Academic–industry collaborations and knowledge co-production in the social sciences

Initiating collaborations – the role of prior relationships

The interview data indicated that a prior relationship between academic researchers and external partners was linked to the willingness of both parties to invest in collaborative work. Having a prior relationship helped to establish a reference point on which to base judgements about the risks and benefits of entering into collaborative research and this was important for industry partners (Bogenschneider and Corbett, 2010; Contandriopoulos et al., 2010; Haynes et al., 2011a, 2011b). For ARC industry partners this was particularly salient because research collaborations can generate scrutiny of their policies and practices. For instance, in all but 2 of the 30 projects studied there was some type of prior relationship between one or more of the chief investigators (CIs) and industry partners. This was typically based on the CI having worked with, or for, an industry partner. For example, one academic CI stated:

Yes, well in particular with [agency name removed] because I’ve worked for them for so many years, we’re friends. In fact, if I ever have a need for an industry partner, they’ll – I can almost guarantee they’ll be on it. But pretty much because we meet regularly, ideas emerge with them and then they’re very good at facilitating other people’s involvement. (CI interview 2)

The importance of ongoing contact was central to building a sense of trust between industry partners and academic CIs. The interview data illustrated that trust in the motives and expertise of academic CIs was particularly salient for industry partners, who sought reassurances that the research was to be conducted with their needs in mind. For example, one industry partner representative made the following statement in reference to reasons for his reported low level of input into and oversight of a Linkage project, referring specifically to the trust he had in the CI:

he’s been involved off and on in a very fundamental way with the functions of the … Council.… I, in fact, hired him myself, not so much for land claim work but for other functions of the … Council…. But he goes back much further than that … if it was someone who wasn’t of his standing, and as thorough and as esteemed as him, I probably would have – either myself or designated someone to have more input.… But I and the Council have a lot of faith in [CI name removed] … that they’re doing things ethically and thoroughly and in accordance with the aims of the … Council. So, we had a little bit of – I had a little bit of input into project design and then otherwise left it to him. (IP interview 19)

Ongoing connections and relationships of trust were also seen as important to the quick resolution of problems related to data collection and in the dissemination of results. Over half of the academic CIs interviewed highlighted this point. However a prior relationship with a partner organisation was no guarantee that a specific Linkage project would have an influence on the policies and practices of industry partners. More about this issue will be canvassed below.

Industry partner involvement in project design and implementation

One largely unexplored issue in the research literature is how dense and reciprocal do research collaborations have to be to conclude that knowledge co-production has occurred. This is linked to the levels of input that partners have into project design and implementation. In the context of academic–industry collaborations this is particularly relevant because academic researchers can place a great deal of value on decisions relating to research methodology and data analysis, given their importance to an academic’s occupational identity and role. Academics may be unwilling to compromise in such circumstances due to a fear it will undermine the rigour of the research process. However, for a collaboration to be a two-way process participants need to be willing to vary their level of control over different stages of the research process.

This issue of input into collaborative research was investigated by a series of questions about the level of input CIs and partner organisations had into the design and implementation of the Linkage project. Academic CIs always reported consistent levels of input, given that they typically led the project design, data collection and analysis phases. For ARC industry partners there were three main types of industry partner involvement that emerged from the data: formal supporter, responsive audience and integral partner (adapted from Ross et al., 2003). Each involved different levels of input into the research process. Table 1 outlines how these models were evident in the current study.

OPEN IN VIEWER

Table 1.

Models of industry partner involvement (adopted from Ross et al., 2003).

Model	Characteristics
Formal supporter	Industry partner(s) support did not go beyond cash contribution and was a passive participant; either serving as a research site or funding project costs e.g. data collection. Industry partner(s) did not have any input into the grant application.
Responsive audience	Industry partner(s) were involved in the research process by responding to researcher approaches with ideas, information or tactical advice on how to execute the fieldwork or address problems occurring during the completion of data collection. Industry partner(s) provided feedback on draft grant applications.
Integral partner	Industry partner(s) shaped research design by helping to write sections of the draft grant application; had input into the formulation of data collection instruments, interpretation and dissemination of results. Senior decision-makers were actively part of the process.

