Technology-enabled deliberative democracy - RSA

Technology-enabled deliberative democracy

Comment 1 Comments

  • Picture of Peter Miles FRSA
    Peter Miles FRSA
  • Deliberative democracy
  • Technology

We are seeing an increased recognition of the need for deliberative democracy. Peter Miles FRSA shares his experience of a technology-enabled structured dialogue methodology that actively supports a diverse group of stakeholders and experts to develop a deep and shared understanding of a complex challenge and then commit with confidence to a way forward.

While there are many group deliberation methods around, most rely on simple technology (such as the humble but ubiquitous post-it note), or apply computing in a passive mode as a way of documenting and sharing text. Almost always there is a tacit assumption that words and sentences have a single meaning.

The structured dialogue methodology introduced here has its roots in Interactive Management, developed in the 1970s by John Warfield and Alexander Christakis in the US. This approach spread around the world under a variety of names, including Structured Dialogic Design, Structured Democratic Dialog and Demosophia. It is designed to avoid ‘cognitive overload’ and reduce distortion due to power imbalances, to integrate diverse expertise and perspectives, and it enables deep mutual understanding as a solid foundation for progress. While the technology plays an active part and the process of facilitation plays an important role, the content is always owned by the participants.

Let us start with an outline of the experience of the participant. First of all, they will be in the room – known as a Colab – for a reason; the participation group will have been carefully designed at an earlier stage of the process, taking account of the overall context, and aiming above all for cognitive diversity. They may be deeply involved in the particular issue and care about any potential changes, as a stakeholder; or they may have sway over resources that will be needed for implementation. They could be an expert of some sort – possibly technical, or financial, or from working in a frontline position – that gives them a specific insight into how things really work.

Participants will be asked to answer a ‘trigger question’, initially working on their own. The question will be a broad one, designed to cover the scope of the situation, and along the lines of: ‘What are the main issues, barriers and challenges we face in achieving X?’ Everyone will then be asked to provide just one answer. That answer will be captured in a software tool, along with answers from everyone else and this process continues until a number of answers are captured. Then participants get the opportunity to clarify the answers that others have given and these clarifications are again recorded and captured, always using their own words. Further understanding emerges during a grouping phase, where answers are contrasted and compared.

The next stage is where the magic happens. The software presents a pair of answers (each one an issue, barrier or challenge), and participants are asked the question ‘Does issue A directly aggravate issue B?’ (Other forms of comparison question can be used – the aim here is to test for causality – for example, does A make B worse, or in some way stop B from being resolved?) We might expect the group to agree on most of these comparisons, but in practise they generally don’t, because they do not all have the same ideas in their head on what A means, what B means, and then whether A does indeed ‘aggravate’ B.

The good news is, they now have a real focus and time to explore and resolve that, while not being distracted about C or D or X, Y, Z (their time will come). And once they have resolved it (which can take seconds or quite a bit longer), a group decision (yes or no) is entered into the tool, which moves on to present another pairing.

The process can sometimes seem painfully slow, but step-by-step progress is always being made, differences in understanding are being cleared out of the way, and the individual mental models of the participants are steadily converging. The result (after what could typically be a very full and exhausting day) is then presented in the form of a graph – an ‘influence map’ – produced not by the facilitator but by the software tool, as a direct result of the group’s collective decisions. Individuals or sub-groups then develop a narrative appropriate to their context, based on the influence map, to explain and represent the outcome to others not present in the Colab (for example, to an important stakeholder group, such as a finance department that will need to fund the outcome).

The influence map shows where the emphasis needs to be, where the priorities need to lie, and reveals the issues that will need to be resolved to make other ‘downstream’ issues tractable. A different form of Colab can then be used to develop a plan of action to tackle what everyone has now agreed is the challenge. Where appropriate, a further form of Colab can be applied as a foundation stage to develop a shared vision, essentially co-creating an up-front structured benefits map.

So where, how and why does it work?

This approach is consistent with a contemporary understanding of brain science, and while utilising the latest information technology, is based on concepts first developed in the 1970s. It has been applied widely and successfully but remains a niche approach. My personal experience spans applications including the implementation of a new business model in a major UK engineering firm, and the development of a new model for children and young people’s services in the NHS. Many hundreds and probably thousands of applications have been notched up by practitioners worldwide.

This approach is powerful for a number of reasons, perhaps the primary one being that everyone is focused on just one pairing at a time, and that task stays within the cognitive limits of each individual’s short-term working memory. Participants are then encouraged to think more deeply, engaging both parts of our ‘fast and slow’ brains. As soon as participants see a pairing question, intuition usually kicks in, but deliberation becomes dominant as discussion takes place (and this is why people are exhausted by the end of the day).

The process also provides a ‘level playing field’ for people who think and talk in different ways, and with different levels of power and status. While there is still some advantage in enabling people to express themselves clearly and forcefully, that does not beat first-hand knowledge of the reality of the situation. Logic, knowledge and critical thinking are given the strongest hand, while waffle and obfuscation hardly stand a chance.

You might expect that participants go into the room with strong initial biases towards a particular outcome. But the process does not start with potential outcomes, it starts with the ‘atoms’ of the complex challenge being faced, and the group co-constructs a picture of the overall challenge from the multiple perspectives of the participants. Because their mental models have evolved in a step-by-step process, the resulting influence map often looks ‘obvious’ to the participants (this is a good time for the facilitator to bite their lip). The typical outcome is one that most people agree with and everyone can accept, because they have heard the arguments and respect the way the process works.

A word or two about the technology, which is used in a way that augments and complements human thinking. Firstly, in a relatively recent development, it can be used for real-time capture of the participants’ words. Then, in more familiar ways it is used to present information to the group and to allow participant voting. At the heart of the method is the algorithm that provides the pairwise comparison, called Interpretive Structural Modelling. In essence this provides an efficient way of handling the comparisons of pairs of issues, which would otherwise take a lot longer. It also acts as a memory of the complexity of the emerging matrix of relationships, and represents that complexity in a readily assimilated graphical influence map.

This methodology has a renewed relevance in the context of increasing calls for deliberative methods to strengthen our democracy. It will not be appropriate for all situations, but those with the responsibility of designing deliberative interventions should at least be aware of it as a capability, and consider applying it in situations where a deep-dive is necessary, or perhaps where other approaches have become stuck. There is also a wealth of learning behind the decades of development that may be appropriate to borrow from and apply in new ways.

I no longer design or facilitate interventions myself and to the best of my knowledge nobody else in the UK is currently doing so. There is however a global network of practitioners and academics to connect with, and the software tools are available from more than one source. The RSA is just one of many institutions calling for and working towards an increase in deliberative methods; if any staff, Fellows or general readers would like to know more I can provide sources of further information and am open to discussion.

Peter is a trustee of a local mental health charity, and otherwise mostly retired. Between 2004 and 2017 he worked in process consultancy and facilitation, helping organisations to take decisions and make plans in complex multi-stakeholder situations. Prior to that he was CEO of an international automotive electronics business. Peter has BSc. and MSc. degrees in Electronics and the CIM diploma in Marketing.


Join the discussion


Please login to post a comment or reply

Don't have an account? Click here to register.

  • How secure would the software be? I don't expect this question to be answered in detail here, because it is a security matter, but to be fully democratic the outcomes must reflect truthful opinions and not those of some hidden code (or manipulative measure).

    Thanks for the article Peter Miles FRSA

    Isabella Wesoy FRSA