The seven ideas of the Delphi method: it takes a village

The Delphi method

The Delphi method uses a guided process of interaction to gather, broaden and deepen the knowledge and understanding of a group of experts, the panel, about the phenomenon under study, which often lies in the future. The features of the phenomenon are peeled back layer by layer and round by round. Unlike in ordinary surveys, participants are led into interaction with one another in a way that can be called shared knowledge-building or collaborative learning.

The facilitator's role in the process is proactive, whether the aim is agreement (consensus) or a plurality of views (dissensus). In the latter, the study explores different possible, probable and desirable alternative futures. In both, what matters are the discussions of anonymous panelists and the reasons they give for their own choices.

The Delphi method was developed at the RAND Corporation in the United States in the 1950s. The aim was a tool for situations that call for decision-making that is fast but considered. Such situations arise especially in crises, where there is a risk that social hierarchies narrow down decision-making. The method borrowed its name from ancient Delphi, whose temple oracle foretold the future for centuries. In today's Delphi, the oracle's tasks are shared by three: the facilitator, who plans and guides the study; the panel, a many-voiced group; and the panelist, who is a member of the panel, an expert, an assessor and a participant in discussion.

The Delphi method treats expert knowledge in a particular way. According to Fred Woudenberg's analysis, the different forms of the method share three features:

• anonymity: instead of persons, their assessments and arguments do the debating
• iteration: several rounds, or continuous interactive commenting within a limited period
• feedback: positions are refined on the basis of information and feedback from other panelists.

What decides the success of a Delphi study is the choice of the right panel and how the facilitator catalyses the panel's work through their own actions.

The Delphi method owes much to many ideas, and they are not easy to put in order. Fortunately, there is no need to. The Delphi method is a hybrid tool whose different forms emphasise different aims and techniques. What unites all Delphi studies, however, is an interest in knowledge and a sense of wonder, which is put in the form of a question. A question signals a starting point of not knowing, the opposite of knowing everything. A question starts learning, which an answer in some sense ends. Seen in this way, the Delphi method is at bottom a technique of learning.

If no questions arise in the mind, it cannot produce answers either. A question opens up active thinking in a way that a ready answer does not necessarily do. Not everything can be known, but anything can be asked, and asking leads us to consider, for instance, what could be otherwise and better than now. Delphi processes begin with an open question or claim that cannot be settled by knowledge alone, because there are several possible answers. The task of the Delphi method is to identify alternatives and to start reasoned reflection on their probability and desirability.

It takes a village

The series grew out of the idea of adapting the Delphi method for schools and school learning. I identified seven ideas, and their relationship is pictured as a village in which each idea lives in a house of its own but deals with all its neighbours. The analogy was sparked by the familiar saying that it takes a village to raise a child.

When one neighbour puts forward a hopeful claim about the village's development, a second sets about developing it and a third comes up with more reasons why the vision can be realised. A fourth has the ability to carry out the project, as long as a fifth manages to bring along those who hesitated at first. In a good village the future is made, and suitable tools are found for it. The village steps into a virtuous circle in which its social and intellectual capital grows faster than any inflation.

The meme could serve as a parallel term for the idea. Richard Dawkins introduced it in The Selfish Gene (1976). He formed the word from the Greek for imitation (mimeme) and shortened it to rhyme with gene. The meme is a broader concept than the idea: besides ideas, it covers styles, patterns of behaviour, knowledge and other cultural phenomena that spread from person to person. According to Dawkins, genes spread biologically and are selected by natural selection, and in the same way memes spread culturally and are selected by how well they replicate and spread. On social media the meme has become carnivalised, but in Dawkins's sense the village of memes is an apt analogy.

The analogy invites us to think that the seven ideas of the Delphi method must be taken to school as a whole village at once. Separated from one another, the ideas do not gain strength. Together they produce renewable energy that achieves things no one previously knew how to dream of. At school it is permitted to build understanding of things that seem distant or useless for one's own life. Such distant skills include respectful conversation with people unlike oneself and the ability to build an active relationship with the democratic system that both regulates and enables our actions. A good village practises wonder, because great learning begins with the thought that something could be otherwise.

