Having appeared in 1950-60, the Delphi method has become a convenient tool for making forecasts in the field of scientific developments and their influence on the strategic principles used in waging war. Often it is called the Delphic method. In the process of its development, RAND specialists took an active part, and Norman Dalkey, Olaf Helmer and Nicholas Rescher are considered to be direct authors. When choosing a name, it was decided to use the name of the Delphic Oracle.
The Delphi method is usually classified as a method of expert evaluation. Its main features are considered anonymity, correspondence, multi-level. The idea of the method is to, using a series of actions, consistently achieve the right solution, or at least maximize consensus. Such actions can be polls, interviews, brainstorming. For the production of analysis, they operate in several stages, the processing of the results is carried out by statistical methods.
The core uniqueness that the Delphi method possesses is its basic principle of using a certain number of independent experts (the best effect is achieved when they do not even know about each other). Repeated use of the method allows us to conclude that evaluating and predicting the result of an unstructured group (collective) of individuals gives higher results. However, there is no open clash between the so-called carriers of opposing opinions. The Delphi method relieves opponents of direct contact, group influence. The survey is conducted extraterritorially; experts are not gathered in one place.
The Delphi method assumes the presence of the following subjects: first of all, this is a group of researchers. Each of this group is required to respond individually and in writing, the next is the organizational group, which should bring together the opinions of experts.
This method has properties that other methods that are part of various types and classes do not possess. Delphi, as a tool for solving complex problems, has several advantages. The main one is complete independence from the influence of the authority of the majority. There is no dispute or discussion; they are replaced by individual expert polls. Step by step, this process looks like a chain of mandatory activities.
At the preliminary stage, a group of experts is selected, as a rule, their number does not exceed 20.
The main stage includes the formulation of the problem, during which the experts get the question and begin to break it down into sub-questions. The task of the organizational group is to select the most frequently asked sub-questions; as a result, a questionnaire appears. This questionnaire is sent to experts who evaluate it for completeness of information and the possibility of adding something else. So there are 20 options that have additional answers with information and aspects, on the basis of which make up the next questionnaire.
An already improved questionnaire is re-sent to experts. Now their task is to give their original version of the solution, plus everything else, it is necessary to consider the most extreme points of view that other experts have provided. Evaluation of the problem takes place in several aspects: resource endowment, efficiency, and the degree of compliance with the statement of the original task.
The appearance of the prevailing judgment of experts, bringing their points of view together, completes the stage-cycle of solving the problem. Having familiarized the experts with the arguments, which have a strong difference, it is proposed to evaluate them with the subsequent opportunity to change their minds. The procedure is then repeated.
The number of stages reaches three, but if necessary, there may be more. After that, they proceed to the final analytical moment. During it, the consistency of the opinions of experts is checked, an analysis of the findings and the development of final recommendations are made.