Information revelation for better or worse recommendation: Understanding Chinese users’ privacy attitudes and practices

Document Type

Conference Proceeding

Publication Date



The performance of a Recommendation System (RS) is mainly determined by how well it ‘understands’ its users: how much information it is able to trace and obtain. The former is largely depended on the algorithmic designs of the recommendation which have been explored for over a decade, while the former is more and more determined by the data owners—the users. In this paper, we presented our study on understanding Chinese users’ information revelation attitudes and practices which has not been fully explored in both the recommendation and online privacy research fields. Specifically, unlike the majority of previous studies that revealed users’ self-disclosure practices and attitudes in SNSs, our study is the first and only an initial investigation into the two aspects of the personal information disclosure and sharing: the differences of personal information shared by and with different types of audiences. Our study revealed that among the five types of recipient’s, students placed the least trust on Advertisers, among four other groups as Close Friends and Family, University Community, Friends on the Social Networking Sites, and Complete (online) Strangers. Overall, students feel more comfortable actively sharing personal information with the types of audiences than being shared by these audiences of their personal details.

Publication Title

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

First Page Number


Last Page Number




This document is currently not available here.