The responsive audience and integral partner model are more aligned with how knowledge co-production is envisaged. However, what became clear from the data was that the categories outlined in Table 1 should not be seen as mutually exclusive, and it was difficult to say with certainty how many of the cases investigated fell within each category. This is because partner involvement tended to shift throughout the duration of the Linkage project. Many of the collaborations studied went through a cycle that saw industry partner involvement shift from an integral partner approach at the beginning, to a responsive audience or formal support model towards the end of the project. This shift resulted from a range of factors, one being that interest in the project on the part of industry partners tended to diminish over time. Also, once data collection was completed, often the task of translation and dissemination was left to the academic researchers. This is a major weakness with academic–industry collaborations.

While the complete withdrawal of industry partner support was not observed in the current study, it must be noted that industry partner participation was in many instances unstable. For example, one problem consistently reported across 16 of the 30 ARC Linkage projects examined was that of staff turnover in industry partners. This particularly was a problem with the nominated partner contact person. This observation was consistently made by CIs working with government departments. Some CIs were more prepared than others to deal with this issue, having strategies to bring new industry partner representatives up to speed with the project, or selecting people in agencies they knew were stable. While the turnover of staff within government agencies is not all that surprising, it does create stability problems for research collaborations when new people come on board; in such cases the collaboration can falter. For instance, in a number of cases the departure of an industry partner contact person proved fatal, because it meant losing a key champion of the project, with it then proving harder for CIs to generate or encourage industry partner input or to sustain interest. One CI described the situation this way:

So the people that I’ve talked to initially about doing this project have long since left [agency name removed] and not only that, but then the people who were there at the beginning of the project have also left. I don’t know if all government departments are like this, but certainly in [agency name removed] there’s been a lot of staff turnover. When we deliver a final report it will be to about the fourth person that I’ve dealt with over the life of the project.… We’re not quite sure who to deliver the report to … the person who was interested in the project at the beginning has left and they’re not quite sure why we’re doing it or what its purpose is. (CI interview 1)

Strategies of communication and dissemination

Strategies concerned with translating and disseminating research are particularly important in collaborations involving academics and industry partners given they have a bearing on research impact. This needs to involve direct and interactive dialogue and engagement with industry partners (Bogenschneider and Corbett, 2010). An important empirical and practical question is how this is best achieved, and what factors inhibit information dissemination and exchange?

In the current study CIs were asked about the methods used to communicate and disseminate results to collaborating organisations, with there being a reliance on a mix of informal approaches (e.g. emails, phone-calls, ad-hoc meetings), organised face-to-face gatherings (meetings and workshops) and written methods (i.e. summaries and reports). Some CIs placed emphasis on informal strategies while others relied mainly on written methods. While no clear strategy of success for communicating and disseminating results emerged, there was recognition that a reliance on written reports was limited in scope. As many CIs commented ‘the industry partner was committed but rarely did anything with a report’. Some industry partners’ cash and in-kind contribution was based upon assisting with dissemination, which involved organising conferences and workshops. This was regarded as helping to expose the research to potential end-users and provided avenues for industry partner representatives and other stakeholders to have input into the interpretation of findings.

