Maps

On the Impact of Real-Time Feedback on Users' Behaviour In Mobile Location-Sharing Applications

Description
On the Impact of Real-Time Feedback on Users' Behaviour In Mobile Location-Sharing Applications
Categories
Published
of 13
8
Categories
Published
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Similar Documents
Share
Transcript
  Open Research Online The Open University’s repository of research publicationsand other research outputs On the impact of real-time feedback on users’ behaviourin mobile location-sharing applications Conference Item How to cite: Jedrzejczyk, Lukasz; Price, Blaine A.; Bandara, Arosha K. and Nuseibeh, Bashar (2010). On the im-pact of real-time feedback on users’ behaviour in mobile location-sharing applications. In: SOUPS ’10:Proceedings of the Sixth Symposium on Usable Privacy and Security, 14-16 July 2010.For guidance on citations seeFAQs.c  2010 ACMVersion: Accepted ManuscriptLink(s) to article on publisher’s website:http://dx.doi.org/doi:10.1145/1837110.1837129http://doi.acm.org/10.1145/1837110.1837129Copyright and Moral Rights for the articles on this site are retained by the individual authors and/or other copy-right owners. For more information on Open Research Online’s datapolicyon reuse of materials please consultthe policies page. oro.open.ac.uk  On The Impact of Real-Time Feedback on Users'Behaviour in Mobile Location-Sharing Applications   ABSTRACT  Effective privacy management requires that mobile systems‟ usersbe able to make informed privacy decisions as their experienceand knowledge of a system progresses. Prior work has shown thatmaking such privacy decisions is a difficult task for users becausesystems do not provide support for awareness , visibility and accountability when sharing privacy-sensitive information. Thispaper reports results of our investigation into the efficacy of real-time feedback as a mechanism for incorporating these features of  social translucence in location-sharing applications, in order tohelp users make better privacy decisions. We explored the role of real-time feedback in the context of Buddy Tracker, a mobilelocation-sharing application. Our work focuses on ways in whichreal- time feedback affects people‟s behaviour in order to identify the main criteria for acceptance of this technology. Based on thedata from a three week field trial of Buddy Tracker, a focus groupsession, and interviews, we found that when using a system thatprovided real-time feedback, people were more accountable fortheir actions and reduced the number of unreasonable locationrequests. We have used the results of our study to propose high-level design criteria for incorporating real-time feedback intoinformation sharing applications in a manner that ensures socialacceptance of the technology. Categories and Subject Descriptors  H.5.2 [ User Interfaces ]: Evaluation/methodology, Graphical userinterfaces (GUI), User-centered design. H.3.4 [ Systems andSoftware ]: Current awareness systems, user profiles and alertservices.   General Terms  Design, Experimentation, Human Factors. Keywords  Feedback, mobile computing, location based services, privacymanagement, social translucence. 1.   INTRODUCTION Many studies have shown that users are not very good atunderstanding the future value of keeping personal informationprivate [1,2]. Some solutions involving location privacy policieshave been suggested (e.g., [3]). However, prior research showsthat end-users have difficulties in expressing and setting theirprivacy preferences, and their privacy policies change onlymarginally, unless they are given privacy tools that help themunderstand future implications of their privacy-related choices[2,4]. Moreover, setting privacy rules is a time-consumingprocess, which many people are unwilling to do until their privacyis violated. We see this as a strong motivation to design tools thathelp users to make informed privacy decisions as their experienceand knowledge of a system progress. In the spirit of Palen andDourish [5], we propose to build privacy-sensitive systemssupporting the continual and selective disclosure of personalinformation by providing real-time    feedback  as the method of informing users about how their location information is beingused. In our work we define feedback to be the notification of information disclosure, where the notification specifies whatinformation about the person is disclosed when and to whom. Thisdefinition is drawn from the work of Bellotti and Sellen who considered feedback as “ informing people when and what information about them is being captured and to whom theinformation is being made available ” [6]. While such feedback-oriented support for privacy management has been studied onconventional (large screen) computers [7,8], innovations areneeded for mobile devices. Therefore we decided to explore therole of real-time feedback in managing privacy in mobilelocation-sharing applications.In this paper we present a location-sharing service grounded onthe concept of  socially translucent systems proposed by Ericksonand Kellogg [9]. Translucency is achieved by real-time feedback providing awareness and visibility in the form of ad-hoc warningsdisplayed on, or generated by, the mobile device. Obviously, information about who has accessed one‟s location information might have a positive effect, i.e. improving the comfort of usinglocation-sharing technologies, openness or fewer privacy concerns[7]. However, our preliminary results suggest that this is notalways the case, especially from the perspective of datarequesters. Our field trial with the real-time feedback featureshowed that people were more accountable for their actions if they knew that the data owner would be notified of their request.This supports the third characteristic of  socially translucent systems : accountability . Providing feedback to those whoselocation is being checked resulted in better awareness andunderstanding of the location requests made by others. Thisresulted in location requests being made only when the requesterhas good reason to do so.We explore the role of real-time feedback in privacy managementin the context of Buddy Tracker, a mobile location-sharing servicewe developed, by asking the following questions: Copyright is held by the author/owner. Permission to make digital or hardcopies of all or part of this work for personal or classroom use is grantedwithout fee.Symposium On Usable Privacy and Security (SOUPS) 2010, July 14-16,2010, Redmond, WA, USA.   