In a bad village, spirals and haste lead downwards. Wonder turns into puzzlement and gradually into condemnation. Dialogues shrink into loud debates until trust in the neighbours collapses and talk stops altogether. New conspiracy memes simmer in the bubbles. Making the future is forgotten, and only high fences are built in the village. Before long, people move away and the village is deserted.

Let us return to the good village. In a good school, a Socratic question, such as a claim about the future, starts not only wonder and thinking but also a dialogue, which in the Delphi method is guided into open speech and discussion. Open speech is democracy's characteristic way of practising a politics of reasons, in which the best argument is assumed to win. Different perspectives and interests are taken into account when the panel brings together the people who have relevant expertise and a stake in the matter. Most Delphi studies examine the future, which makes the method a central method of futures studies.

Not everything can be known, but anything can be asked

Anything can be asked and claimed, but the world holds countless phenomena that cannot simply be known or proclaimed true. Actions and aims rarely have only good or only bad sides; the two are inseparably mixed. What is useful to some can be harmful to others. Economic success may be achieved at the expense of the environment, and a decision that looks positive now may prove harmful in the future. Even when there is agreement about a fact, interpretations of its causes and consequences vary.

A basis in knowledge is the starting point of understanding, but knowledge alone is not enough to navigate the world sustainably. In the Delphi method, the choice of panelists ensures that existing knowledge is available when an open question is explored. Knowledge is anchored in a view of how the world works, a worldview and an ideology, but also in the horizon in which one's own and others' benefit and good are assessed, that is, interest. The plurality and opposition of interests produce many voices, which are given room especially in policy Delphi studies that aim at dissensus.

Facts can be established beyond dispute, but the causes and motives behind facts cannot be treated as facts. Our perspective determines what interpretations we make, and over time the interpretations change too. Democratic communities are marked by a plurality of views and reasons. Interpretations and stories are built from real and imagined materials, and through them we come to understand what has happened and why. A fact says nothing certain about the forces behind a phenomenon or the motives of actors, but a story does. The future scripts, or scenarios, built with the Delphi method are such stories, as are political visions.

The future as an object of knowing and making

The future is an object of knowledge of its own, because there can be no certain knowledge of it: it has not yet happened. Yet the future is always present in human life as hopes, fears, plans and goals. The further ahead we look, the more varied the possible future can become. From the point of view of futures studies, it makes no sense to limit knowledge of the alternatives. From the point of view of making the future, however, it is valuable to consider which futures are desirable and probable, and under what conditions.

The Delphi method practises wonder through the questioning method that Plato made famous in his Socratic dialogues. A well-formulated, tricky question inspires a search for answers. Often the answer remains incomplete and gives rise to new questions. Even if the answer remains partly open, the layers of questions and answers have produced growing understanding among the participants of possible, avoidable and desirable alternatives. A survey collects information, as the Delphi method does. The difference is that in a survey information is collected once and in one direction, whereas in the Delphi method it is processed in many directions and many times.

In 2023 I wrote that the wild card of future panels is artificial or supportive intelligence, as Osmo Kuusi calls it in connection with the Delphi method. I anticipated that panels would in some form be joined by an AI expert whose computing capacity is many times that of the rest of the panel, and which arranges information into a form that the human panel can use. The wild card has since been played. In Pronoia the AI assistant Kastalia talks with panelists before the round, and Kassotis opens up the material in the analysis phase. AI carries the water, but seeing is human work.

Argument in the service of politics

When we shift our gaze from questions to answers and solutions, our attention turns to their reasons, but also to the people and techniques used in the answers. This side of human activity too was described in lasting terms in antiquity. The three means of Aristotle's rhetoric are ethos, pathos and logos, and his enduring insight is that politics is, to the highest degree, the making of the future. Politics can hardly be practised successfully without reasonable rhetorical and argumentative skills.

In politics, decisions are made in the present, but most of them aim at the future. In the ancient city-state, people gathered to decide whether to go to war, free slaves or levy taxes, and in these decisions rhetoric played a decisive part. For the Delphi method, by far the most important of the means of rhetoric is logos, reasoning. The Delphi method emphasises good reasons and seeks to reduce the influence of the other two means. Arguments are made to compete in human interaction whose course and outcome are unknown at the outset, even though a Delphi process is guided and facilitated more actively than an ordinary survey. The weight of ethos and pathos is diluted above all by anonymous participation.

