UBC Theses and Dissertations
Towards usable end-user privacy control for social software systems Najafian Razavi, Maryam
The recent growth and wide adoption of social software systems have transformed the Web from an information pool to a platform for communication and social interaction. While often times social software systems are used with the goal of sharing information, studies have shown that many users struggle to properly manage selective sharing of the vast and diverse information artifacts they dispose in such tools. Most existing social software systems define privacy either as a private/public dichotomy or in terms of a “network of friends” relationship, in which all “friends” are created equal and all relationships are reciprocal. These models fail to support the privacy expectations that non-technical users bring from their real-life experiences, such as segregating one’s disparate groups, enabling different degrees of intimacy within one’s network, and providing flexible, natural means of managing the volatile social relationships that social software systems confront. Furthermore, both models suffer from lack of empirical grounding and systematic evaluation. The research described in this thesis employs a qualitative research methodology to deepen understanding of the information sharing process in the context of social software systems, in order to propose guidelines for building privacy management mechanisms in this domain that provide users with more control over privacy, and yet, are intuitive and easy to use for the average, non-technical user population of social software. The research is based on a grounded theory study of users’ information sharing behavior in a social software tool, and offers several contributions, including clarifying users’ privacy needs, concerns, and strategies, and identifying factors that affect users’ decisions regarding sharing various information artifacts with different audiences. The findings lead to the development of several design heuristics and a general framework for usable privacy in social software domain, which inform the design of OpnTag's, a novel prototype that facilitates creation, organization, and sharing of information for an individual operating in various social contexts. Results of an empirical evaluation of OpnTag’s privacy management mechanism show that our proposed privacy framework is flexible enough to meet users’ varying information sharing needs in different contexts while maintaining adequate support for usability.
Item Citations and Data
Attribution-NonCommercial-NoDerivatives 4.0 International