You may notice some images loading slow across the Open Collections website. Thank you for your patience as we rebuild the cache to make images load faster.

Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Towards understanding how users decide about friendship requests in Online Social Networks Rashtian, Hootan 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


24-ubc_2014_september_rashtian_hootan.pdf [ 3.15MB ]
JSON: 24-1.0167341.json
JSON-LD: 24-1.0167341-ld.json
RDF/XML (Pretty): 24-1.0167341-rdf.xml
RDF/JSON: 24-1.0167341-rdf.json
Turtle: 24-1.0167341-turtle.txt
N-Triples: 24-1.0167341-rdf-ntriples.txt
Original Record: 24-1.0167341-source.json
Full Text

Full Text

Towards understanding how users decide about friendshiprequests in Online Social NetworksbyHootan RashtianB.Sc in Computer Engineering, Isfahan University, 2011A THESIS SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THE DEGREE OFMaster of Applied ScienceinTHE FACULTY OF GRADUATE AND POSTDOCTORALSTUDIES(Electrical and Computer Engineering)The University Of British Columbia(Vancouver)June 2014c© Hootan Rashtian, 2014AbstractAccepting friend requests from strangers in Facebook-like online social networksis known to be a risky behavior. Still, empirical evidence suggests that Facebookusers often accept such requests with high rate. As a first step towards technologysupport of users in their decisions about friend requests, we investigate why usersaccept such requests. We conducted two studies of users’ befriending behavior onFacebook. Based on 20 interviews with active Facebook users, we developed afriend request acceptance model that explains how various factors influence useracceptance behavior. To test and refine our model, we also conducted a quantita-tive study with 397 participants using Amazon Mechanical Turk. We found thatfour factors significantly impact the receiver’s decision towards requests sent fromstrangers, namely, knowing the requester’s in real world, having common hobbiesor interests, having mutual friends, and the closeness of mutual friends. Based onour findings, we offer design recommendations for improving the usability of thecorresponding user interfaces in order to help users make more informed decisions.iiPrefaceThe chapter 4 and 5 of this thesis have been published. The author of this thesisperformed the users studies presented in chapter 4 and chapter 5. He also ana-lyzed the data from those studies. He authored the corresponding paper, underthe supervision of Dr. Konstantin Beznosov who provided feedback and guidancethroughout the research process. Below is the details of the published paper:• Hootan Rashtian, Yazan Boshmaf, Pooya Jaferian, Konstantin Beznosov.(2014,July). To Befriend Or Not? A Model of Friend Request Acceptance on Face-book. In Proceedings of the 10th symposium on Usable Privacy and Security.ACM.Two user studies were conducted as part of this research. For the first study(explained in chapter 3), we submitted a human ethics application with the BREBnumber of H13-01452 to UBC Behavioural Research Ethics Board. For the secondstudy (explained in chapter 4), we submitted an amendment (with the same BREBnumber) to the first study application. The ethics application and its amendmentwere approved by UBC Behavioural Research Ethics Board.iiiTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.1 Overview of friendship acceptance in social science literature . . . 42.2 Overview of friendship acceptance in online social networks liter-ature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.1 Design of the qualitative study . . . . . . . . . . . . . . . . . . . 12iv3.1.1 What is Grounded Theory? . . . . . . . . . . . . . . . . . 133.1.2 Different versions of Grounded Theory . . . . . . . . . . 133.2 Design of the quantitative study . . . . . . . . . . . . . . . . . . 144 Exploratory Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164.1 Grounded Theory approach . . . . . . . . . . . . . . . . . . . . . 164.2 Research questions . . . . . . . . . . . . . . . . . . . . . . . . . 164.3 Participant recruitment . . . . . . . . . . . . . . . . . . . . . . . 174.4 Data collection . . . . . . . . . . . . . . . . . . . . . . . . . . . 204.5 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224.6 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405 Quantitative Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475.1 Research question . . . . . . . . . . . . . . . . . . . . . . . . . . 475.2 Survey execution . . . . . . . . . . . . . . . . . . . . . . . . . . 485.3 Sample representativeness . . . . . . . . . . . . . . . . . . . . . 495.3.1 Age comparison . . . . . . . . . . . . . . . . . . . . . . 495.3.2 Gender comparison . . . . . . . . . . . . . . . . . . . . . 495.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 696 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 726.1 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 766.2 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 767 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79vBibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82A Interview Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . 89B Survey Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98viList of TablesTable 4.1 Demographics of interview participants . . . . . . . . . . . . . 24viiList of FiguresFigure 4.1 Mediator role . . . . . . . . . . . . . . . . . . . . . . . . . . 18Figure 4.2 Volunteer role . . . . . . . . . . . . . . . . . . . . . . . . . . 19Figure 4.3 Theoretical saturation of interview data . . . . . . . . . . . . 22Figure 4.4 Online lifecycle of Facebook friend acceptance (OLFFA) model.Shaded components on the top are the internal factors andcomponents with hyphenated borders are the external factors.The middle box, which includes 3 components, represents thedecision making process. The dashed arrows represent deci-sion making flow. The solid arrows represent the impact ofcomponents on each other. . . . . . . . . . . . . . . . . . . . 25Figure 5.1 Comparison of our sample to Facebook population in terms ofage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50Figure 5.2 Comparison of our sample to Facebook population in terms ofgender. