Application and consensus building
There are three major characteristics that separate Delphi from other survey methods. First, the panelists involved in the survey are experts in the issue concerned. Second, the responses of the panelists are treated anonymously. In studies of sensitive issues, even the identities of the panelists are kept anonymous from each other. Third, the study is conducted in repeated survey rounds interspersed with controlled opinion feedback (Weaver, 1971; Dalkey & Helmer, 1962).
A panel of experts is selected as the respondent group, based on the areas of expertise required. They are then provided with a questionnaire on a certain subject and are required to give their opinion. Their responses are anonymous except to the moderator. The moderator collects the questionnaires, summarizes the results and then develops a new questionnaire for the respondent group. Before answering the second round of questionnaire, the panelists usually receive a summary of opinions, which they can use to re-evaluate their original answers based upon examination of the group responses. In each succeeding round of questionnaires, the range of responses by the panelists gets smaller and converges toward the "best" response through this consensus process (Taylor & Judd, 1994; Linstone & Turoff, 2002).
It is important to note that consensus, in and of itself, is not the ultimate goal of the Delphi technique. The value of the method is not merely its ability to induce consensus, but also its ability to highlight a diversity of underlying assumptions (Nolan, 1994). This would suggest that in the case of identifying criteria to assess creativity, a Delphi should not only result in an agreed upon set of criteria - but on a set that no single expert would have been able to arrive at individually.
