Date of Award

August 2022

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Management Science

First Advisor

Atish Sinha

Committee Members

Arun Sen, Mark Srite, Yang Wang

Keywords

online customer reviews, online healthcare reviews, online ratings, review helpfulness, text mining, topic modeling

Abstract

As e-commerce platforms keep growing in popularity, online customer reviews, which represent users’ evaluation of products or services, have become a crucial information source in consumer decision-making. Online reviews have proven to have a major impact on various critical aspects of a business, such as reputation, sales, and product returns. The goal of the three-essay dissertation is to investigate the influential antecedents and consequents of online customer reviews in business and healthcare. Essay I explores the effects of product exposure time on review content and review helpfulness. We find that the descriptions of utilitarian attributes in a review increase with product exposure and mediate the relationship between product exposure and review helpfulness. To test the effects of product exposure, we first extract latent topics from review content and then identify utilitarian topics from them. Next, we build a regression model to test the utilitarian information’s relationship with product exposure. The results support our central thesis that product exposure has a significant positive influence on review helpfulness, and this relationship is mediated by the utilitarian information in a review. We also find that users who have prior knowledge of the domain the product belongs to do not need long exposure times to write helpful reviews. Our findings demonstrate the need to account for product exposure and domain knowledge when examining online review helpfulness. The finding that early reviews tend to be less helpful because they contain less utilitarian product information has important implications, both for research and for practice. Essay II focuses on investigating how medical performance factors affect hospitals’ online ratings. We find that readmission, mortality, safety of care, and time in emergency department significantly influence a hospital’s online reputation. We also extract three influential review content factors: reviewer medical knowledge, medical quality evaluations, and service quality evaluations. This is the first study to investigate the effect of a set of representative hospital medical performance factors on online ratings. Furthermore, it is the first attempt at examining the roles of reviewer medical knowledge and different types of experiential quality evaluations in the online healthcare review domain. We also find a significant influence of CMS overall quality star rating on a hospital’s online reputation. The findings provide valuable inputs into a hospital’s marketing strategies and have important managerial implications for providers, patients, and online platforms. In Essay III, we propose a hybrid aspect-based sentiment analysis (ABSA) framework that mines patients’ online evaluations of a hospital from different aspects. We then integrate the extracted average sentiment polarities into regression models, where the numeric online rating is the dependent variable. The results show that including the aspect categories’ polarities dramatically increases the models’ fit. The standardized coefficients reveal that “Staff,” “Nurse,” and “Doctor” are the three most influential aspects of hospitals’ online review ratings. Our results prove the necessity of adopting ABSA in the online healthcare review domain. They also have practical implications for patients, healthcare providers, and online review platforms.

Available for download on Sunday, September 01, 2024

Share

COinS