The Delphi technique is a quantitative option to generate group consensus through an iterative process of answering questions.
After each round, the responses are summarised and redistributed for discussion in the next round. A consensus is reached through a process of convergence involving the identification of common trends and inspection of outliers.
In its original form, question rounds are administered in writing, for instance, distributed by email. The technique has been adapted for use in groups face to face with the heart of the process remaining intact, allowing individuals time to reflect and an equal opportunity to contribute, ‘using disagreement as a trigger for deeper analysis’ (Dick 2000). For instructions on facilitating a face-to-face Delphi exercise, see Bob Dick’s (2000) detailed reflection.
The Delphi technique originated in the forecast of future trends. Its unique contribution is the ‘boiling down’ of differing expert opinions or other stakeholders into consensus for decision making – without creating confrontation or allowing strong individuals to dominate the process (as often happens in face-to-face discussions).
This technique can be used in various evaluation tasks to predict future data. For instance, assessing the ‘problem’ the program aims to solve and forecasting future changes. This may support an evaluation in the ‘describe’ stage, help with the setting of indicators, and with the development of monitoring plans or journals. Alternatively, the Delphi technique can also be applied in retrospect if monitoring and evaluation data is lacking and an evaluator seeks to gather it from stakeholders. Due to its quantitative, expert-based nature, it is especially useful in judgment and forecasting of situations in which pure model-based statistical options are not practical.
Delphi, a dialectic consensus method, seeks to incorporate opposing views into a new solution in contrast to other consensus techniques, through which ‘decisions are most often taken by an arbitrator or leader or by majority vote. To oversimplify the issue somewhat, one of the expressed views prevails’. In contrast, Delphi is useful in complex situations where there is little overlap between opposing stances and negotiation, or the ‘winning’ of one side (by way of the majority of power) would grossly displease another (Dick 2000).
Advice for choosing this method
- This method is useful for soliciting advice from geographically dispersed, time-poor experts.
- It allows for ‘aggregation’ of quantitative data in the form of differing opinions.
- It is time-intensive. The process usually includes three or four rounds, each of which must allow experts time to comment. It is therefore likely to take weeks.
Advice for using this method
- Delphi is a useful complement to a range of other evaluation measures. Expert consensus produced by Delphi can, for example, inform the monitoring journals or the design of surveys for a larger audience.
Dick, B. (2000). Delphi face to face [On line]. Resource Papers in Action Research Available at http://www.aral.com.au/resources/delphi.html
'Delphi study' is referenced in:
- Communication for Development (C4D) :
- Rainbow Framework :
- Rainbow Framework :