基于误判率的最大差异度量法  

Best-Worst Scaling Based on Misjudging Rates

作　　者：陈陌[1,2] 杨靖[2] 赵威[2] 

机构地区：[1]天津大学管理与经济学部,天津300072 [2]中国汽车技术研究中心,天津300300

出　　处：《统计与决策》2017年第4期32-35,共4页Statistics & Decision

摘　　要：最大差异度量法在市场研究中越来越受到重视,但是其问卷设计中包含的误判信息并没有被有效利用,文章在定义了顺序比较和三角误判的概念后,提出了计算个人误判率的方法,统计出的个人误判率可以作为排除误判率较高评价者的依据。在此基础上,结合汽车消费者购买因素调查的实际案例,验证了基于误判率的最大差异度量法的有效性,同时利用不同购买因素之间的有效对比计数和概率信息,得到了个体的最优排序,为后续研究提供了更有价值的数据信息。The use of Best-worst scaling is more and more important in market research, but the misjudging information is underutilized in the questionnaire design. Based on the definition of the order-comparison and the circular triads, the computing method of individual misjudging rates is proposed, and the results can be as a criteria for removing some evaluators with high misjudging rates. Combining the case about purchasing factors of vehicle consumer, the validity of best-worst scaling based on misjudging rates is verified. Simultaneously, aiming at different purchasing factors, the actual comparison counting and probability are used to judge the optimal purchasing factors ranking for every individual, which is more valuable in future research.

关 键 词：最大差异度量 误判率 顺序比较 三角误判 个体最优排序 

分 类 号：C934[经济管理—管理学]