In the above case, how did Rand use Delphi technology to make a prediction?

In the above case, RAND Company designed the procedure of Delphi consultation and the choice of interviewees. In the choice of respondents, the list is carefully selected from various departments such as industry, government, universities, research institutes and business circles, and the professional composition of respondents is reasonable. Each of them received an invitation letter to participate in Delphi forecast, a postcard to be sent back to the management team indicating whether they would like to participate and the first round of inquiry form, all of which strictly followed the anonymity principle of Delphi method. At the same time, three rounds of repetitive procedures are adopted and the principle of cyclic repetition is followed. The design of the inquiry sheet was made by the management team after listening to the suggestions of experts in titanium alloy industry.

In the first round of questionnaire survey, RAND explained the requirements for respondents, provided background materials and three questions. At the same time, two professionals and some temporary assistants who do secretarial work are hired to collect and sort out the answers of each round, calculate the quartile interval and collect the comments given by the comprehensive image paper.

In the second round of questionnaires, the answers of the first round are published, and new questions are given to grade the answers of the first round. At the same time, the reasons of the first round of prediction are scored or supplemented, and the answers of the second round are summarized to investigate the changes and convergence between the second round of answers and the previous answers.

The third round of questionnaires explains the distribution and changes of the answers in the second round, supplementing the opinions provided by experts, and all the opinions with great differences require experts to re-evaluate. Finally, the third round of responses is summarized, and the changes and coordination of responses are investigated. In these three rounds of programs, some statistical methods are used in each round, and finally the prediction results are sorted out.

In the comparison between Delphi method and expert group research, it can be seen that Delphi method provides an opportunity to point out the doubts and differences related to the subject, and at the same time, it does not lose the accuracy and consistency of expert group prediction, which is convenient for investigating abandoned ideas and is an effective intuitive prediction method.