By Gabriele Bammer.
In research tackling complex societal and environmental problems, what is involved in taking context into account when aiming for more comprehensive understanding of a problem, as well as research support for policy and practice change so that the problem is better dealt with?
As outlined in my work developing Integration and Implementation Sciences (i2S), considering context involves dealing with circumstances at three levels:
- social or big picture, eg., the historical, political, economic, cultural, geographical and other circumstances
- organizational, especially the structure and culture of the research and stakeholder organisations involved
- individual, especially the positionality or identity of researchers and stakeholders.
Each of these dimensions of context has widespread ramifications, affecting:
- how a problem manifests and is understood, including what is studied and how, as well as the unknowns considered to be pertinent
- what actions are likely to be considered and to have a chance of being successful.
Interestingly, there is no field of enquiry that sets out to deal with context in a comprehensive way.
In addition, context can never be perfectly or completely understood. Not only is there just too much to take into account, but understandings of context and their importance will also often be contested. That is one important reason why I argue that imperfection is inevitable when tackling complex problems.
Nevertheless, those tackling complex societal and environmental problems require expertise in addressing context, both “know-that” to understand the different dimensions of context and “know-how” to deal with those dimensions of context effectively. The first of these is dealt with here.
Social or big picture context
Social or big picture context takes in:
- the history of the problem, including how it has been understood, researched and acted on, as well as key events and the evolution of the problem and actions on it
- the politics of the problem, including legislation and other policy actions, recognition in political party platforms; political enquires; interest group and lobby group positions; and sources of power, including authorisation or legitimacy, for recognising and acting on the problem
- the economics of the problem, including what the problem costs society and specific social groups, as well as costs of options for addressing the problem and cost-benefit assessments
- the cultural aspects of the problem, including effects on particular societies or groups, plus different cultural perceptions and actions on the problem
- the geography of the problem, including the localities and environments it occurs in, such as local, national, global and sometimes even extra-terrestrial levels, as well as place-based constraints on action.
This list does not aim to be exhaustive, instead seeking to make the point that there are multiple social or big-picture circumstances that come in play when tackling complex societal and environmental problems. Two other key points are that:
- these social or big-picture circumstances interact, for example the history of the problem is often an account of the politics, economics, cultural aspects, geography and more
- while assessment of what is known and done (or attempted) is important, equally important, but harder to assess, is what is not known, ignored and undone.
Organisational context
Organisational context concerns the research and stakeholder institutions involved in developing more comprehensive knowledge, as well as in supporting policy and practice change, in addition to the specific government, business and civil society organisations involved in enacting change.
It is useful to tease apart organisational context into structure and culture.
Structure includes composition, hierarchy, roles, responsibilities, connections, funding, physical layout and available resources. If we consider research organisations, the different kinds of disciplinary experts employed, for example, will probably influence how a problem is tackled. Similarly, how well knowledge brokering occurs is likely to be affected by whether there are designated roles for this activity. As a final example, physical layout, including proximity of individual working spaces, natural gathering places, meeting rooms and opportunities for privacy, are all likely to have an impact on, for instance, opportunities to exchange ideas and work collaboratively.
Culture, particularly the customary ways of doing things in organisations, is also impactful. For example, decision makers may favour particular research groups or may eschew research input; community groups may have clear protocols for working with researchers or may be hard to contact and get involved; research organisations may have favoured epistemological approaches. Culture in research organisations also includes expectations about participation in research group meetings, seminars and social events; whether participation in public debate or expression of controversial views are encouraged; and how power plays out.
Individual context
The positionality of individual researchers and stakeholders can also be usefully seen as an aspect of context. Key aspects of positionality include the person’s relationship to the problem, to those involved in understanding and looking for actions to deal with the problem, and to the institutions and individuals in a position to influence the problem.
It is useful to ask, for example, how individual positionality affects:
- the approach to the research or potential actions
- conflicts of interest
- useful perspectives and blind spots
- expectations of those tasked with addressing the problem.
Concluding questions
Do these three dimensions of context look useful for your work? Do they miss critical areas? Do you have additional perspectives to share?
