Delphimeetod

Delphimeetod is a structured forecasting and decision-making technique aimed at gathering expert judgments while reducing the influence of group dynamics. It was developed in the 1950s at the RAND Corporation by Olaf Helmer and Norman Dalkey to forecast technological and strategic developments. The method relies on a panel of experts who participate in multiple rounds of questionnaires with anonymized responses and feedback.

In practice, the process begins with selecting a relevant set of experts and designing an initial questionnaire. After the first round, the organizers summarize the responses and provide feedback to the participants, including the reasons behind different opinions. Participants then reconsider their views in light of the group feedback and may revise their answers in subsequent rounds. This iterative cycle continues until a predefined criterion is met, such as stabilization of opinions or attainment of consensus. Variants include the real-time Delphi, which accelerates rounds, and the policy Delphi, which examines options and policy implications without forcing a single consensus.

Applications of the Delphimeetod span technology forecasting, long-term planning, risk assessment, and priority setting in research and policy. Strengths of the method include the ability to elicit informed judgments from diverse experts, maintain anonymity to reduce bias, and progressively refine uncertainty. Limitations involve reliance on the quality and selection of experts, potential attrition across rounds, the time required to complete multiple iterations, and the possibility that consensus may reflect a false agreement rather than true accuracy.