地域環境評価のための集合調査法の有効性に関する実証研究

土浦市内の道路周辺地域住民による会議実験

抄録

The paper reports on the application of real time Delphi method to group social survey in the evaluation of residential environment -roadside areas with heavy traffic as a case.
The performance of social survey methods is appraised with respect to those characteristics as representativity of samples, identifiability of the sampled people, stability of responses, accuracy in communicating the meaning of the questions, quickness of response and economy of execution. Subjectives' accessibility to objective facts and to other people's opinions are judged to be essential to secure the accurate communication and high stability of responses.
A group survey method experimented here employed real time Delphi method where inquiries and answers were iterated along with the group's total opinion fedback to the participants. The effectiveness of this information exchange scheme process were examined in comparison with the results from personal interviews executed beforehand.
In June 1981, four experimental conferences were held at the Evauation Laboratory of Man Environment Systems (ELMES) set up in NIES, with the aim of evaluating the environment of roadside areas. In those conferences, 17 to 24 citizens of Tuchiura City, Ibaraki Prefecture, selected randomly from the subjectives interviewed three months before, gave weights by rating method (the values vary from 0 to 100) to 7 items which represented characteristics of roadside areas' environment, such as noise, vibration, air pollution, dust, litter, pedestrian danger and difficulty in crossing the road. Electronic voting device, which accumulates the participants' votes immediately, indicated to them the average points scored by the group after every voting. This procedure were iterated three times with discussions among participants in between. After the weighting of indivisual items, participants are asked to assign grades in five steps to their nearby environment. This process were iterated twice with discussions in between.
Conclusions conducted from those experimental conferences are as follows: 1) In examining the results from interview and conference, the scored values concerning the participants' nearby residential environment were not so stable. And the rating method was judged to be unstable for weighting in comparison to the pairwise rating method (the linear weighting method), based on large variation observed among participants' opinions. 2) Employment of real time Delphi in a group conferences leads to gradual stabilization of serial responses. 3) On the assumption that the comprehensive environmental evaluation of roadside areas were estimated by the linear combination form of the 7 items, the coincidence of the regression model with the participants' total evaluation on the environment were examined. The coefficient of determination (R²) of 0.949 estimated from the conference, in comparison to 0.753 from the personal interview, showed the effectiveness of real time Delphi method in evaluating the residential environment comprehensively.
In addition to those results, discussions held in the conferences clearly showed the differences of the environmental perceptions among participants. And answers to the inquiries asked after the conferences indicated that the information exchange scheme employed to the conferences promoted participants' motivation to participate in this kind of group conferences.