From interaction to learning together

The Socratic method requires a capacity for interaction, which in an open society appears as open speech. Jürgen Habermas made its social character and significance well known. The open society, in turn, is Karl Popper's concept of a community in which those in power can be criticised and institutions reformed gradually without violence. The Delphi technique makes possible open speech, participation and interaction in an iterative process that resembles Socrates' dialogues with his students. A question sets off commenting and discussion, which produces not only alternative answers but also new questions.

A Delphi panel brings people together on the basis of their merits. The aim is a group in which all the expertise and interests related to the phenomenon under study are represented. In a discussing panel the group is kept small enough that its size does not prevent dialogues from arising. The more interaction the process contains, the better the arguments put forward for different alternatives are weighed.

The amount of dialogue and the changes in panelists' views show that the panel is moving from collecting knowledge to creating it. I call such a panel a learning panel. It can arrive at views and solutions that none of the participants held when the panel began. Its capacity to learn arises partly from bringing together people who would otherwise communicate little with one another. When people approaching a phenomenon from different directions give themselves over to interaction, collaborative learning arises.

John Dewey saw the school as a rehearsal stage for social and communal learning. Paulo Freire took his social analysis further in a more radical direction, in a way that suits the study of great upheavals whose resolution will fall to the generations attending school at any given time. The Delphi method suits a school attuned in a Deweyan and Freirean spirit, and it offers a chance to practise many key skills. These include social and interaction skills, logical and rational decision-making, and the ability to communicate and to form a systemic overall picture of phenomena, the environment and society. Key skills also include futures consciousness, which Sanna Ahvenharju has structured in her doctoral research into five dimensions: a time perspective that extends beyond the present, agency, openness to alternatives, systems perception and concern for others.

Working together and creating knowledge together are on the increase. They have already turned working life towards teams and even democratised it, and they are gradually influencing education as well. More and more often it is realised that the ultimate learner is not the individual but a team, a community, an organisation or even a whole nation. Alongside learning between brains, another upheaval is under way, in which learning and knowledge creation take place entirely outside human brains. The information flow outside the human body, the exosomatic flow, is currently growing fastest of all. It would be strange if it were left outside Delphi processes, and it has not been.

The dialects of the Delphi method

In connection with the Aavistus studies on the future of the Delphi method, Linturi and Vepsäläinen (2021) mapped the purposes and implementations of the method and drew up a table of the variations in its current use (Delphi variations 2030). Within the bounds of its defining features, anonymity, iteration and feedback, the purposes of the method, and partly its forms of implementation too, have grown richer throughout its history. The table structures its use along five dimensions: purpose, target (consensus or dissensus), the basis for assembling the panel, the mode of facilitation and the method of analysis. The purposes group into three main motives:

• Research. Forecasting and technology foresight often aim at consensus and are analysed quantitatively, for example statistically and through trend analysis. Social foresight, foresight more broadly and barometers more often aim at dissensus and use qualitative analysis, such as argument analysis. The panel is assembled on the basis of expertise and interests.
• Organisational development. Action research, strategy work and the preparation of decisions most often aim at consensus, although strategy work also aims at dissensus. The panel is assembled on the basis of membership, status or role, and analysis uses, for example, systems analysis, CATWOE analysis or cross-tabulation.
• Social development. Debate and argumentation, increasing participation, inquiry-based learning and pedagogy, and group therapy and identity building can aim at either. The panel is open or targeted at a specific group, facilitation and participation are active, and analysis is typically qualitative, for example argument or discourse analysis.

Most Delphi studies are academic research and theses. Scientific publication databases contain thousands of books and articles, most of which have aimed at consensus. The share of dissensus panels has, however, grown in recent years, and in Finland there are already clearly more of them than of Delphi studies aiming at consensus. In such cases the Delphi process is often continued with scenario work, the most typical method linked to the Delphi method.

Besides research, the Delphi method is used to develop organisations and communities. Often this means strategy work or decision-making, which is indeed one of the method's original purposes. Much less is known about this use than about research, because few open publications describe organisations' development processes. On the borderland between research and development, technology foresight studies have been carried out, typically with large panels.

