1. Background and objective This study examines the methodology of Causality analysis on environmental evaluation. The purpose of this paper is to propose a new method to grasp causes and effects of "attractiveness(subjective satisfaction)" and "dissatisfaction" based on environmental evaluation data.
2. Basic idea First, it is necessary to conduct a survey that respondents evaluate the environment using the following three kinds of evaluation items: i) Evaluation items of each viewpoint (several to dozen), ii) Evaluation of "attractiveness" and "dissatisfaction" (2 items), iii) Comprehensive evaluations (1 or more items).
Then, the following causal model is applied to the obtained evaluation data:
{Cause variables: evaluation items of each viewpoint} → {Intermediate variables: "attractiveness" and "dissatisfaction"} → {Result variables: comprehensive evaluations}
Based on the obtained causal model, causal effects from "attractiveness" and "dissatisfaction" to comprehensive evaluations are grasped. And cause variables are classified using Kano model as follows.
“Attractive Quality”: Items that have effect to "attractiveness" and no effect to "dissatisfaction".
“One-dimensional Quality”: Items that have effect to both "attractiveness" and "dissatisfaction".
“Must-be Quality”: Items that have effect to "dissatisfaction" and no effect to "attractiveness".
3. Case study In order to examine more concrete survey and analysis methods, analysis based on the proposed method was applied to the following three survey cases. Case1: survey for regional environment, Case2: CS-survey for pablic facilities, Case3: survey for large-scale park.
4. Findings The most important outcome of this research is that we were able to show that it is possible to grasp the cause and effect of "attractiveness" and "dissatisfaction" based on evaluation data of case study survey.
Furthermore, the following findings regarding methodology and evaluation mechanism were obtained.
As the first step of analysis, how to apply multiple regression analysis to select variables used for analysis was proposed. As the next step, moodel building process in SEM was proposed.
The effectiveness and significance of the technique using intermediate variables "attractiveness" and "dissatisfaction" as latent variables was discussed. This technique has the significance of preventing deterioration of goodness of fit due to poor reliability of "attractiveness" and "dissatisfaction", and the significance of relaxing the assumption required for the actual evaluation structure represented by the model.
It was shown that causal model with hierarchical structure is sometimes useful, sometimes not useful. The latter case is when Must-be Quality items and Attractive Quality items coexist in lower level items affecting the same upper level items.
As a visualization of the effects of "attractiveness" and "dissatisfaction", factor loading plot of "attractiveness" and "dissatisfaction" was proposed. This graph is also useful for comparison between cases.
5. Future tasks and prospects The future tasks are to apply the proposed method to many cases, and to reconsider relationship between the proposed method and the Kano model (The item classification by the Kano model is defined by function form of causal effect to result variable.).
抄録全体を表示