Do People Protect Their Privacy on Facebook? Measuring Facebook Privacy Concerns Indirectly Group Members Yousra Javed Hakim Touati Melissa Lennon Agenda • • • • • Background Problem Statement & Motivation Methodology Results Conclusion & Future Work Background Privacy is a very emotionally charged topic for most people. Invoking this emotional response by explicitly mentioning content sensitivity leads people to exaggerate their concerns To reliably research privacy concerns, study questions should be neutral enough to elicit responses that reflect the true values and potential actions of the survey takers. Facebook is the ideal context for this subject of study because the ongoing privacy concerns do not seem to make much of an impression on its millions of users. How do Facebook users truly feel about their privacy? Background (Prior Research) Debatin et al.[3] explored user attitudes toward social networks such as facebook in their paper Facebook and online privacy: attitudes, behavior and unintended consequences . Liu et al.[5] explored the users expectation vs. the reality when analyzing privacy settings. Braunstein et al.[1] addresses the fact that privacy is a very emotionally charged topic for most people. Specifically mentioning content sensitivity invites exaggerated reporting of concerns. Their paper Indirect Content Privacy Surveys introduced the indirect survey method but studied generic objects. Problem Statement & Motivation Investigate the privacy language affect on Facebook subjects through indirect questioning Indirect survey questions diminish any emotional response because they do not explicitly mention privacy and security. We believe Facebook users are less private than what they state. This might be due to: § False perceptions of privacy § Lack of awareness about Facebook privacy issues § Users' attitude toward social networks Methodology • • • • • Initial Survey Compilation Pilot Study Survey Revision Based on Feedback Deployment and Data Collection Data Analysis Methodology • • • • • Initial Survey Compilation Indirect Survey Q 1 and 2 : Frequency of use Pilot Study Q 3 and 4 : Frequency of sharing Survey Revision Based on Feedback Q 5 : Potential for embarrassment upon content exposure Q 6 : Importance to the user Deployment and Data Collection Q 7 : Content ranking based on Data Analysis importance Direct Survey Q 1 : Potential financial loss Q 2 : Potential Embarrassment Q 3 : Ease of access to data + Demographics and Westin Questions Methodology • • • • • Participants: Students of this class J Initial Survey Compilation Indirect Survey (First) and Pilot Study Direct Survey afterwards Survey Revision Based on Feedback One week gap Deployment and Data Collection Results: Lots of comments!!! Data Analysis Methodology Revised the questions based on suggestions Categorized direct questions Merged the surveys into one • • • • • Three parts: Facebook usage Facebook privacy Demographics &Westin Survey Questions Initial Survey Compilation Pilot Study Survey Revision Based on Feedback Deployment and Data Collection Data Analysis Methodology • • • • • Initial Survey Compilation Pilot Study Survey Revision Based on Feedback Deployment and Data Collection Data Analysis Link shared on Facebook and through emails 34 users completed the 40 question survey Methodology • • • • • Initial Survey Compilation Pilot Study Survey Revision Based on Feedback Deployment and Data Collection Data Analysis Data Analysis • Comparison between the responses of the direct and indirect questions for each group of related questions • Frequency of use • Frequency of sharing • Importance to the user • Content-type ranking based on its direct and indirect privacy score • Indirect Pscore = 1/4( A1+A2+A3+A6)+A4+1/2(A5+A6) [1] • Direct Pscore = 1/2(A3+A4)+ 1/4(A1+A2+A3+A4) • Subject classification using Westin s index criteria and responses of direct and indirect survey [2] Results User Demographics !"