In the first of two practice articles in the special issue, Bammer1 acknowledges that there is an extensive literature on the related topics of co-creation, co-production, co-design, stakeholder engagement and participatory research.
She cautions, however, that although some authors recognise non-linear chains of causation, complex interdependencies and other manifestations of complex problems, the literature is mostly based on problems being treated as ‘tame’, that is, those with a clear mission and solution.
What is different about co-creation when the problem under consideration is complex, or ‘wicked’? Examples of complex problems include: mitigating the effects of global climate change, reducing social and health inequalities, and combatting organised crime.
To explore this question, Bammer’s first step is to transform the characterisation of ‘wicked’ problems from the planning context in which they were conceived to a research context. Because the term ‘wicked’ is not easily translatable into many other languages, the term ‘complex’ is preferred.
Five key characteristics of complex problems are proposed for the research context:
- Difficult to delimit. Complex problems are open systems problems where everything is connected to everything else, therefore there are no clear boundaries to the problem.
- Contested problem definitions. Different disciplines and stakeholders will define the problem differently and the definition will also change as the context (for example, political, economic or geographic) changes.
- Multiple uncertainties plus unresolvable unknowns. Everything cannot be known or found out about complex problems. Further, ignoring unknowns is a source of major adverse unintended consequences and nasty surprises.
- There are real-world constraints on what can be done. Policy and practice responses to complex problems – whether they are by government, business or civil society – are constrained by political, economic, historical, cultural and other circumstances.
- Solutions are not true/false, but good/bad, and they are partial and temporary. The interrelationships among the characteristics of complexity described above mean that perfect solutions are not possible.
Dealing with any complex problem requires understanding how these five characteristics play out for that problem, as well as how actions can best be shaped around them. In turn, this requires inputs from multiple disciplines and stakeholders. The focus of Bammer’s paper is limited to stakeholder inputs.
Bammer alerts that it is also worth noting that simple (or tame) and complex (or wicked), should be seen as endpoints on a continuum rather than as a dichotomy. Further, researchers may embrace or choose to ignore elements of complexity in their approach to a problem. For example, improving a treatment service, which is often approached as a simple problem, may have elements of complexity. Service failures may be systemic; different stakeholders may define the problem differently; some unknowns may be hidden (unknown unknowns); the history, geographical location, and/or cultural context may be significant; and the solution may involve trade-offs.
The aim of Bammer’s paper is to start to lay out what complexity means for co-creation, as well as best practices for co-creative research on complex problems. A framework (the Integration and Implementation Sciences or i2S framework) is introduced that allows the elements of complexity to be systematically unpicked and dealt with. The paper then takes one tool – for thinking about how stakeholders are engaged in co-creation – which is relevant to the whole i2S framework and examines three aspects of the use of this tool when the research problem is complex.
The development of Integration and Implementation Sciences (i2S) had its origins in a seven-year research programme that Bammer led from 1990 to 1997. This involved examining the feasibility of prescribing pharmaceutical heroin as a treatment for heroin dependence in Canberra, Australia’s capital city. Reducing illicit drug use is a complex problem. Although the research was focused on one potential response, that is, trialling pharmaceutical heroin prescription, the feasibility research was structured to explore the benefits and risks of that response, rather than seeing it as a perfect solution.
In this programme, the research team worked closely with stakeholders affected by the problem, specifically illicit drug users, ex-users, the families of drug users, and the general public, as well as stakeholders in a position to do something about the problem, namely police, treatment and other service providers, and policy makers. The research was unusual in that there were no major funding or time constraints, meaning that the research team was not limited in which stakeholders they could engage or how. This gave the team a unique opportunity, within one programme of research, to learn multiple lessons about stakeholder engagement and co-creation. Although the research was conducted more than 20 years ago, the lessons are still pertinent and provide a range of examples for the key points of the paper. The i2S framework has subsequently been tested against other complex problems relating to population health, the environment and national security, but none of these has provided a richer set of illustrations for co-creation in research on complex problems.
What does complexity mean for co-creation?
Bammer advises that different stakeholders will have different, often unique, perspectives on the five elements of complexity, that is, the problem’s connections with other problems, different ways the problem can be defined, important uncertainties and unknowns, some of the real-world constraints on what can be done, and what will make a good solution.
For example, in the heroin prescription feasibility research, illicit drug users made the research team aware of the connection between drug dependence and, for some of them, childhood sexual abuse. Some parents of illicit drug users defined the problem as lack of support by intervention services for young people who had lost their way, while others saw the problem as a failure by society to shield young people from accessing drugs. The police drew the research team’s attention to a critical unknown, namely how to prevent an influx of drug users to a city where heroin was provided on prescription. Policy makers pointed out the real-world constraint that heroin prescription would have to fit within Australia’s international treaty obligations. Some of the general public thought that making heroin available on prescription was a good option, others opted for strengthened law enforcement, yet others wanted to focus on different treatment options, and others again wanted to see heroin use legalised.