Lukasz Jedrzejczyk*, Blaine A. Price*, Arosha K. Bandara*, Bashar Nuseibeh* †   *Department of Computing, The Open University, Milton Keynes MK7 6AA, UK{L.Jedrzejczyk; B.A.Price; A.K.Bandara; B.Nuseibeh}@open.ac.uk † Lero, University of Limerick, Republic of Ireland  1.   What is the impact of real- time feedback on users‟  behaviour?   We investigated users‟ reactions to thistechnology and how it affects users‟ behaviour. 2.   What are end-users ‟  criteria for socially accepted real-time feedback system ? We were interested in how to build acontext-aware real-time feedback manager system for supporting awareness that meets users‟ needs.  The next section of the paper discusses related work relevant tothe concept of feedback (Section 2), followed by our classificationof feedback mechanisms for mobile applications (Section 3).Next, we present the technical details of the Buddy Trackersystem that we built for the purpose of our research (Section 4).This is followed by the presentation of our findings from a focusgroup session, interviews, and the field trial of Buddy Trackerwith real-time feedback feature. Section 6 describes some high-level design guidelines for incorporating real-time feedback intosystems, which we have developed on the basis of ourexperimental results. Finally, Section 7 summarizes our resultsand describes the most pressing research issues related to real-time feedback, which form the basis of our future work. 2.   RELATED WORK Feedback can be viewed from different perspectives, such assupporting machine learning [10], maintaining privacy policies orsupporting collaborative work [6,11]. In this paper, we considerfeedback from the perspective of sensory representation forpersonal privacy support.Previous work on feedback has produced a wide range of resultsfor different contexts and activities. Bellotti and Sellen [6] studiedthe use of feedback to show users of the RAVE environment thatthey were being recorded. They found that feedback in the form of an LED light is a good communication tool but that displaying thefull information about people watching is too intrusive. To provide a “just -in- time” descriptions of who is requesting information and why, Hong proposed the concept of  accessnotifications represented as a dialog window with additionalcontrols for accepting, denying or ignoring request.  Accessnotifications support plausible deniability and also act as a privacymanagement tool [12]. Sellen et al. [13] proposed a novel designfor a situated device, The Whereabouts Clock  , presenting real-time info rmation of “what the group is up to” based on a fictionaldevice described in J.K. Rowling‟s  Harry Potter  books.Another attempt to provide feedback about location requests waspresented by Sadeh et al. [4]. They proposed a design for bothreal-time and aggregated feedback mechanisms. The first was abubble notification (as used in the Windows Operating Systems);and the second was a location request history list, showing whohad access to what information and when. The bubble was foundas a minimally disruptive method for supporting awareness, whichis one of the goals of our research. A similar interface waspresented by Lederer et al. [8] who also designed an interface fora disclosure log to help people understand their privacy policies.A shortcoming of the latter design is that it does not provide amechanism for making suggestions and refining privacypreferences ad-hoc. Tsai et al. [7] proposed a similar interface fora disclosure log. They found that feedback is a very important design feature supporting user‟s acceptance of location -sharingtechnologies and improving the comfort of sharing location. Theyalso highlighted the correlation between the feedback availabilityand openness.Raento and Oulasvirta [14] discuss the concept of historicalfeedback in the form of disclosure logs for location informationon mobile devices. Their interface provides both coarse-grainedlocation request information, and fine-grained view available ondemand.The work described so far has focused strongly on visualfeedback, which is not appropriate in all contexts. Previous work in using feedback for privacy support also does not address theissue of contextual real-time feedback, and using alternativesensory representations (not only visual) for supporting privacy.For example, vibro-tactile and auditory feedback has been usedsuccessfully in other domains such as mobile search [15],navigation [16,17] or supporting visually impaired people inreading graphs activities [18]. We are interested in supporting visibility , awareness and accountability by extending thetraditional ways of communication through new interactions. 3.   