#$% !"#$"%& 01"& ,(;"899<&-"-8"%=>?@& E$F;(B?9#& ! &'()*(+% '()"& ,"-()"& *23.4& .6357& +/&(#$&089:"& A"==&B>(#&*&C"(%& *&B9&.&C"(%=& .&B9&5&C"(%=& 5&B9&+&C"(%=& '9%"&B>(#&+&C"(%=& A"==&B>(#&G?1>&H;>99)& .&C"(%=&9I&;9))"1"& +&C"(%=&9I&;9))"1"& '9%"&B>(#&+&C"(%=&9I&;9))"1"& ,-.%-/%01$21% *+& ./& *5& *6& 4& *& +& D& **& **& /& /& */& .+& Results Cont d… User Classification Class Westin Indirect Direct Fundamentalist 2.94% 0% 0% Pragmatist 94.11 100% 85.2% Unconcerned 2.94% 0% 14.8% Results Cont d… Content-type ranking based on its direct and indirect privacyRanking score of content types based on privacy score Indirect Questions Direct Questions User stated Ranking Photos, wall posts, status updates (14 users) Photos, status updates, wall posts (10 users) Privacy Score 1.25 0.25 Photos Status Updates Content Type Wall Posts Results Cont d… Comparison between the responses of the direct and indirect questions for each group of related questions 5 Average Average reported frequency of use for each content type Reported Frequency of Use for each Facebook Content Type 0 1 2 Rating 3 4 Indirect Direct Photos Status Updates Content Type Wall Posts Content Type Indirect µ Direct µ P-Value Photos 4.147059 2.117647 2.413e-09 Status updates Wall posts 3.882353 2.352941 1.110e-06 3.735294 2.235294 6.681e-07 Results Cont d… Comparison between the responses of the direct and indirect questions for each group of related questions Average reported frequency of sharing for each content type 5 Average Reported Frequency of Sharing for each Facebook Content Type 0 1 2 Rating 3 4 Indirect Direct Photos Status Updates Content Type Wall Posts Content Type Indirect µ Direct µ P-Value Photos 4.823529 2.823529 7.317e-10 Status updates Wall posts 4.088235 2.441176 6.063e-07 4.058824 2.735294 3.671e-05 Results Cont d… Comparison between the responses of the direct and indirect questions for each group of related questions 5 AverageAverage reported ofFacebook eachContent content type Reportedimportance Importance for each Type 3 Indirect µ Direct µ P-Value Photos 4.176471 1.411765 2.2e-16 Status updates Wall posts 2.470588 1.705882 0.01650 3.000000 1.617647 2.572e-05 0 1 Rating Content Type 2 4 Indirect Direct Photos Status Updates Content Type Wall Posts Conclusions & Future Work • What we accomplished • • • • • • Built a pilot survey to evaluate users privacy concerns and privacy wording effect. Refined the survey questions. We analyzed the results of our survey. Users have less perceived usage and sharing habits than in reality. Why it is hard to evaluate privacy importance in SN? • • Because it is relative to general privacy. Because it is relative to content desirability and importance. Future work will consist of • • Refining the survey questions and adapting them to Facebook and real word content. Study the relative privacy importance when compared to real world importance. References [1] Alex Braunstein, Laura Granka, Jessica Staddon, Indirect Content Privacy Surveys: Measuring Privacy Without Asking About It , Symposium on Usable Privacy and Security (SOUPS), 2011 [2] Ponnurangam Kumaraguru, Lorrie Faith Cranor, Privacy Indexes: A Survey of Westin s Studies http://www.casos.cs.cmu.edu/publications/papers/CMU-ISRI-05-138.pdf [3] Bernhard Debatin, Jennette P. Lovejoy, Ann-Kathrin Horn M.A., Brittany N. Hughes, Facebook and Online Privacy: Attitudes, Behaviors, and Unintended Consequences http://onlinelibrary.wiley.com/doi/10.1111/j.1083-6101.2009.01494.x/ full [4] Emily Christofides, M.Sc., Amy Muise, M.Sc., and Serge Desmarais, Ph.D, Information Disclosure and Control on Facebook: Are They Two Sides of the Same Coin or Two Different Processes? , [5] Yabing Liu, Krishna P. Gummadi, Balachander Krishnamurthy, Alan Mislove, Analyzing Facebook Privacy Settings: User Expectations vs. Reality http://conferences.sigcomm.org/imc/2011/docs/p61.pdf Questions
© Copyright 2025 Paperzz