Use of Artificial Intelligence (AI) Statement: Artificial intelligence was not used in the development of this i2Insights contribution or the work on which the contribution is based. (For i2Insights policy on artificial intelligence please see https://i2insights.org/contributing-to-i2insights/guidelines-for-authors/#artificial-intelligence.)
Biography: Gabriele Bammer PhD is Professor of Integration and Implementation Sciences (i2S) at the National Centre for Epidemiology and Population Health at The Australian National University in Canberra. i2S provides theory and methods for tackling complex societal and environmental problems, especially for developing a more comprehensive understanding in order to generate fresh insights and ideas for action, supporting improved policy and practice responses by government, business and civil society, and effective interactions between disciplinary and stakeholder experts. She is the inaugural President of the Global Alliance for Inter- and Transdisciplinarity.
Thank you, Gabriele – your clear consideration of context is useful. With regard to individual positioning in terms of useful perspectives as well as potential blind spots, I would suggest two stages sitting underneath that single heading: a gatekeeping stage, where positioning determines the literature a researcher is exposed to in the first place, and a subsequent processing stage, where positioning shapes how material already accessed through citations gets read and engaged with.
This is a pattern I suspect is more common than it appears. While literature reviews offer a particularly clear example, these patterns likely operate across other research practices, in research design, collaboration, supervision, and peer review. I have seen it in my own field and also in transdisciplinary synthesis: a piece of literature in a particular tradition cites an earlier piece of literature from a different tradition, usually more comprehensive and integrative, but engages with it for only a limited purpose, not taking into account the full argument being made. The earlier piece of literature is cited multiple times, but from a disciplinary perspective it engages with the argument selectively, at a narrower level than that at which it was originally formulated. Sources from outside an immediate tradition can clear the gatekeeping stage, since they are found and cited, but it is at the processing stage that a researcher’s own disciplinary training shapes what can be drawn from them. These patterns appear to involve not only individual positionality but also organisational incentive structures.
This example looks more like selective engagement than “not known, ignored and undone”, because the material exists and is used, just not fully. Bob Williams’ point that deciding the boundary between context and what is not considered context is one of the most important decisions taken in any intervention seems to apply here at a citation level too. The question of how much of a paper is engaged with, and how much is treated as irrelevant, is a small version of the same problem.
This seems especially relevant for transdisciplinary literature reviews, since the gatekeeping and processing problems feel inherent to such reviews. Reviewers necessarily work with material coming from disciplines outside their own, and the opportunity for deeper engagement feels particularly important since transdisciplinary literature reviews are often a reader’s only means of getting acquainted with another literature tradition. This gap may not be noticed even by peer reviewers, as they usually come from different traditions and tend to scrutinise citations from their own tradition most closely.
The two stages likely call for different responses. Addressing blind spots would mostly involve broadening exposure to sources, which works partly at the individual level but also requires organisational shifts, such as making time for reading across fields, valuing transdisciplinary breadth in hiring and promotion, creating cross-disciplinary spaces where different traditions can be encountered. Addressing selective engagement offers the valuable opportunity to develop the capacity to interpret material according to an unfamiliar paradigm. Selective engagement is also sometimes simply appropriate, as part of necessary scoping, and I have not yet found a clean criterion for telling productive and unproductive instances apart, beyond reflection in hindsight. Perhaps there is at least a partial, forward-looking signal worth exploring – whether the source being drawn on narrowly is one whose fuller argument might enrich or complicate the citing researcher’s own framing, if engaged with more closely.
In know-that terms, this would mean seeing gatekeeping and processing as operating within the existing heading of positionality and blind spots, while recognising that these individual-level processes are nested within organisational structures and broader social contexts that enable or constrain them. If this distinction holds, it may point to three different kinds of know-how that could be worth developing further, such as broadening exposure at the individual level for blind spots, shifting organisational incentives and cultures to support deeper engagement across disciplinary lines and attending to the social and historical patterns that have made certain traditions more visible and resourced than others. I would be curious whether this distinction might be useful for transdisciplinary synthesis work.
