Durham E-Theses Systematic Measurement of Centralized Online Reputation Systems

Background: Centralized online reputation systems, which collect users’ opinions on products, transactions and events as reputation information then aggregate and publish it, have been widely adopted by Internet companies. These systems can help users build trust, reduce information asymmetry and filter information. Aim: Much research in the area has focused on analyzing single type systems and the cross-type evaluation usually concentrates on one aspect of the system. This research proposes a systematic evaluation model (SERS) that can measure different types of reputation system. Method: From system perspective, all reputation systems can be divided into five underlying components. Input refers to the collection of ratings and reviews; Processing is the aggregation of ratings. Output publishes the information. Feedback Loop is the collection of the feedback of the review, which can be seen as the ‘review of the review’; Finally, Storage stores all the information. Therefore, based on each component’s characteristics, a series of benchmark criteria can be defined and incorporated into the model. Results: The SERS has defined 29 criteria, which can compare and measure different aspects of reputation systems. The model was theoretically assessed on its coverage of the successful factors of reputation systems and the technical dimensions of information systems. The model has also been empirically assessed by applying it to 15 commercial sites. Conclusion: The results obtained indicated that the SERS model has identified most important characteristics that have been proposed by reputation systems literature. In addition the SERS has covered most dimensions of the two basic technical information system measurements: information quality and system quality. The empirical assessment has shown that the SERS can evaluate different types of reputation systems and is capable of identifying the weakness of current systems.