As these examples illustrate, stakeholders have an essential and intimate role in making complexity evident. As a consequence, co-creation on complex problems requires a specific research mindset towards stakeholder involvement. Stakeholders help co-create an understanding of the problem, as well as charting a path forward in dealing with it.
Bammer alerts that this contrasts with the way most researchers have been trained, which is that the problem, and how it will be dealt with, is defined by one or more disciplines. Further, the understanding and path forward will vary depending on the mix of stakeholders involved in the research. In the heroin prescription feasibility research the team aimed to be comprehensive, and their appreciation of the problem and recommendations would have been different if they had only worked with treatment providers, or if they had confined input to a range of stakeholders committed to continuing prohibitionist policies.
How can research take into account multiple stakeholders with multiple relevant inputs? Bammer advises that the Integration and Implementation Sciences (i2S) framework is specifically designed for dealing with complex problems, and begins by considering three core research domains:
- synthesis of existing knowledge
- developing an understanding of critical unknowns and figuring out what to do about them
- making the research evidence available to support policy and practice change.
For each domain a set of fine-grained considerations is then teased out to guide the researchers in planning or documenting the research overall (Figure 1). When used for documenting the research, the questions elicit decisions made about a) the problem (for example, ‘how was the problem defined and framed’), and b) the methods employed in the research (for example, ‘which systems methods were used’), as well as prompting self-evaluation. These in turn help the researchers capture what they engaged with stakeholders about and the specific methods they used.
In planning or documenting research on complex problems, the fine-grained questions and sub-questions can generally be asked in any order. The key issue is that all questions are addressed.
What does complexity mean for how stakeholders are engaged?
Two frameworks are widely used to describe how stakeholders are engaged: Arnstein’s ladder2 and the International Association for Public Participation (IAP2) spectrum of participation3. It is noteworthy that neither was developed for the research context; instead they address public participation in government decision making. Most research studies have used these frameworks to identify where in the ladder or spectrum they are operating, rather than in any more analytic way.
Bammer advises that rewriting the IAP2 spectrum for the research context provides a useful framework for exploring how stakeholders are engaged, including on complex problems (Figure 2).
The five types of engagement in the original spectrum can be easily translated from public decision making to research. Only one modification is required to make the spectrum relevant for complex problems. In the stakeholder participation goal for ‘collaboration’, ‘each’ is replaced with ‘salient’ as in ‘Researchers partner with stakeholders for salient aspects of the research process’. While full collaboration may be possible on a straightforward research question (with, for example, few stakeholder groups and one simple form of data collection), it is unlikely that any stakeholder groups will be able (or even want) to contribute meaningfully to each aspect of the research on a complex problem.
Three issues are particularly pertinent when the research problems are complex. These are the ability to:
- involve different stakeholder groups in different ways
- think deeply about which forms of stakeholder engagement are likely to be most appropriate
- take into account when researchers do and do not have control and different categories of stakeholder groups.
Involving different stakeholder groups in different ways
The particular value of the research-relevant modified IAP2 spectrum when the research problem is complex is that – given that such research generally involves a large number of stakeholder groups – it makes it possible to consider that different groups (separately or together) can be involved in different ways. Further, over the course of the study or in different aspects of the study, the same stakeholder group can also be involved in different ways.
A useful illustration comes from the inclusion of illicit drug users and police in the heroin prescription feasibility research. Each group was involved in three major ways. First, both groups, as a whole, were informed about the research, using a variety of mechanisms tailored to the specific group. Second, police representatives (chosen to represent a spectrum of views about heroin prescription) collaborated in designing and interpreting a survey of police; similarly drug user representatives collaborated in a survey of their peers. Finally, representatives of each group were consulted about how the research findings would be presented to policy makers.
Thinking deeply about which forms of stakeholder engagement are likely to be most appropriate
The modified research-relevant IAP2 framework not only provides a way of describing different levels of engagement, but also provides a way of thinking deeply about which forms of engagement are likely to be most appropriate, given the different aspects of complexity the stakeholders can provide information about, including delimiting the problem, problem definition, relevant unknowns, real-world constraints, and possible ways forward. The point is to think about these in the overall context of the research and the mix of stakeholder groups involved.
For example, in the heroin prescription feasibility research, the research team recognised a major conflict between those who thought action was required to strengthen prohibition (and therefore opposed heroin prescription) and those who argued for drug law reform (and therefore supported heroin prescription). This reflected different values about self-direction, hedonism, conformity and benevolence. The team wanted to be even-handed, respecting both value systems, and that affected their collaboration with different stakeholder groups. ‘Inform’, ‘consult’ and ‘involve’ were all relatively unproblematic, but the team had to be much more careful when they wanted to ‘collaborate’ with or ‘empower’ stakeholder groups, because these are much harder to implement in a neutral fashion.