A useful break down Gabriele. I often find myself thinking about how the phenomena and scale of interest for analysis is somewhat disciplinary – so fields like psychology and behavioural economics resolve a lot of detail in internal and immediately external dynamics to the individual, but tend to relegate everything else as “context”, whereas sociologists or political scientists, depending on their area of interest, might do the same or instead emphasise broader phenomena and scales of analysis, and lose detail on the individual and group level elements. And all might neglect biophysical, or economic phenomena and scales of analysis without interdisciplinary integration. From this point of view, focusing on a relevant boundary object, like “place”, or “behaviour” and “behaviour change” are quite helpful, encouraging a more open ended approach. We engaged with many of these issues in proposing a socio-ecological behavioural framework in this publication:
https://link.springer.com/article/10.1007/s42532-025-00227-y
Thanks Stefan. Extrapolating a little from what you said, in a disciplinary setting, it can be useful for everything that’s not discipline-specific to be treated as context. That way it at least gets some consideration. If it is banished as not being relevant, it just gets ignored. The idea of using a boundary object to bring more contextual considerations into play is an interesting one.
One of the serious issues about ‘context’ is that it is like pinning a tail on a donkey blindfolded. Everyone seems to have a different understanding of what it is, let alone how it affects .. um well everything that’s not context. So I respond to Gabriele’s comment that there isn’t any field of enquiry that sets out to deal with context in a comprehensive way in two ways. One is that – metaphorically speaking only – you are trying to research in a fog. The other is that there are plenty of fields that have a decent stab at it. I’m familiar with the strategy, organisational development, evaluation and systems discourses and each have found ways of managing context from their own particular field’s perspectives. The lesson I take from the systems field is that exploring where the boundary lies between context and whatever is not considered context is one of the most important decisions taken any intervention. As Gabriele points out, there is so much stuff out there, that in order to do anything you have to ignore nearly everything. Where that line is drawn is critical and so significant are the consequences that it needs to be carefully deliberated upon.
Many thanks, Bob. In the development of Integration and Implementation Sciences (i2S) it has become clear that there are some key areas where individual fields of enquiry have a perspective, but it’s no-one’s job to try to bring them together. Unknowns and change are two others and I’ve been involved in getting different fields to talk to each other (references below for anyone interested). Thanks for identifying some of the fields “that have a decent stab at” context. The systems lesson that you point to is a thought-provoking and important one. I’m not sure about the fog analogy, although it works in terms of not being able to see clearly – for me it’s more of a haze than a dense fog. Is that what you mean?
References:
Bammer, G. and M. Smithson (eds) 2008 Uncertainty and Risk: Multidisciplinary Perspectives, London: Earthscan.
Bammer, G., (ed) 2015 Change! Combining analytic approaches with street wisdom. ANU Press, http://dx.doi.org/10.22459/CCAASW.07.2015 (open access).
I think I’ll stick with fog. Take the concept of environment for instance. In some fields the words context and environment are treated as synonyms (in English). In the field I was originally trained in, ecology, it means pretty much everything. In some fields it depends which bit of the field you are operating in. For instance, in the soft systems domain ‘environment’ means everything that the system cannot influence or control but which influences or controls everything the system does, while in viable systems domain, an intervention’s environment can be considered very much part of the system. So to a large extent, at least in the assortment of fields I connect with, context tends to mean whatever you want it to mean or however that particular field understands the notion. It makes trying to unify the idea of context somewhat challenging. Having said that I do like your meta, meso and micro distinctions.
Fair enough! Those are really useful examples – thanks.
Gabriele, the concept of “context” has been at the forefront of my mind over the past 20 years. Your three-layer (i.e., three level) framework provides a starting point for using “context” as an explanatory power. I would add a second intersecting dimension to each layer as a complete-as-possible or realistic set of factors that can occur within the three layers: behavioral, cultural, social, technical, environmental, religious, ethical, political, and so on. I would also add a third dimensions that intersects the other two: the extent of complexity, interdependencies, and causal interactions between layers and factors. Understanding context, for me, is like using the Rubik’s Cube to identify the characteristics of the context in question. It’s a three-dimensional cube. I have come up with term “outer boundary of context”—which is a tentative label—to identify the best description of the characterization of “context” in the sense of the three-dimensional cube. It’s a way to have a discussion about to what extent one’s situational awareness of what is going on varies between skimming the surface to fully analyzing what is actually happening.