Large numbers of panelists are also brought together in participatory Delphi processes, whose use is growing. Participatory panels enrich the picture of the situation and increase cohesion and social capital, for example in municipal and regional development. Another target of social development, and at the same time of building competence, is teaching and learning, with the school as their setting. It was especially in mind when this series was first planned.

The dialects tell us what the Delphi method is used for. In Pronoia the same method is carried out in four variations, which tell us in what cadence the rounds proceed: Argument Delphi, Classic Delphi, Real-Time Delphi and Barometer Delphi. They are presented in the facilitator article The four variations.

The essays of the series

Wonder can continue in the essays, each of which focuses on one idea, or cluster of ideas, behind the Delphi method that guides its use:

• The idea of Socrates: a question opens what an answer closes
• The idea of dialogue: in conversation, something always changes
• The idea of argument: reasoned speech in every age
• The idea of open speech: deliberated democracy
• The idea of expertise: an evolution in three layers
• The idea of the future: a Bakhtinian theory of knowledge
• The idea of futures studies: visiting the future to learn what is worth doing now

The guiding ideas of the first and third essays are dug up from antiquity. The idea of Socrates describes the dialogical method, which as a process started by a question can be seen as a forerunner of the Delphi method and is still a valid technique of knowledge creation. The idea of argument borrows the threefold division of Aristotle's rhetoric and examines how people seek to influence one another. For the Delphi method, logos, word and reason, stands above the other skills of rhetoric.

The second and fourth essays look at the ways people communicate with one another. They emphasise open and dialogical forms of conversation in which participants can form their opinions and argue for them without pressure and without being pressured. Open speech is at bottom a form of communal and social learning on which lasting democratic forms of government are built.

The last three essays connect the Delphi method to social phenomena, politics broadly understood, and to the layers through which humans have learned to acquire knowledge and understanding on the basis of expertise. Throughout its history the Delphi method has been distinctly an expert method, but its development is influenced by the fact that the understanding of expertise has broadened and deepened considerably since the modern Delphi method was developed.

The essays on the future link the Delphi method to an epistemological view according to which a significant part of reality is not only to be known but also to be discussed and negotiated. The historically shaped present and the possibilities for future action influence each other in both directions. Futures are worth taking into account when decisions are made in the present, that is, when politics is practised. This gives the Delphi method, and futures studies more broadly, a practical and political quality in addition to an interest in knowledge. The future is not only studied and anticipated; it is also made every day. Our actions are guided by discussion and deliberation, but also by the notions we have of how the world works now and what it is becoming. The last essay examines the big picture of futures studies and its methods through the views of a few futurists, and finally returns to what was anticipated in 2023 and where we have ended up. Each essay ends with its sources.

Sources

• Ahvenharju, Sanna (2022). Futures Consciousness as a Human Anticipatory Capacity: Definition and Measurement. Doctoral thesis, University of Turku.
• Dawkins, Richard (1976). The Selfish Gene. Oxford University Press.
• Dewey, John (1916). Democracy and Education. Macmillan.
• Freire, Paulo (1970). Pedagogy of the Oppressed. Herder and Herder.
• Kuusi, Osmo (1999). Expertise in the Future Use of Generic Technologies. VATT Research Reports 59.
• Linturi, Hannu & Kuusi, Osmo (2022). Tulevaisuuksia ennakoiva Delfoi-menetelmä [The Delphi method for anticipating futures]. In Aalto, H.-K., Heikkilä, K., Keski-Pukkila, P., Mäki, M. & Pöllänen, M. (eds.), Tulevaisuudentutkimus tutuksi: perusteita ja menetelmiä [Introducing futures studies: foundations and methods]. University of Turku.
• Linturi, Hannu & Vepsäläinen, Jukka (2021). Delphi variations 2030. Metodix.
• Woudenberg, Fred (1991). An evaluation of Delphi. Technological Forecasting and Social Change 40(2), 131–150.

Read the essays

The introduction is followed by seven essays, one for each idea.

1. 1The idea of Socratesa question opens what an answer closes
2. 2The idea of dialoguein conversation, something always changes
3. 3The idea of argumentreasoned speech in every age
4. 4The idea of open speechdeliberated democracy
5. 5The idea of expertisean evolution in three layers
6. 6The idea of the futurea Bakhtinian theory of knowledge
7. 7The idea of futures studiesvisiting the future to learn what is worth doing now