In this article, Miriam Edelman considers the dynamics of group decision-making, and the ways in which traditional approaches can fall short. She introduces a tornado model of decision-making, and its application in various case studies.
- Group decision-making involves complex processes and potential pathologies, and frequently creates conflict, disappointment and inefficiency.
- Rationalising group-decision by using various tools can help, but often doesn’t allow for the complexity of human decision-making, which isn’t always rational.
- The tornado metaphor encompasses elements of chaos in the decision-making process to ensure that the group has an understanding of shared values, hopes and priorities.
Consider the following three cases based on true stories. Please note that the details have been changed.
- An investment fund business wants to introduce a new product. Their service provider and a consultant design a new operating model, but the management team won’t approve it. The project is put on hold. A few months later, thinking they have the answer, they call the consultant back, but again a new model is rejected. This cycle repeats twice more.
- A religious congregation needs a new building. They commit to honouring their values: ethical building standards, environmental responsibility, full access for the differently abled, teaching/learning areas, an inspirational worship space, and a design that encourages community. As the design develops, however, budget constraints arise. Several people feel the values they care about are being short-changed or abandoned. The steering committee tries to make compromises but feels attacked.
- People at a university want to develop a new facility. They invite representatives from across the university to a meeting in a similar space. Some think the meeting’s purpose is to make decisions about the project; they feel frustrated that nothing seems to be decided. Others think they should explore ideas as input into designing the facility; they’re irritated when conversations devolve into budgets and scheduling. Still others wanted to experience what the facility would be like, but feel that didn’t happen. The project languishes.
Typical bureaucratic nonsense. Nothing to be done? Not for an organisational psychologist? Each of these cases is a dysfunction in decision-making. I would suggest that all could be ameliorated with a similar prescription, through the tornado model of decision-making I developed and use regularly in my practice.
Decision-making is a particular interest for organisational psychologists, with the complex processes and potential pathologies involving decision-makers, implementers and other stakeholders. It frequently creates conflict, disappointment and inefficiency.
Traditional approaches, particularly from the 1970s through the 1990s, sought to rationalise group decision-making, and utilise tools and technology in decision-support systems to improve outcomes. Based on multi-attribute utility theory, tools helped groups to identify their criteria (or ‘preferences’), assign weighting to each requirement, and reduce the decision to a mathematical formula. Where a sensitive facilitator remembers that the system is a tool, not the only answer, these tools can be very helpful.
For instance, a family is deciding where to spend their holiday. They want acceptable cost, limited travel time, a pool and good food, and assign weights to each requirement to reflect their priorities. Now they rate each option: the villa in Spain gets 3 out of 5 for cost, 5 out of 5 for food. Because cost was the most important factor, the rating is multiplied by its weight of 5, for 3 x 5 = 15 points. Food only has a weight of 2, so that adds only 10 points. When all is totalled, the model reveals the mathematically best answer: a self-catering flat in Blackpool. But rather than feeling triumphant and excited, the family feels disappointed, frustrated and flat. They check their spreadsheet; the calculations are correct.
The problem is that decision models, whether simple or technologically sophisticated, easily run into the same problems as our cases at the beginning, because the human processes involved are more complex than formulas. As Daniel Kahneman and the late Amos Twersky, among others, have shown us, our thinking processes individually, and certainly with the additional complexities of a group, are biased, flawed and not perfectly rational. Bemoaning human nature isn’t particularly helpful, though.
The tornado model
The problem, in all of these cases, is that the groups moved too fast into evaluating options or even solving the problem, without exploring underlying forces that have now tripped them up.
The tornado model of decision-making (based on work developed by Dr Patrick Humphreys, formerly of the London School of Economics) describes group decision-making in five levels, as shown in the graphic.
© Miriam Edelman
- The very top, and widest, level (number five) is critical to set boundaries, by exploring fantasy scenarios and dreams, without formalisation or structure. Here, the group considers their values and priorities, conflicts and constraints, fears and hopes.
