讲座预告 | 何洁:Treatment of “Don’t-Know” Responses in Contingent Valuation Surveys
发布时间:2017-10-19
十大信誉赌博官网平台环境与资源经济学前沿系列讲座
Treatment of “Don’t-Know” Responses in Contingent Valuation Surveys:
Random Valuation Model Revisit
时间:2017年10月25日下午2:00 -- 4:00
地点:十大信誉赌博官网平台实验楼201会议室(校友之家后小楼二层)
主讲:何洁(Jie He),加拿大Sherbrooke大学经济和管理学院经济系教授。
主讲人简介:
何洁教授,加拿大Sherbrooke大学经济和管理学院经济系教授,于法国奥弗涅大学国际经济发展研究所(CERDI)经济学获得博士学位。曾任比利时鲁文天主教大学(2006)、香港理工大学(2009)、中山大学(2013-2015)访问教授,曾就职于世界银行(2004-2005, 2008-2009)及联合国环境署(UNEP,2005)。担任四十多家国际学术期刊的匿名评稿人,2014年起每年担任魁北克社科研究基金(FRQ)的评审委员会成员。多次为法国国家研究基金(ANR)做相关项目的专业评审, 曾主办加拿大环境资源经济学会年会,任China Economic Review期刊客座编辑, International Journal of Health Preference Research编委会委员, 留美经济学会(Chinese Economist Society)理事(2011-2012, 2013-2014),环球中国环境专家学会(Professional Association for China’s Environment, PACE)理事(2009-2018)。自2015年以来参与筹建中国环境自然经济学会(Chinese Association of Environmental and Resource Economics, CAERE)。
何洁教授的研究一方面关注中国发展过程中经济增长、对外开放(贸易, FDI)和环境的关系, 另一方面主张将非市场价值评估方法应用到公共产品(环境质量和自然资源以及医药和医疗方法)的价值评估中。有近40篇学术论文在国际专业期刊上发表,被引用次数高达2300余次(google scholar, 中国知网)。主持参与世界银行、加拿大国家社科基金(SSHRC)、加拿大魁北克省研究基金(FRQ)、加拿大基因组(genome Canada)、加拿大自然资源部(Nature Resource Canada)、加拿大区域气候联盟(OURANOS)、加拿大国际发展研究所(IDRC)资助的20多个研究项目。近年来也和十大信誉赌博官网平台、中山大学、南京大学、暨南大学等的中国同事合作,推广环境自然资源非市场价值评估在中国的应用。
讲座摘要:
美国NOAA条件价值评估方法小组建议在主要的投票评估问题中,明确允许出现“是”和“否”选项之外的“不回答”(No Answer)或“不知道”(don’t know,DK)选项(Arrow等,1993)。 这个建议的一个原因在于调查中有相当一部分受访者回答典型的态度问题时选择了这一选项。 然而,关于如何使用这种中间回应尚未达成一致的意见。在条件价值评估(CV)中处理“不知道”选项的方式,密切取决于其背后造成“不知道”回答的原因。
本文中,我们提出了一个新的假设,即若受访者已充分明确其潜在的支付意愿(WTP)数值,则被调查者在封闭式支付意愿问题中的“不知道”回答将更有可能遵循Wang(1997)提出的“中间回应假设(middle-response assumption)”,而在其他情况下,其“不知道”回答可能更类似于“无回答”。为检验上述假设,我们实施了三次关于公共物品的条件价值评估,通过设计具体的支付意愿不确定性问题,来推断“不知道”回答的意义。问卷从一个标准的封闭性支付意愿问题开始,首先要求受访者就给定的投标价格在“是”、“否”和“不知道”选项之间进行选择,随后请受访者在附加的确定性测度问题中表达他们的不确定性情况,该问题提供了从“绝对确定”到“绝对不确定”等五种口述确定性水平。此外,参照Haener和Admowicz(1998)的方法,无论受访者在支付意愿问题上选择了哪一个选项,他们都被要求在一个开放性问题中给出其最大支付意愿,同时也会附加有另一个确定性测度问题。基于上述问卷设计,我们设法将选择“不知道”回答的受访者区分为两类:一类是能够充分明确自己在开放式问题中的支付意愿数值的受访者,另一类是最大支付意愿不够明确的受访者。初步结果表明,将“不知道”作为“无回答”的简单二分选择(DC)模型与将“不知道”作为中间回应的Wang(1997)模型,他们两者之间在估计效率上的比较,仅在因不同潜在原因导致的“不知道”回答被更为清晰地界定时才有意义。
The U.S. NOAA panel on contingent valuation method endorsed the reference approach but suggested that a “no-answer” or “don’t know” (DK) option to be explicitly allowed in addition to the ‘yes’ and ‘no’ vote options on the main referendum valuation question” (Arrow et al., 1993). One reason for this recommendation is the recognition of a sizable portion of respondents took such option while answering typical attitude questions in surveys. However, there is not yet a consensus about how to use this middle response. The choice about how to treat the “don’t know” (DK) answers in a CV survey depends closely on the reason lying behind that motivate such answers.
In this paper, we propose a new hypothesis which assumes that, if a respondent has made sufficient effort to identify his/her implicit WTP value, his/her DK answer in the close ended WTP questions will be more likely to follow the middle-response assumption provided in Wang (1997), in other cases, his/her DK answers may be similar to a No answer. To test such assumption, in three public good contingent valuation surveys, we propose a specific WTP uncertainty question design that can be used to induce the meaning of DK response. Starting with a standard close-ended referendum WTP question that asks respondents to choose between the “Yes”, “No” and “DK” options for the proposed bid price, our questionnaire also invite respondents to express their uncertainty in a follow-up certainty scale question, where we offered five verbal certainty levels varying from absolutely certain to absolutely uncertain. In addition, following Haener and Admowicz (1998), no matter which response option was chosen for the WTP question, respondents were asked to report the amount of their maximal WTP in an open-ended question, followed by another certainty scale. Based on such questionnaire design, we managed to distinguish the respondents choosing DK responses into two categories: those making enough efforts therefore providing certain open-ended WTP amount and those making insufficient efforts therefore giving uncertain maximal WTP amount. Our preliminary results showed that the estimation efficiency between the simple dichotomous choice (DC) model recoding the DK as No and the Wang (1997) model, which consider DK as a middle response became comparable only when the DK responses due to different motivations were more clearly identified.