【作者】 李雨洁；

【导师】 廖成林；

【作者基本信息】 重庆大学 ， 企业管理， 2014， 博士

【摘要】 随着网络购物的兴起，其弊端也日益凸显，尤其是网络的虚拟性导致消费者难以判断商品的真实质量，而解决该问题的常见方法是借鉴其他消费者的在线商品评论。但是，在线商品评论的真实与客观性却备受质疑。倘若在线商品评论不能代表大多数消费者的真实感知，传递出不准确的商品信息，则会增加其他消费者的交易风险，同时误导企业制定错误的营销策略。因此，需要判断在线商品评论偏差的存在性。若存在偏差，则需分析导致偏差的因素，不同的影响方式，以及针对性的纠偏措施。鉴于此，本文拟以“C2C模式下在线商品评论偏差的成因与影响机理”为题进行研究。从在线商品评论偏差的存在性研究入手，分别从消费者、网站和卖家三个视角分析偏差的成因，重点分析消费者的个体行为因素、C2C网站的评价机制和卖家的操纵评论行为对在线商品评论真实性的影响机理，探寻纠正评论偏差，获取真实商品信息的方法。首先，判断我国C2C网站的在线商品评论是否存在偏差。实证分析获取我国C2C网站的在线商品评论的整体数据特征。由于该数据特征与已有研究结论存在差异，为解释冲突的原因，通过行为实验模拟网络购物与评论的过程，排除消费者评价缺失、网站评价机制、卖家操纵评论等干扰因素对评分结果的影响，以获取消费者对商品真实的评价。通过比较分析C2C网站数据、实验数据，以及国外相关研究成果，发现并论证了我国C2C网站的在线商品评论偏差的存在性。其次，基于在线商品评论偏差的数据特征，从在线商品评论的三大主体――消费者、网站和卖家，系统研究在线商品评论偏差的成因及各因素的影响机理。第一，消费者个体行为因素对在线商品评论偏差的影响机理研究。首先，从消费者的视角，基于在线商品评论偏差的数据特征，筛选出导致评论偏差的影响因素，包括消费者的自我选择偏差行为、未主动评价行为和评价行为趋同性。其次，利用分布函数刻画出消费者在线评论行为特征对评论偏差的影响，给出反映在线商品评论特征的密度函数及该分布的期望和方差。进一步，讨论在线商品评论均值的期望反映商品真实质量的理论条件，并讨论各因素对评论偏差的影响机理。研究结果加深了对消费者视角下在线商品评论偏差形成过程的认识，为该视角下评论偏差的纠正提供理论基础。第二，C2C网站评价机制对在线商品评论偏差的影响机理研究。首先，基于在线商品评论偏差的数据特征，立足于我国C2C的情境，筛选了导致在线商品评论偏差的影响因素，包括“默认好评”机制和“退货交易关闭”机制。其次，建立反映该机制的在线商品评论模型，利用模型中分布函数的期望和方差，研究了各因素对评论偏差的影响机理。再次，得出该视角下在线商品评论反映商品真实质量的理论条件。研究结果进一步阐述了评价机制与评论偏差的关系，为该视角下评论偏差的纠正提供理论基础，也为中外跨文化比较提供新的出发点。第三，卖家操纵评论行为对在线商品评论偏差的影响机理研究。首先，筛选出卖方视角下，导致在线商品评论出现偏差的影响因素。其次，根据在线商品好评率的数据特征，通过数学推导解析数据成分，对C2C网站的卖家操纵评论行为进行识别，判断淘宝网卖家操纵评论行为的存在性。再次，通过建立在线商品评分模型反映卖家操纵评论行为特征，探讨随商品上架时间变化的卖家操纵评论行为对偏差的影响机理。最后，结合淘宝网某店铺的现实数据，对在线商品评论模型的参数进行拟合求解，完善模型，最终达到纠正在线商品评论偏差的目的。研究结果对网站改进评价机制和方法，提高在线商品评论真实性提供了理论依据。最后，针对以上三个视角的研究得出本文的研究结论。同时，提出对策建议，帮助网站、消费者和卖家克服自身缺陷，纠正评论偏差并维护良好的交易环境。更多还原

【Abstract】 With the rise of the online shopping, the disadvantages are also growing.Especially the virtual of the internet makes consumers difficult to determine the truequality of the commodities. The common measure to solve the problem is to learn fromthe existed online product reviews. However, the authenticity of online product reviewsis still questionable. If the online product reviews do not represent the real perceptionsof the mass and couldn’t convey the accurate information of the commodity, which willincrease the risk of the other consumer’s transactions and will mislead enterprisesmaking correct strategy. Given this, the paper concentrates on the topic of “TheFactors and Influence Mechanism of Online Product Reviews Bias Based on C2C Mode”to make a deep research from three different angles to explore the sources of the bias.The paper focus on the influence mechanism of the three aspects including theconsumer’s self-selection bias, the C2C site’s appraisal mechanism and the seller’smanipulation behavior to the online product reviews to explore the methods to correctthe biased online product reviews.In the beginning, measuring whether there is bias in the online product reviews ofC2C sites in China. Empirically analyzes the overall data features of the C2Ce-commerce site. Through the behavioral experiment, the paper simulates the real sceneof the online shopping to acquire the experimental data features. The C2C sites’ data,the experimental data and the existed research results are compared to explore theexistence of the bias in the online product reviews of C2C sites in China.Then, from three angles of the online product reviews main body—the consumer,the site and the seller, the paper explores the sources and the influence mechanism ofthe biased reviews.Firstly, the influence mechanism of consumer’s individual preference to the onlineproduct reviews bias. In the first place, the paper discusses the factors of the biasedreviews from the angle of consumer’s individual preference, including the consumer’sselection preference, the inactive review behavior and the consumer’s conformity ratingbehavior. Then, based on the data features of the online product reviews, we constructthe model to reflect these factors influence mechanism. At last, we put forward theconditions of the mean to be the unbiased estimator of the product quality based on thisangle so as to overcome the bias of the consumer’s individual preference. Secondly, the influence mechanism of C2C sites’ appraisal systems to the onlineproduct reviews bias. In the first place, the paper discusses the factors of the biasedreviews from the angle of the C2C site’s appraisal system, including the “DefaultGood Reviews” mechanism, and the “Return Goods” mechanism. Then, based onthe data features of the online product reviews, we construct the model to reflect thesemechanism’s influence. At last, we put forward the conditions of the mean to be theunbiased estimator of the product quality based on this angle so as to overcome the biasof the C2C sites’ appraisal system.Thirdly, the influence mechanism of seller’s manipulation to the online productreviews bias. In the first place, the paper screens the influence factors from the seller’sangle. Then, based on the data features of the feedback rate, the paper recognizes theexistence of the seller’s manipulation behavior through the mathematics deduction.Then, we construct the online product reviews model to reflect this manipulatebehavior’s influence. At last, combing with the Taobao’s actual data, we derive the ofcurvilinear equation of the seller’s manipulation and acquire the real score of the onlineproduct reviews so as to reflect the influence mechanism of the seller’s manipulationbehavior along with the time and to achieve the goal of overcoming the bias of theseller’s manipulation behavior.At last, based on the previous research, the paper reaches the conclusion. Besidesthis, we put forward the related suggestions on how to conquer the defects of theconsumer, the C2C e-commerce sites and the seller.更多还原