初始投标值数量和样本容量对双边界二分式CVM的影响

双边界二分式条件价值评估法是当前广泛应用于评估生态环境和自然资源价值的陈述偏好法。为了研究双边界二分式CVM中,不同分布中初始投标值数量及样本容量对WTP的影响,应用蒙特卡洛模拟和支付意愿函数模型相结合的方法,用概率分位数设计投标值,分别选择数据来源不同的Weibull分布和对数Logistic分布,动态模拟不同初始投标值数量和样本容量对WTP期望值产生的影响。模拟结果表明,对不同的WTP概率分布,初始投标值数量和样本容量对WTP的影响趋势是一致的,当初始投标值数量和样本容量分别大于5和500时,初始投标值数量和样本容量的增加,对WTP估计值的影响程度越来越小。研究结果表明,在二分式CVM研究中,当以WTP期望为代表值时,其初始投标值数量和样本容量至少为5和500。研究结论为二分式CVM问卷设计中投标值数量及样本容量的设定提供参考依据。

Influence of initial bid number and sample size on the double-bounded dichotomous choice contingent valuation method

The contingent valuation method (CVM) is a survey-based method that uses random samples of families or individuals to reveal consumer preferences for goods or services, such as environmental resources, in a hypothetical market. Through questionnaires, CVM identifies people''s preferences and expresses them in dollar terms as willingness to pay (WTP) or willingness to accept (WTA). WTP (or WTA) can be used to gain insights into the potential economic benefits or losses brought about by a planned project. A great variety of questionnaire formats have been developed and applied, among which the dichotomous choice CVM is considered to be one of the most advanced methods. In recent years, the double-bounded dichotomous choice contingent valuation method (DBDC-CVM) has been widely used to value the ecological environment and natural resources. It can identify respondents'' true WTP and reduce the deviation of assessment results, thereby making the results more accurate and reliable. The initial bid number and sample size are very important determinants of WTP in DBDC-CVM. Fewer initial bids and smaller sample sizes will reduce the validity and reliability of the resultant WTP. In contrast, more initial bids and larger sample sizes increases the difficulty and cost of the survey. The number of initial bids is always obtained using a pre-test, while the sample size is always determined by field research or online surveys. This study investigates the dynamic effects of initial bid number and sample size on mean WTP, using the willingness to pay (WTP) function model and a-posteriori Monte Carlo simulation. We use Weibull and log-logistic distributions of WTP to simulate two different sources of data. In order to simplify the Monte Carlo simulation, we quantize the probability distribution and make a hypothesis to design bids. We produce four graphs to visualize the simulated results and use these to identify the optimum number of initial bids and sample size. Initial bid number and sample size had almost the same influence on mean WTP for both WTP distributions. When the initial bid number t and sample size exceed 5 and 500, respectively, the influence on mean WTP is almost the same. Based on this result we suggest that at least 5 initial bids and 500 samples are needed in order to ensure the validity and reliability of estimates of mean of WTP generated by Monte Carlo simulation. These results provide a quantitative basis for CVM questionnaire design and have the potential to improve the validity and reliability of WTP.