The team did engage through collaboration and empowerment when it worked with the local Aboriginal and Torres Strait Islander community, however. Not only was this the most appropriate and respectful way to work with them, but they also represented both sets of values, so that maintaining neutrality was not an issue.
Take into account when researchers do and do not have control and different categories of stakeholder groups
One aspect of co-creation that is made evident by the i2S framework but rarely discussed is that while researchers have a high degree of control over the processes pertaining to knowledge synthesis, and understanding and managing unknowns, they have little control over supporting policy and practice change. Knowledge synthesis and understanding and managing unknowns are the core business of research, and it is up to researchers to cede some control to ‘let in’ stakeholders. On the other hand, policy and practice change is not core research business, and it is up to decision makers to cede some control to ‘let in’ researchers. Researchers can determine what they set out to do, but not whether any necessary interactions with decision makers are achieved or how their inputs are acted on.
It can also be helpful to specifically consider stakeholders in two categories:
- those affected by the problem
- those in a position to do something about the problem.
This is not a hard and fast distinction, but rather a useful heuristic.
The research-relevant modified IAP2 spectrum works for both categories of stakeholders, but there is an additional set of considerations under ‘empower’, especially for stakeholders in a position to do something about the problem. Researchers can seek to increase the power of decision makers to achieve change. Most commonly, this involves applying pressure from outside the policy or practice systems – through advocacy – with the aim of strengthening the influence of decision makers arguing for particular action on a problem.
Researchers can do this by advocating for particular outcomes themselves or they can seek to strengthen the advocacy of stakeholder groups affected by the problem, by giving their perspectives legitimacy through including them in the research. Researchers can also take the additional step of activism, where they specifically contribute to the lobbying and other actions required to make change happen. There is a substantial discussion to be had about the appropriate roles of advocacy and activism by researchers and how closely they should engage with decision makers in the research process, but that is beyond the scope of Bammer’s paper.
In the heroin prescription feasibility research, the starting point was to keep decision makers informed about the research, and – where they were willing – to consult, involve and/or collaborate with them. The research teamchose a neutral stance, reviewing carefully both the potential costs and benefits of heroin prescription, and therefore eschewed advocacy and activism.
These distinctions about control and different categories of stakeholders are also useful in considering how the specific elements of the research-relevant modified IAP2 spectrum play out in practice. Combining them, as shown in Figure 3, provides a useful way to tease out key questions to guide researcher engagement with stakeholders. The point here is not to be comprehensive, but to illustrate that the type of stakeholder involved, and where in the research process they are involved (understanding or action) influences the considerations that are relevant.
By and large these questions are self-explanatory, but it may be helpful to illustrate four of them with examples from the heroin prescription feasibility research.
First, on the question of action and urgency from the first quadrant of Figure 3, the stakeholder groups mostly agreed that illicit drug use was a problem that needed action. But there was disagreement on urgency, for example families of drug users tended to see the need for considering heroin prescription as urgent, whereas some treatment providers thought that existing treatment services should be improved first.
Second – relevant to quadrant 2 in Figure 3 – although this was not the case in Australia, in some countries it would not be possible for government policy makers to be seen to liaise with ‘criminals’ such as illicit drug users, so that research may be the only way for these stakeholders to be heard. This also relates to the issue of accessibility and organisation in quadrant 1 of Figure 3.
Third – relevant to quadrant 3 in Figure 3 – in the heroin trial feasibility research, policy makers from the affected jurisdiction gave the research legitimacy by putting discussion of the progress and findings on the agenda of an influential national policy committee, ensuring that the committee stayed abreast of what was in train and eventually made a decision.
Finally – relevant to understanding trade-offs important for decision makers (quadrant 4 in Figure 3) – early in the research, key politicians made the research team aware of the perspectives that would be most influential in their decision making, including those of the police commissioners and the key medical professional association, so that they were able to ensure that the team included those perspectives in the research.
Article source: Adapted from the paper Key issues in co-creation with stakeholders when research problems are complex published in the Evidence & Policy special issue Co-creative approaches to knowledge production: what next for bridging the research to practice gap?, CC BY-NC 4.0.
Acknowledgements: This series has been made possible by the publication of the special issue as open access and under a Creative Commons license. The guest editors and paper authors are commended for their leadership in this regard.
- Bammer, G. (2019). Key issues in co-creation with stakeholders when research problems are complex. Evidence & Policy: A Journal of Research, Debate and Practice, 15(3), 423-435. ↩
- Arnstein, S.R. (1969). A ladder of citizen participation. Journal of the American Planning Association, 35(4), 216–224. ↩
- International Association for Public Participation (IAP2) (2014). IAP2’s public participation spectrum. ↩
Also published on Medium.