FEEDBACK CLASSIFICATION Our work seeks to find appropriate location privacy feedback mechanisms for a variety of contexts. We have designed a modelfor studying the role of feedback in location privacy managementby classifying feedback along three dimensions: sensory,interaction and time. Consider the following example contextscenarios: SCENARIO 1:    Alice and Bob are users of the Buddy Tracker application. Bob checks o n Alice‟s location when she is giving a presentation in a meeting. A blue LED light on her phone started  flashing when she was presenting her slides (the blue light indicates that someone is checking on e‟s location). She glanced her phone and after the meeting Alice checked who was checkingher location.   SCENARIO 2:  Alice is playing the favorite game on her mobile phone. While she was playing, a warning pop-up appeared sayingthat  „   Bob  just checked her location‟  . The game paused. She felt very annoyed as she lost her place in the game due to the alert.  These sample scenarios show a positive and negative example of how we can incorporate the real-time feedback within thespectrum of mobile privacy interaction. They also show howfeedback can be delivered, describe the time when information isdelivered, and also what triggers delivery of real-time notification.In order to support our studies on feedback in privacymanagement we have distinguished the following three feedback dimensions. 3.1   Sensory Dimension The sensory dimension (S) relates to the feedback representation,describing how information will be communicated to users. Wehave identified three subgroups of the sensory dimension:    Auditory feedback S(A) describes any audio interactionbetween the system and the user, which has been recognizedas an intuitive and unobtrusive medium for communication[11]. It can be as simple as a distinct musical tone playingwhen the event occurs or it can incorporate fully descriptivenatural language feedback.     Visual feedback S(V) relates to any visual element or featureon a mobile device that supports interaction including GUIelements used in ad-hoc communication. It can be used torepresent the current state of the system, and also to displayaggregated information based on historical data, i.e. icons,warnings, dialog boxes, privacy critics [19], disclosure logs[4,8,7], or map visualizations. Visual feedback can be alsorepresented via hardware features, which relates to any visualfeature of mobile device design that can be managedprogrammatically and used for communication (e.g. the LEDlight in HTC G1 Android phone 1 ).    Tactile feedback S(T) describes the vibro-tactile interactionbetween the system and the user such as the phone vibratingwhen an event occurs.   3.2   Interaction Dimension The interaction dimension (I) describes how sensoryrepresentation of feedback is triggered. Feedback can be releasedautomatically or on demand.    Automatic feedback I(A)   is released without user‟s intervention, every time the event occurs. Example: as soon asBob checks the current position of Alice her phoneimmediately vibrates and plays a sound.    On demand feedback I(OD) refers to a manual request madeby the user, e.g. Bob shakes his phone to display a list of friends that accessed his location within last hour, or hechooses a menu option to list everyone . 3.3   Time Dimension The time dimension (T) describes the temporal freshness of theinformation communicated via feedback mechanismscharacterized by the sensory dimension. It can be divided into twocategories:    Real-Time Feedback T(RT)   is designed to support users‟ awareness and visibility by providing timely information.    Aggregated feedback T(A) relates to any aggregatedinformation based on historical data from disclosure logs. 3.4   Mobile Interface Elements for Feedback Below we describe examples of mobile interface elements thatcould be used to provide feedback. Each interface elementsupports different sensory representations of real-time feedback,including real-time and aggregated information delivery throughautomatic interaction.    Dialog box    –  pop-up like window provides controls forspecifying privacy choices. When the dialog box is open theuser can not perform any action until it is closed (Figure 1a).    Toast  –  small pop-up displaying few lines of text in thebottom of the screen which disappears automatically after 2seconds. It does not prevent user from using the phone.    Notification bar  –  notification on the status bar (top part of the screen), adds an icon indicating type of event, with anoptional ticker-text message. It does not prevent user fromusing the phone (Figure 1b). 1 http://www.htc.com/www/product/g1/overview.html Figure 1. Selected visual representations of real-time feedbackinterfaces.      LED Light    –  flashing LED light, in Buddy Tracker blue lightmeans that s omeone is checking user‟s location (hardware specific).    Flashlight    –  screen flashes a few times and then goes back tothe previous state.    Vibration    –  special pattern indicates a location-checkingevent.    Sound    –  feedback is represented as a distinct musical toneplaying when the event occurs or it can incorporate fullydescriptive natural language feedback, e.g. playing synthesized or recorded speech: “  Bob is checking your location .”      Security alert  –  special type of dialog box incorporating mapvisualization to convey richer feedback information. This typeof visual feedback is used to display aggregated information in the event of unusual events, i.e. user X has checked Y‟s location 50 times in last two days (Figure 1c). This type of representation can be used to present both real-time andaggregated information.In this section we presented a classification of feedback that weuse in our research model for studying the role of feedback inprivacy management for mobile location-sharing applications.We also presented example interfaces supporting our feedback classification. The work presented in this paper does not cover allpossible variations of interaction, time and sensory dimensions.Here, we focus only on exploring implications of real-timefeedback through automatic interaction on users‟ beha viour inBuddy Tracker, mobile location sharing-application. 