Identifying the causes and effects of poor privacy practices by online social network users: The potential of advisory monitoring software in changing user behaviour

In this thesis the behaviour of online social network users is examined, with a particular focus on user activities that preserve or compromise personal privacy. A two-fold approach to data collection was used, with a large-scale online survey (n=90 respondents) that measured user's online social network use (using Facebook as a case study) and their personality traits (via Cattel's 16 Personality Factors questionnaire), followed by a smaller number of semi-structured interviews (n=14 participants). A hypothesis under examination was that personality traits serve to predict the behaviour of online social networks users, and have the potential to identify those users at risk of either carelessly or unknowingly compromising their personal privacy (and thereby placing themselves at risk of identity theft and other malice), and to inform the design of training materials that attempt to encourage those users to protect their personal information more carefully. Survey data were subject to correlational analysis to examine the association between the 16 Personality Factors, and the Big Five personality traits derived thereof, and the information visibility settings of the respondents (viz., in the online social network selected as a case study, the settings of "public", "friends only", "custom", "only me" or "not supplied" for each information type). To identify at-risk participant groups (i.e., those making sensitive personal information publically available), cluster analyses were performed by partitioning respondents into their dominant Big Five traits grouping. For survey respondents with the least cautious settings, it was found that the most highly correlated Big Five traits were Independence (Agreeableness) and Self-control (Conscientiousness), with a smaller effect of Tough-mindedness (Openness). More fine-grained analyses were performed that related user behaviour to facets within each trait. Follow-up interview data indicates that the true settings of at-risk group was indeed similar to those reported in the survey, but also that a significant proportion of participants were not aware of settings available to protect their privacy and so did not make use of them. Finally, a set of recommendations in relation to privacy settings has been constructed, in addition to detailed guidance of how to behave safely on the case study online social network (Facebook). A set of recommended actions for Facebook themselves are also included.