就雙界二分選擇法的詢價情況，起價點配置的不同，如何影響願付價格的評估？再者，評估願付價格時，採用Tsai(2005)所提出的三要素模型來進行分析，是否會較傳統單一要素模型來的有效率？本文將透過模擬實驗的方式，來就前述兩項議題作分析探討？模擬結果顯示在評估願付價格平均數時，我們應該採用等距的設計方式設計起價點；在評估願付價格中位數時，則應該採用等百分比的設計方式。而第二次詢價採用提高為原始價格兩倍或降低二分之一的方式，應該是可以接受的。就模型的比較而言，我們發現不論資料的型態為何，三要素模型都較單一要素模型來的有效率。 / Although the double-bounded dichotomous choice approach is commonly used in a survey to estimate WTP values for a non-market good, the impact of the initial bid values is still not quite clear. Furthermore, models commonly used to analyze the data assume that all the respondents are willing to pay a reasonable price. Tsai(2005) presents a three-component mixture model that relaxes the assumption. How efficient is the model compared with the conventional “one-component model” is also unknown. We focus on these two issues in this study. Simulation results show that equally-spaced design is good for estimating the mean WTP, while designs based on percentiles should be considered as far as the median WTP is concerned. And, the conventional approaches that choose the follow-up bid values to be either double or one half the initial bid values appear acceptable. In addition, regardless the data structure, the three-component models are more efficient than the conventional one-component models.