Across all the collaborations studied there existed a tension between the need for academics to publish in traditional academic outlets and the translation of results to industry partners and other end-users. The two were not often seen as compatible. In some instances CIs made it clear they would ‘only go so far’ with disseminating and communicating research findings to industry partners and other stakeholders. One CI described this tension in the following way:

historically we’ve been very keen to produce outputs that are quickly available and accessible to practitioners and to policy people. Well those are not Grade A journals or typically books.… But over the past three years I guess in the time of this project we have felt considerable pressure … to have outputs that are A-level journals. In fact that sort of conflicts with why we think we’re here and it conflicts with what we have experienced as the potential impact of this research we’re doing. (CI interview 22)

Hence cultural and institutional pressures and demands within academia (i.e. to publish in peer-reviewed journals) created disincentives against investing heavily in research translation to industry partners. This has been observed in other studies (see Cherney et al., 2012a, 2012b). This result gives weight to arguments that closer interactions may not bridge the divide between academics and potential end-users of social research, because organisational incentives can generate opposing priorities, especially academic contexts where dissemination and translation to non-academic end-users is not seen as a legitimate form of scholarship.

The impact and benefits of collaborations

To examine the issue of social research uptake this study divided data into the categories of conceptual, instrumental and political impact that is often cited in the literature (Nutley et al., 2007). Conceptual impact was consistently reported across all 30 projects. CIs and industry partners made reference to the Linkage project as having had an ‘impact on thinking’, or that it ‘provided insights’, ‘changed understanding’ and ‘clarified issues’ relating to the research problem or the industry partner role. Industry partners consistently reported forms of conceptual impact without the collaboration having a direct influence on policy or practice. For example one industry partner stated:

No, it has had no real impact … or hasn’t had it as yet.… We feel happy that we’ve managed to contribute to furthering knowledge but it hasn’t had any great trickle-through effect … it changed conversations around the table. (IP interview 1)

When it came to instrumental forms of impact, CIs and industry partners made reference to a number of instances where projects had led to concrete changes in policy or practice. Across the 30 projects studied, 15 reported examples where the collaboration had directly influenced policies and practices within industry partners. This included changes in the training of industry partner personnel, improvements to the way industry partners interacted and supported clients, and reforms to legislation and policy in state and federal government. For example, one project examining the problem of whistleblowing reported a diverse range of outcomes ranging from impacts on legislative design and policy, with one industry partner reporting:

Well the research, apart from talking about what the legislation should look like, also talked about what good practice would involve in relation to the sorts of policies and procedures that should be in place. We have modified our model, internal reporting policies in the light of that.… We’re modifying our protected disclosures guidelines in the light of that…. It then talked about what is actually good practice in the management of whistleblowers…. We are building that into our guidelines, into the training of our staff. I also use a lot of the research results in the training that we’ve been doing for many years now of public officials across [identifying information removed] in the better management of protected disclosures. (IP interview 2)

When prompted about the types of outcomes reported in the quote above, the industry partner interviewee stated the project ‘is my baby’ and that it was not ‘just an academic exercise’, commenting the CIs were ‘not just a bunch of academics thinking about it [i.e. the problem of whistleblowing] theoretically’, they were not ‘hung up on the theoretical’. These statements highlight the important role of organisational champions in helping to facilitate the impact of research collaborations internally within industry partners. This result though was not unique to the specific project referred to above; it was a consistent theme across collaborations that reported direct forms of influence on policy or practice. For example, an interviewee from a collaborating organisation that reported impacts on their training and educational practices stated unambiguously that he was ‘personally passionate about the area’ the ARC Linkage project was addressing. This respondent reported he was the main industry partner contact person for the length of the project, despite organisational restructuring that occurred during the duration of the ARC Linkage.

This is an important finding because it points to what may differentiate successful collaborations from unsuccessful ones. It provides support for the conclusion that the capacity of academic–industry collaborations to shift organisational processes and transform policy and practice, can only be achieved with the assistance of agency personnel who understand the rationale and justification of the research collaboration, and are willing to act as change agents (Cherney and Head, 2011; Mitton et al., 2007; Meyer, 2010; Rickinson et al., 2011). Among the projects studied, these individuals were not junior personnel, but typically carried weight and influence within their respective agency.