Thanks – I agree that intersections are important and the different kinds of intersections that you mention (extent of complexity, interdependencies, and causal interactions) are, of course, really useful to think about. The first set you mention (behavioral, cultural, social, technical, environmental, religious, ethical, political) largely overlaps with what I think of as big picture context, so I don’t really understand how they would work as an intersecting dimension in the framework I suggest. In any case, as you say, there are lots of moving parts to take into account and the Rubik’s cube analogy is useful for that. Where the Rubik’s cube analogy falls down for me is that it can be solved, whereas I would argue that an understanding of context can only ever be incomplete and imperfect. That doesn’t downplay the important of doing the best possible analysis of context.
I don’t see layers (levels of analysis) as only operating within any layer. What happens in one layer affects other layers. So there is a “verticalness” to characteristics between layers affecting other layers. I have seen this in my research. As to the analogy of the Rubiks’ Cube, right, its limitation as an analogy is exactly what you said: Rubik’s Cube implies a solution. I was using the Cube analogy to more easily represent multiple dimensions interacting in a location. Setting aside these two minor issues, the idea of context is very powerful and I continue to be amazed at how often the meaning is never explained, only assumed. That’s why I was excited by your post.
Thanks for those clarifications – indeed we are in furious agreement. Interestingly, “context” is the topic that has had the least contributions in i2Insights of all the 11 elements of Integration and Implementation Sciences (https://i2insights.org/2024/04/16/i2s-updated-framework/), with 28 contributions over 10 years compared with 160 for “stakeholder engagement” and 141 for “teamwork.” To see what else has been written on context, scroll through: https://i2insights.org/category/main-topics/context/. And contributions are always welcome.
This is interesting. In another perspective, in evaluating the quality of development research, Research Quality Plus (https://idrc-crdi.ca/en/rqplus), and Research Quality Plus for Co-Production (https://link.springer.com/article/10.1186/s12961-023-00990-y), both take context into account as part of understanding the quality that is produced. As you note, the dimensions of context can differ so there is some fluidity in how it gets defined according to the areas and priorities of research.
In most cases, the dimensions include (there is more depth and detail to each, so just a summary here):
– Maturity of the research field: are we looking at an issue within an established field that may not have been addressed in a specific context, or some dimension of the field is poorly explored to date? Or are we working in a new field, building a field of research and learning through development research how it affects a region or country – a good example would be the work IDRC [International Development Research Centre] has done on AI well before the field was mainstream to work with researchers in low- and middle-income countries to understand the impacts it can have on their countries and where influence needs to brought to bear
– Data environment: is it relatively strong or relatively weak, or unavailable to researchers in universities, NGOs and think tanks? How reliable is the data?
– Political environment: what is the impact the political system has on the research? Is it a supportive environment or one with little to no interest in having the issues explored?
– Organizational research environment: I think you have covered that one well – how strong are the research centres in supporting the research and bringing the technical skills to bear? What inhibits the organizations from operating effectively?
– Capacity for research (in some cases, capacity for partnered research): this reflects the individual dimension but also looks at the priority on capacity strengthening in the research itself, through the engagement of PhD students, through research skills development, but also at research administration skills, and the skills for facilitation and networking.
An area we regretted not including in a recent assessment was a focus on environmental context: what environmental factors have affected the research such as cyclones in the research zone, drought forcing migration, among other factors.
Great – thanks! I wonder whether the elements you describe are additional or would fit in the framework I describe. Environmental context would fit under “big picture” context, as would political environment. I would probably include maturity of the research field, data environment and capacity for research in an expanded organisational context. In any case, they are all valuable aspects of context to be taken into account.