- Level four is deciding how to decide. Here, problem-structuring language defines the problem and sets frames for proceeding.
- Level three allows structure building, such as identifying the relevant attributes to be used as criteria, or creating an influence diagram for probabilities.
- Level two is for sensitivity analysis, such as testing options against the criteria.
- Once the model is complete and solid, level one, the point of the tornado, allows for making best assessments – the actual decision.
Five levels of decreasing scope seems like a linear progression – 5, 4, 3, 2, 1, impact! That rarely happens. We usually need to move around through the levels, in multiple passes. Some have proposed a circular model, rotating between options, values, and processes. But rather than circularity, with connotations of fruitless repetition, a more appropriate description is a spiral. Imagine the levels as a three-dimensional cone, with spiralling cycles moving helically through in any direction: the tornado model of problem-solving and choice.
The tornado metaphor encompasses elements of chaos in the decision-making process. A tornado forms from disturbances in temperature to create a spiral rotation descending from the clouds as it develops into a funnel to the ground. Tornadoes must reach a critical stage of development before the actual funnel (our lower levels of the process) appears; in many cases, development ceases before the funnel stage is reached. Once the funnel does form, air moves helically throughout the cone with powerful internal momentum, with the point inflicting active damage. The spiralling movement in many possible directions in the funnel extends indefinitely, without a centre or pre-defined exit. Without taking the tornado metaphor too far – for, after all, most decision-making is at least benign if not beneficial, rather than violently destructive – the tornado image emphasises the non-linear, iterative, multi-level thinking processes involved.
The question for the psychologist then becomes how best to assist groups who choose to enter the tornado to navigate its paths.
Diagnosis and intervention
The tornado model helps us to understand that in all of our examples, the group has proceeded into lower levels of the funnel before fully completing the necessary work at the top, level five. Unless a shared ‘small world’ has been discussed, so the group has an understanding of shared values, hopes and priorities, the process is likely to founder. At level five, the group makes explicit the unspoken expectations and understandings individual participants assume everyone shares – but often that is not the case at all.
Our vacationing family had other priorities that were less ‘rational’, like glamour, excitement and fun, which hadn’t been considered in their mix. Maybe some members were more concerned about certain values, like cost, than others, and averaging everyone’s input disguised these different priorities. I’d suggest expanding the list of criteria by having imaginative conversations about how they want to feel during their time away. Then, with a longer list, have everyone assign their own weights and ratings – to do the analysis independently, before then discussing their results together.
The university story doesn’t have a happy ending, as the project continued to languish. Some efforts arose in a different form, but the system-wide acceptance for a large facility never happened. The core instigators might have achieved a better result by staying at level five a little longer. They were unclear about what they wanted to achieve, and should have built their own shared understandings before reaching out to wider stakeholders, who should also have stayed in the dreams and fears space instead of moving straight to practicalities.
The congregation is a special case because nearly everyone involved is a volunteer. The organisational psychology literature is, understandably, heavily focused on the world of paid employment. Much does apply to volunteerism, but with caveats or adjustments. The steering group needs to return to level five, with transparency, to prioritise between their potentially conflicting values, then to level four to agree how they will decide between them. While some may still be disappointed in certain outcomes, the congregation will have clarity about their process.
Luckily for the investment manager, a new consultant was appointed who was an organisational psychologist (me), who recognised a tornado failure. We stopped the project, returned all the stakeholders to the room, and identified priorities and values. It turns out that assumptions had been made on behalf of a team key to the process, and that these assumptions were completely backwards. With that clarified, a new operating model was easily approved, and implemented within a month.
Groups often feel powerful pressures – cultural, institutional, social, or other – to move down into more concrete levels of the tornado funnel as fast as possible. We want to get on with it, get to the decision. What we’ve seen, though, is that this counterintuitively wastes time, forcing us back to the less comfortable exploration of level five, before we can truly ride the tornado – and decide.