Abstract
Online consumer reviews (OCRs) enable consumers to publicly communicate post-purchase evaluations, satisfaction, and electronic word-of-mouth (eWOM). This study investigates the associational relationships between multi-dimensional product and service evaluations, expectations, and overall consumer ratings. The unit of analysis is the individual OCR. A sample of 2300 OCRs of 20 best-selling leggings was drawn from a brand-owned website. Data were analyzed using topic modeling with human validation of topic labels, sentiment analysis, and regression models to account for product-level dependence and statistical uncertainty. Topic modeling extracted themes related to expectations, multidimensional product and service performance, confirmation/disconfirmation, satisfaction, and eWOM expressions. Sentiment analysis indicated that expectations, aesthetic product performance, and recommendations leaned positive, whereas service and extrinsic product performance exhibited neutral to negative sentiment. The regression results demonstrated positive associations between overall satisfaction and four dimensions: product performance, service performance, confirmation/disconfirmation, and post-purchase outcomes.