a time-consuming, long-term proposition that requires multi-faceted engagement with human services industry partners. It is estimated to take an average of seventeen  years for research evidence to reach clinical practice (Morris et al., 2011). It is always somewhat arbitrary to state when work started (some programmes of research in our case studies commenced in the 1980s and it was several years before a first grant was awarded) and when first impact occurred. Impact is a process, rarely an event that can be marked at a point in time. It follows that documenting and achieving impact costs money. It requires infrastructure within universities, human services organisations and the profession, building the skillsets of all involved.

Limitations and conclusion

The findings from the sample of twelve case studies were limited to three fields of practice and are not generalisable. Nonetheless, they present a detailed picture of the scope and types of engagement and impact that have occurred. A limitation of the study is its use of researcher self-reports about how useful their research was to policy and practice. The impact was not verified, and researchers may have overclaimed or underclaimed their impact.

The value of social work research goes beyond its instrumental and short-term uses. There are many and varied ways that research can influence policy and practice over time, and this study details the range of strategies that social work researchers have used to engage with research end-users. Because it is an iterative process, an external orientation to the field is advantageous to having an impact on policy and practice. The research demonstrates the benefits of collaborative, developmental research designs that incorporate long-term engagement from the problem formulation stage to dissemination and implementation. It puts the focus on programmes of research (rather than single studies), on the importance of research partnerships to achieve impact, and on the process of knowledge production—engaging with research partners in policy and practice to design and do research that matters. These imperatives hold regardless of government policy pertaining to impact assessment. Nevertheless, because research engagement and impact are being measured and assessed by government and universities, and the process of measurement assigns value, social work researchers should record their engagement activities and document evidence of impacts, both direct and indirect.

Funding

Funding for this research was granted by Australian Research Council DP170102412 Advancing the production, utilisation and impact of social work research to generate innovation in human services.