DELPHI

An experiment was conducted comparing the performance of structured face-to-face discussion groups and the anonymous questionnaire technique. The experiment was guided by two presumptions 1 that in a face-to-face situation, information transfer is likely to be much greater than in the anonymous controlled communication situation and 2 that the effect of undesirable social interactions could be meliorated by imposing a specific format for the discussions. The presumption to be tested was that a structured conference of this sort would produce more accurate estimates than the questionnaire technique. The major outcome of the experiment was that the presumption that the structured discussion would turn in a better performance was not born out in fact, the questionnaire responses were, if anything, somewhat more accurate than the structured conference responses.