4.   BUDDY TRACKER ARCHITECTURE Figure 2 illustrates our architecture for Buddy Tracker, a mobilelocation-sharing service that supports the three characteristics of  social translucence : visibility , awareness and accountability byincorporating real-time feedback. Our application provides severaloptions for managing privacy, which are described in section 4.3.Firstly we briefly describe technical details of Buddy Tracker andthen present the functionality of the application.    Figure 2. Buddy Tracker architecture.   4.1   Technical Details We combined several separate services in our design, allowing usto develop prototypes quickly, deploy them automatically, andupdate services for users in the field without user intervention.Our application uses the Navizon 2 service which updates ourserver every 10 minutes using the most accurate positioningsystem visible to the device at the time: GPS, Wi-fi, or cell-id. Wefound this service on the Apple iPhone to be the most accurate,easiest to set up, and had the best power economy of all of hardware and services we considered. 4.1.1    Buddy Tracker Client  The Buddy Tracker client application is implemented as a webapplication, which appears and functions much like a nativeapplication on the iPhone, using the jQTouch library 3 . Theinterface can be also used on other mobile devices which supportWebKit 4 engine for rendering web pages, such as Google Androidpowered phones. This allowed us to activate and deactivatefeatures instantly by changing the files on the server. It alsoallowed us to monitor usage of the system in order to send usersinstant experience sampling requests and to send real-timefeedback to people whose location had just been viewed (a featureabsent on all the other mobile location sharing services weconsidered).A user of the client application (U1) sends a request to view thelocation of a fellow user (U2) to the Buddy Tracker server. The server generates a response containing U2‟s location information and sends it to U1. Additionally, the server generates a feedback response, which is sent to U2, informing them that U1 viewedtheir location. Both the data requester (U1) and data owner (U2)are users of Buddy Tracker client application. The diagram alsoshows the integration with Navizon server, which is being used asa positioning service in Buddy Tracker architecture.   4.1.2    Buddy Tracker Server  The server implements three modules (Security Manager, PrivacyManager and Real-Time Feedback Manager), and uses four datarepositories (Users Information, Location Information PrivacyPolicy Repository and Query Log). The User Informationrepository contains information about users, such as their name,login, and password. The Location Information repository stores the users‟ positioning data as triple: time, location and user  2 http://www.navizon.com 3 http://www.jqtouch.com 4 http://en.wikipedia.org/wiki/WebKit reference. Users‟ privacy preferences and real -time feedback preferences are stored in a Privacy Policy Repository and theQuery Log contains information about location requests. This lastrepository is used by the aggregated feedback module provided inBuddy Tracker (Figure 5) to enable users to view who hadaccessed their location in the past.We will now explain functionality of Buddy Tracker modules byillustrating an example location request, in which one user looksup location of another user in Buddy Tracker.The first module that takes part in that request is the SecurityManager; it is responsible for each user‟s authentication. After  asuccessful check of a user‟s details in the Users Informationrepository, the location query is forwarded to the Privacy PolicyRepository which analyzes the data owner‟s privacy policy . Thes ystem sends a response to the user based on requester‟s detailsand data owner‟s privacy policy. Information about the location query (data requester, data owner, location, granularity level of disclosed location) is then forwarded to the Real-Time Feedback Manager. The Real-Time Feedback Manager first checks the data owner‟s preferences for real -time feedback and then sends thefeedback notification based on that information. Secondly, theReal-Time Feedback Manager saves the location requestinformation in a Query Log for future reference. 4.1.3   Positioning service  –  Navizon Server  Buddy Tracker uses Navizon for user‟s location positioning,which provides a user‟s current location information. It is a third party service; therefore we had to develop a connector thatintegrates the Users Location repository with Navizon ‟s database.  Navizon is configured to update the user‟s position in it‟s server  repository every 10 minutes. The Buddy Tracker server sends arequest to the Navizon service at the same frequency and retrieves an XML file containing the user‟s location information.   4.2   Social Translucence in Buddy Tracker Our main objective when designing Buddy Tracker was to support the data owner‟s privacy. To this end we have created a system that helps peopl e understand each other‟s actions with respect to their privacy and social relationships . Buddy Tracker‟s architecture is grounded on the concept of  social translucence,which has been highlighted as a method supporting awareness, a shared knowledge that enforces accountability by making  people‟s actions visible one to another. Figure 3. Social Translucence in Buddy Tracker.  
Search
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x