| Description |
In this research, we investigate how reviews written by different types of reviewers can have different impacts on sales and on other reviewers. To investigate the role of review source, we use Amazon reviewer ranking to identify the two groups of top-ranked and bottom-ranked reviewers. In the first essay, we use a weekly panel dataset of 182 music albums over the first 8 weeks after their release date along with their weekly DMA-level sales and the demographic data to study the impacts of each reviewer group's product ratings on sales. Specifically, we have the following key research questions in the first essay: Do top-ranked reviewers or bottom-ranked reviewers have a bigger impact on product sales? What is the role of review content in explaining the differential impact of the two reviewer groups on product sales? (What is the sensitivity of different groups of customers, based on their demographics, to the product ratings of top-ranked reviewers and bottom-ranked reviewers?) How will the impact of the two groups change) when there is more uncertainty and ii) when the product is new? We test the robustness of our results by repeating the analysis using data from a different product category - cameras. Also, we use Amazon sales rank instead of DMA iv sales data. We use the instrumental variables for endogeneity correction of the key online WOM measures. In the second essay, we use a daily individual reviewer level data for music albums over 60 days after their release date. We are interested in uncovering new insights about how the review generation process is influence by top-ranked or bottom-ranked reviewers. The research questions of the second essay are as follows: How does each reviewer group impact product ratings awarded by subsequent to pranked and bottom-ranked reviewers? (How will such impact on a subsequent reviewer change considering) the product type (i.e., genre of the album) and the product price? What is the role of review content, particularly power and achievement, in explaining the differential impact of the two reviewer groups on a subsequent top pranked and bottom-ranked reviewer's rating? We use Gaussian Copulas to correct for endogeneity of the key measures. |