An unexpected result that arose in the interviews with collaborating organisations was the diverse way non-academic partners understood research impact. Industry partner interviewees stated that the Linkage project had the benefit of enhancing their internal capacity to access expertise and research, and in some circumstances linked the industry partner with other agencies. While the research collaborations studied clearly enhanced the networks of academics, they also provided the same benefit to industry partners. This only came to the fore once the interview moved beyond examining specific types of impact and focused on the benefits arising from participating in a Linkage project. One industry partner stated:

It’s made us more aware of our lack of skills.… It’s given us confidence to be a partner in another Linkage.… The outcomes of the research was one thing.… For us to build those relationships and have us associated with this research and people understand that our work in communities is taken seriously and gives us good credentials … and being part of the project has given us a different set of skills and stakeholders we can tap in to … both individually it’s brought us a great deal of benefit. (IP interview 6)

Another unexpected result was the fact a number of industry partners stated that the ARC Linkage was the conduit through which agencies for the first time actually began to work together (i.e. it generated joined-up work). These results show that gains from collaboration go beyond purely instrumental outcomes (i.e. influence on policy) to include impacts relating to capacity building.

Collaboration was not simply generated by a desire to enhance organisational capacity to improve policy or practice, but was also driven by strategic considerations. With it increasingly the case that government and non-government agencies lack sufficient funds and expertise to conduct their own ‘in-house’ research, ARC Linkages were seen by industry partners as a way of broadening their evidence-base to influence policy decisions. Through the mechanism of collaboration the evidence produced was seen by industry partners as far more persuasive, credible and legitimate, given that it involved academics removed from the political process and could be less subject to criticisms of being biased. The quote below illustrates this point:

we were more interested in getting the right answer, getting a real answer rather than just dodging. Like, we’d do our own surveys but they’re not particularly robust in terms of their outcomes…. But a key part of the research was also then communicating externally as well. So we needed something that was above board and clearly credible. So their ability [in reference to the CIs] to provide the level of statistical analysis and also being hands off from us was also a [unclear audio], whereas if it had just been research by us, even if we’d done it completely and appropriately there was always going to be a question about whether it was just us producing the answers … in relation to engaging with the State Government and the bureaucracy, having that level of credibility was important. (IP interview 20)

The data above show how organisational processes and political imperatives influence the motivation to participate in research collaborations. This can though potentially compromise their integrity and the objectivity of results (Keiser and Leiner, 2009). No academic respondent stated that they felt pressured to produce favourable results for industry partners. Rather, industry partners were mindful of not wanting to tarnish the integrity of data collection or analysis, fearful that doing so would weaken the case for using project results to influence key decisions.

Even if academic–industry collaborations help to generate practical evidence that has implications for policy, it does not mean it will have an immediate impact. The study data indicated that for collaborations to be effective consideration needs to be given to how co-produced knowledge can be integrated into policy and practice – this relates again to the practicalities of translation and uptake. For example, many CIs stated that the uptake of project results was inhibited by organisational imperatives and contingencies such as timing – ‘the agenda has moved on’ (CI interview 1) – with the project competing with day-to-day priorities: ‘you’ve got to insert yourself into their schedules and reacquaint them with your agendas, because they’re not at the forefront of their daily work’ (CI interview 28). Industry partners also recognised that internal organisational processes could inhibit research translation and uptake, with procedures absent to ensure that the findings arising from the ARC Linkage were followed through and exploited fully. The quote below illustrates this point:

After the report was submitted to us, we actually undertook a fairly detailed review of the findings…. We then spent a couple of days with a cross-section of people from inside our business at a workshop where we sort of reviewed all the recommendations and findings and worked out where to from here.… There was a project team, if you like, because the recommendations spanned a number of different disciplines inside our business…. We developed really a programme of activities or actions arising from the report that were then fed back into different streams in the business to follow through…. I guess we wanted to capture all the value out of it that we could.… I think there’s been a big gap between even that review workshop where we worked out what is the value of this to us, what we should do about it and then the implementation of the actions that arose out of that too. Because then particular people were tasked with responsibilities for following certain things through…. I think where we probably fall down is in the next loop back to close off on some of those actions. (IP interview 3)