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50Figure 5.3 Distribution of the sample based on the levels of education. . . 51viiiFigure 5.4 Distribution of the sample based on different employment sta-tuses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51Figure 5.5 Distribution of the sample based on length of membership onFacebook. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52Figure 5.6 Distribution of the sample based on frequency of Facebookusage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52Figure 5.7 Distribution of number of Facebook friends among participants. 54Figure 5.8 Distribution of frequency for receiving friend request. . . . . . 55Figure 5.9 Employment of friendship factors by participants. . . . . . . . 56Figure 5.10 Four groups discussed in the analyses. . . . . . . . . . . . . . 58Figure 5.11 Comparison of friendship factors employment between G1 andG2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60Figure 5.12 Comparison of friendship factors employment between G1 andG3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63Figure 5.13 Comparison of friendship factors employment between G1 andG4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Figure 5.14 Comparison of friendship factors employment between G2 andG3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67Figure 5.15 Comparison of friendship factors employment between G2 andG4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68Figure 5.16 Comparison of friendship factors employment between G3 andG4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70ixAcknowledgmentsI would like to offer my gratitude to my academic supervisor, Dr. Konstantin(Kosta) Beznosov, for his support and mentorship during my M.Sc.I would like to thank Yazan Boshmaf and Pooya Jaferian for their constructivefeedback throughout the project. Also, I would like to thank my friends at theLaboratory for Education and Research in Secure Systems (LERSSE) who havebeen great friends since I joined LERSSE.Lastly, I would like to thank my family, my parents in particular, for the sup-port and encouragement through my entire life. Indeed, words are not capable ofconveying my love to them.xTo my beloved parents, and my dearbrother.xiChapter 1IntroductionUsers of Facebook-like online social networks (FOSN) are not careful when ac-cepting friend requests from strangers, i.e., those who they do not know in reallife or online communities [13, 51]. This behavior can be exploited by an attackerto run an infiltration campaign in a target FOSN [17]. Such malicious campaignsare a growing cyber-security threat [22], where an attacker controls a set of useraccounts and exploits them to befriend a large number of benign users.Large-scale infiltration has three alarming implications [17]: First, the socialgraph of the target FOSN is compromised and polluted with a large number ofnon-genuine social relationships. This means that third-party services and web-sites have to perform appropriate “cleaning” to mask out fake accounts and theirrelationships before integrating with or using such a FOSN. Second, and otherthan online surveillance, the attacker can breach the privacy of users and collectlarge amounts of personally identifying information (PII), such as email addresses,phone numbers and birthdates, which have considerable monetary value in the In-ternet underground markets [16]. This could happen by having access to PII after1making connections with other accounts. In addition, this information can be usedto run follow up, highly personalized e-mail spam and phishing campaigns [39].Third, the attacker can exploit the infiltrated FOSN to spread misinformation as aform of political astroturfing [56], or even influence algorithmic trading that usesopinions extracted from FOSNs to predict stock markets [12, 15].Preventing large-scale infiltration, or at least limiting its scale and impact, isimportant not only to users but also to FOSN operators and social media-basedbusinesses. Improved technology support for FOSN users in helping them to makebetter decisions in regards to friend requests is expected to reduce the associatedrisk. This, however, requires a better understanding of user’s befriending behaviorin FOSNs, particularly what makes them to accept or decline friendship requests.Our research bridges this knowledge gap. In particular, we aim to answer thefollowing general research question: Why do FOSN users accept friend requestsfrom strangers? In our studies, we focused on the scenario where a FOSN userreceives a friend request from another, a stranger in particular, and investigated thefactors that influence the user’s decision on whether to accept this request. More-over, we also studied the process that users go through, when accepting friendrequests, including identity verification, new friend management, and privacy set-tings updates.In order to understand users’ behavior in FOSNs, we designed two studies: aqualitative, exploratory study and a quantitative, confirmatory study. We receivedan approval for both studies from our university’s research ethics board.First, we conducted a set of semi-structured interviews with 20 active Facebookusers (Section 4). The goal of conducting this exploratory study was to understandusers’ behavior in FOSNs in response to friend requests, and explore the factors2that influence their decisions. To the best of our knowledge, there is no relatedqualitative work to support our research questions. Therefore, we used GroundedTheory [21] in our exploration to develop a model that captures such a behavior.In the confirmatory study (Section 5), we refined and partially tested the de-veloped model, by conducting an online survey among 397 Mechanical Turk (M-Turk) workers. The goal was to identify prominent factors that highly impactedusers’ decisions in practice.Based on our findings, we offer recommendations on designing FOSN inter-faces for reviewing and responding to friend requests (Section 6). While defendingagainst large-scale infiltration is challenging [18], we hope