Research On The Formation Of Online Review Dynamics And Its Influence Mechanism

In recent years, the rapid development of electronic commerce has prompted the popularization of online shopping. The virtuality and uncertainty of online shopping has increased the importance of online product reviews in consumers’purchase decisions. As an important product information source, online product reviews not only play a decisive role in consumers’purchase intentions but also provide an effective way for firms to increase sales. The value of online product reviews depends on whether the reviewers provide objective and accurate product evacuations that can help potential consumers make effective purchase decisions. However, the extant research suggests that online product reviews do not always accurately reflect the product’s true quality but rather be a biased estimator of its true quality. An important reason is that online consumers’review behavior will be affected by prior product reviews, rendering the formation of product reviews to be dynamic. The dynamic process in online reviews can be described as that a reviewer’s rating behavior will be affected by prior reviews, and his/her provided review together with other prior reviews will then affect subsequent reviewers’ rating behavior.In general, the existing studies on online review dynamics still have the following drawbacks. First, the formation mechanism of online review dynamics is still unclear. Extant studies look at the effect of prior reviews on either the opinion formation part or the opinion expression part of subsequent reviews, but fail to distinguish or integrate these two effects. The empirical evidence has also been mixed. Second, research on which form of review dynamics (conforming with or deviating from prior reviews) is more helpful remains insufficient and controversial. Third, whether social influence on product reviews (a mechanism of review dynamics) will affect consumers’perceptions of review helpfulness is not well studied at present.Having considered the above mentioned research weaknesses, this dissertation mainly focus on the following three questions. First, how do prior product reviews affect subsequent reviewers’ rating behavior? Second, when a review conforming with or deviating from the majority of prior opinions will be perceived as more helpful? Third, if a subsequent review is subject to social influence from prior reviews, will this social dynamics affect the helpfulness of this subsequent review? To answer these three questions, this dissertation has conducted three sub-researches.The first sub-research draws on an integrative view of information diagnosticity and social influence theory to analyze the effects of prior product reviews on subsequent reviewers’opinion formation and opinion expression, and examine the moderating effects of reviewer factors (review connectedness and reviewer expertise) and rating contextual factors (volume of prior ratings and variance of prior ratings) as well as their interaction effects on the relationship between the average rating of prior reviews and a subsequent rating. With an analysis of 70,410 reviews across 100 restaurants collected from Yelp.com, the empirical results show that both the volume and variance of prior ratings exert negative moderating effects on the positive relationship between the average of prior ratings and a subsequent rating while the moderating effects of reviewer connectedness and reviewer expertise are not supported. In addition, when reviewer connectedness or expertise increases, the negative moderating effect of the volume of prior ratings becomes weaker but the negative moderating effect of the variance of prior ratings becomes not significant.The second sub-research examines the moderating effect of deviation direction on the relationship between rating deviation and review helpfulness and whether this moderating effect is contingent on product type. With an analysis of 80,400 reviews from 100 Yelp restaurants and 117 Amazon tablet computers, the empirical results indicate that deviation direction and product type jointly affect the relationship between rating deviation and review helpfulness. Specifically, for search products, the more a review positively deviates or the less it negatively deviates from prior reviews, the more helpfulness it will be perceived. In contrast, for experience products, the less a review positively deviates or the more it negatively deviates from prior reviews, the more helpfulness it will be perceived.The third sub-research introduces the concept of review order and draws on social influence theory to examine the relationship between review order and review helpfulness, as a means to validate the effect of social dynamics of online reviews on review helpfulness. In addition, three types of moderators-reviewer characteristics (reviewer connectedness and expertise), review characteristics (review rating and length), and a temporal factor (time distance between a review and the first review of a product), are proposed to provide additional support to our social influence perspective on the effect of review order on review helpfulness. With an analysis of 70,610 reviews across 100 restaurants collected from Yelp.com, the empirical results indicate that, all else being equal, the order of a review has a negative effect on review helpfulness. Moreover, the negative effect of review order on review helpfulness becomes weaker when the reviewer has more connectedness or a higher level of expertise, or when the review is more negative in content or more recently posted.This dissertation contributes to the previous literature on review dynamics and review helpfulness in the following ways.First, within an integrative view of information diagnosticity and social influence theory, this dissertation explains how prior reviews will affect subsequent reviewers’ rating behavior both in the opinion formation phase and opinion expression phase. It uncovers the underlying mechanism for the formation of online review dynamics. In addition, this dissertation provides insights on how reviewer characteristics and rating environment work in the dynamic process of online reviews, which advances our understanding on the formation of online review dynamics.Second, this dissertation found that deviation direction and product type jointly affect the relationship between rating deviation and review helpfulness. It helps reconcile the mixed findings in the literature on the effect of rating deviation on review helpfulness.Third, this dissertation is the first to examine whether the social dynamics in online reviews will have a detrimental effect on review helpfulness through investigating the effect of review order on review helpfulness.