Analysis on Star Rating System in Delivery Platform: Question on Scale Equivalence

In online platforms that sell products and services, reviews provide consumers with indirect experiences and contextual information, enabling them to make more informed decisions. Among these reviews, star ratings, which are expressed through various scales, summarize extensive review texts to support consumer decision-making and serve as key indicators for improving services such as platform recommendations and search algorithms. However, despite the possibility that star ratings may be biased due to user subjectivity, contextual factors, and psychological influences, there has been insufficient discussion on whether star ratings maintain interval scale properties. If star ratings do not maintain interval scale characteristics, statistical interpretations based on these ratings may be distorted, leading to errors in both consumer decision-making and corporate strategic planning. Therefore, this study examines the interval scale consistency of star ratings using user review data from Baedal Minjok, a major online food delivery platform in Korea. Reviews of chicken and jokbal (braised pig’s feet) collected from January to April 2024 were analyzed, and a survey was conducted with 120 participants. Participants were asked to read 30 randomly selected reviews with star ratings and reassign perceived ratings. By comparing the means and variances between actual and perceived ratings and conducting bias and median analyses, the study verified the interval scale consistency and biases of the star rating system. The results revealed that the mean intervals between star ratings were not equal, indicating that the star rating scale does not satisfy interval scale properties. This study highlights the necessity of improving star rating systems in online review platforms and suggests that developing more sophisticated evaluation methods can enhance consumer trust and decision-making processes.