摘要
本文根据中国的商业环境,对淘宝网数据进行实证分析,发现线上商品评论呈近似反"L"型的非对称分布,且评论的均值不是商品质量的无偏估计量。这与现有研究认为线上商品评论呈正态分布并假设商品评价的均值是无偏估计量的结论不同,与线上评论呈双峰分布的国外的实证结论亦不相同。为了深入研究线上商品评论存在偏差的原因,以及反"L"型分布的特征,本文构建了理论模型,该模型得出商品评价的均值作为无偏估计量的条件,并进一步揭示出中国电子商务网站"默认好评"机制和消费者主动评价偏差是线上商品评论存在偏差的重要原因。
Based on the Chinese business environment, the paper firstly conducts an empirical analysis using data from Taobao. Results are the online product review follows a reversed ＂ L＂-type distribution, of which the mean is a biased estimator of the product＇s true quality. These findings are different from the exist^d results that online product reviews follow a normal distribution. They are also distinct from the empirical results of foreign e-business data. In order to study the reasons for the biased online product reviews and the features of the reversed ＂L＂-type distribution, the paper constructs a theoretical model to obtain the conditions for the mean of the online product reviews to be an unbiased estimator. The results demonstrate that the ＂default good reviews＂ mechanism and the deviation of consumer＇s initiative evaluation are the main reasons for the biased online product reviews.
出处
《统计研究》
CSSCI
北大核心
2013年第4期46-51,共6页
Statistical Research
基金
国家自然科学基金(重大研究计划培育项目)(90924009)
国家自然科学基金(面上项目)(70871126)资助
关键词
线上商品评论
商品质量
分布
偏差
纠正
Online Product Reviews
Product Quality
Distribution
Deviation
Correction
