By Jeff Knezovich, FHS Policy Influence and Research Uptake Manager, Institute of Development Studies
[Editor's note: This blog is part of a series of reflections emerging from a workshop on complex adaptive systems research methods held in Baltimore in June 2014.]
Many will recognise the causal loop diagram above, outlining counter insurgency dynamics in Afghanistan. The New York Times reported that when then General Stanley McChrystal, who was leading the American effort in Afghanistan at the time, saw this diagram presented he quipped: ‘When we understand that slide, we’ll have won the war [in Afghanistan]’.
And while the room was reported to have erupted in laughter around him, it begs a more serious question: is it possible to communicate complexity without being ridiculed?
Can we even communicate complexity?
While some have risen in defence of the diagram and of its creators, PA Consulting Group, others have been less kind. Alexander Galloway (2011) in his article ‘Are some things unrepresentable?’ cites it as an example of a critical tension in communication where the more information that is represented the less information is actually conveyed:
“Despite an overwhelming amount of detail, the PowerPoint slide is not easy to digest. In fact, the high level of detail seems to hinder comprehension rather than aid it. Unlike realism in painting or photography, wherein an increase in technical detail tends to bring a heightened sense of reality (at least in the traditional definition of aesthetic realism that has held sway more or less since the Renaissance), the high level of technical detail visible here overwhelms the human sensorium, attenuating our sense of reality. Rather, like a fractal whose complexity does not decrease when viewed through a magnifying glass, the information contained in McChrystal’s PowerPoint does not grow more coherent the longer one inspects it. Eschewing lucidity, the diagram withdraws from the viewer’s grasp, effectively neutering its capacity as a vehicle for information. One is left wondering what exactly McChrystal’s PowerPoint slide is meant to communicate. Is it communicating America’s military strategy in Afghanistan? Or the reverse: is it communicating how difficult such strategies are to communicate in the first place?”
He goes on to dub this inherent contradiction of communicating complexity ‘McChrystal’s Law’, and then proceeds to suggest that such visualisations contribute to a political violence committed against the viewer, in part because the aesthetics of the diagram overstate its ability to represent. Yikes!
Unfortunately, McChrystal’s Law is just the tip of the iceberg when it comes to communicating complexity. At a three-day workshop jointly convened by Future Health Systems and the STEPS Centre examining complex adaptive systems (CAS), we had an interesting discussion about some of those challenges.
Challenges in communicating complexity
Some of the key points from the facilitated discussion included challenges like:
- The specialised language of the complexity science hinders comprehension: It’s not just the ‘known knowns’ and the ‘known unknowns’ that throw lay audiences for a loop, the whole language of CAS has a glossary that is not easily understandable. And perhaps we have Malcolm Gladwell to blame – people think they understand CAS terms like ‘tipping points’ because they've skimmed one of his books, but there’s often a lot more to it than that. For those who work with causal loop diagrams, for example, they see meaning embedded in the Afghan slide that a lay viewer might not – like the double cross-hatches in some of the loops which indicates feedback delays, or even the embedded stock and flow diagram.
- Mental models are difficult to challenge: People hold on to their own models and don’t embrace understanding alternative models, especially when they challenge assumptions. CAS modelling forces people to declare assumptions too, which isn’t necessarily a comfortable process.
- Making abstract concepts tangible: Especially CAS approaches that rely on modelling, like agent-based modelling or causal loop diagrams, it can be difficult to relate models to real life scenarios and to make them tangible.
- Western narrative traditions aren’t necessarily suitable to CAS stories: It’s not true everywhere, but in many Western traditions, we’re taught that stories having clear beginnings, middles and ends and clear causal links. These stories work well when relying on Newtonian science where A leads to B, but when working with CAS it’s not the same – A might lead to B which then leads back to A and then C, or not. The message gets blurry. Perhaps CAS researchers should be exploring other types of narrative structures (like cyclical narration, which is more typical of the Indian style of writing).
- We need to be clear what we’re trying to communicate and to whom: Is it research findings? Is it methods? Is it trying to encourage others to attempt similar approaches? If it’s the former, is it really that much different from communicating any other sorts of research findings?
For my part, challenges I highlighted included:
- The need to present combinations of different types of data/information: When it comes to complexity, there’s not just one story to tell. This means combining different types of visualisations, different types of data (qualitative and quantitative) and different representations of a story (through photos or graphics, for example). This can be hindered at basic levels like computer processing power, but also in terms of collating and co-locating these different formats in a single space or platform.
- The static and linear nature of traditional publication formats (yes, academic journals, that means you!): Many of the modelling presentations that we saw during the workshop produced dynamic and interactive visualisations, for example a really interesting model of an airborne contagion spreading through Las Angeles, CA. But when you get to the article it presents one screengrab of what is effectively a minute-long video simulation. The amount of information that strips from model is unforgivable!
- Changing patterns of information consumption: The overall trend in communication activities is toward shorter and more easily digestible snippets of information. Think of the Buzzfeedification of news. Even online videos are getting shorter, with Twitter’s six-second Vines. It’s also moving off of the printed page and onto screens – where people have tended to skim rather than to read in-depth. This does not necessarily lend itself well to communicating complexity.
Tools and approaches to overcome those challenges
Clearly communicating complexity is not short on challenges. But if that’s the case, what are some of the approaches and tools that we can use to help us to, in little ways, overcome some of these challenges?
- Information layering is critical! One doesn’t have to tell the whole story all at once. Causal loop diagrams, for example, can be broken down and explained by sub-system. Although it’s not necessarily talking about a complex system, I really like the way Mapping Czech Crime conveys (but also hides) a lot of information in layers. It does this in terms of granularity, showing information at a provincial level first before allowing users to dig further into the departmental and municipal levels, but also with clicks and buttons that pull up more information.
- Tell an effective story: Be sure to frame it properly so that people know what you’re going on about. Explain pieces of the story, but use digital technologies to help skip around while still showing relationships and connections between the elements. Prezi, for example, can help zoom in and out and jump around in a way that breaks out of the linear straightjacket of PowerPoint or Keynote.
- Remember that data is not information: Just because you’re trying to communicate a complex entity does not mean that you can convey it all. One way of approaching this is through layering, but it’s also worth remembering that data don’t inherently have meaning and what you’re really trying to convey is information. Think about what that means in terms of the stories you’re trying to tell and present the most relevant information for that context.
Overall, I'm optimistic that our ability to communicate complexity will grow. That's not just because non-linear thinking is already starting to permeate research approaches and managing change. It's also because technologies that can help are developing quickly. New free and low-cost tools that can help to visualise data and models are appearing seemingly every day. We used Vensim and Netlogo to create models and casual loop diagrams during the workshop. We created social network diagrams in Google FusionTables and Gephi. But that's just the beginning. Easy-to-use programmes like Tableau Public are also available. So get out there and get communicating complexity!