that progress in thisresearch direction will lead to the improvement of existing security defenses andmake them less vulnerable to both human exploits (i.e., automated social engineer-ing [38]) and technical exploits (i.e., platform hacks [63]).3Chapter 2Literature ReviewIn this chapter, we aimed to investigate the literature in order to point out the previ-ously work done regarding friendship formation. Since we used Grounded Theoryapproach, we were not allowed to conduct comprehensive literature review in or-der to avoid having any bias in developing the OLFFA model. As the literaturereview, we cover related work in usable security and privacy community as well associal science community. Please note that regarding the novelty of our researchproblem, literature review does not give us the ability to make specific comparisonbetween our work and results with previous work. However, we try to highlight thesimilarities and differences between friendship acceptance in FOSNs and real life.2.1 Overview of friendship acceptance in social scienceliteratureMajority of empirical studies on friendship concentrate on the friendship choicesof individuals [33]. It is also worth to mention that although the context of thesestudies is the real life friendship, the model of friendship that was adopted by them4is similar to the model of friendship in FOSNs where in the first step, somebody(such as a user in FOSN) wants to be friend with another person. Therefore, herequests a friendship and then will wait to receive the answer. For adopting suchmodel, there is theoretic work in the literature, which conceptualizes the friendshipformation as “a series of tentative moves from one person to another one” [33].The previous work also point out to the following interesting point. Basically,exchange theorists believe that an important point regarding the process of friend-ship formation is the analysis of a trade-off between cost and benefit, which meansthat in order to make a decision if having a friendship relationship is too costlyor not [14, 31, 37, 65]. This implies the existence of a decision making process,which we also had in our model. Also, it implicitly indicates that the process offriendship formation includes a request and acceptance/rejection similar to nowa-days friendship formation in FOSNs. Overall, these empirical studies investigatethe distribution of friendship choices across a group or class. In other words, thegoal is to find out if friendship offers are accepted or rejected and how this processtakes place. The process is discussed such that in order to form a friendship, thereceiver should know that there exists a friendship offer. This could happen by di-rect communication or by an indirect or unclear message to avoid embarrassmentof receiving rejection. However, this is not how communication of sending a friendrequest occurs in FOSN as the sender should explicitly send a friend request to thereceiver.Then the next step is to make a decision about the friend offer after recognizingit. As mentioned before, this process is the same as the “decision making” processin our model. However, there are differences in the execution of this process. Thedecision is highly dependent on the benefits that will be obtained after having a new5friendship such as making a connection with somebody who has a higher status inspecific directions such as common interests and social connections. However, theliterature includes work that shows issues in friendship formation of people withunbalanced statuses [14]. For instance, it is usually difficult to ignore the largegap between two people even if there is other attractiveness between the two. Thisis also aligned with some of the friendship factors we have such as knowing theperson in real life (KRL) or closeness of friendship in real life (CFR). Anotherinteresting point during the literature review was that we found the environmentalfactors among the factors that impact the process of decision making about friendoffers. For example, in the case that a person receives many friend offers, timeconstraint makes it difficult to decide about all the offers in the sense that it isnot possible to accept all of them and also thinking properly to make an accuratedecision [34]. In addition to time limitation, there are other factors such as initiatorattractiveness that are considered for deciding about a friend offer. In general, weinterpret these factors as a subset of friendship factors revealed from our interviewstudy.2.2 Overview of friendship acceptance in online socialnetworks literatureThere are many research studies that have been done to examine the Facebookusage by its users. If we want to categorize them based on the subjects it in-cludes a long list of topics such as characteristics of profile elements [26], self-presentation [64], social capital, social grooming, and privacy concerns [32, 41,55]. By reviewing the online social network literature, we could find previouswork related to friendship formation (not particularly friendship acceptance) and6as the result, we highlight the most related points here.An important point from the results of these studies is that they show negativeand positive sides of using online social networks. Examples of negative side arethe possibility of stalking and identity theft while the positive side includes findingnew friends as well as getting in touch with old friends, family, and other users aswell as increasing knowledge and information in different contexts by seeking andsharing information. It is mentioned in the previous work that emergence of theinternet and then online social networks help people to keep up their existing con-nections and relationships [42, 54]. In addition, online social networks have servedas a tool to form romantic connections [50] as well as build up virtual communi-ties [57]. In particular, online social networks such as Orkut [6], Facebook [3],Twitter [8], Pinterest [7], MySpace [5], etc. allow users to create accounts in orderto share their information with other users including their family, relatives, friends,acquaintances, and even people they might not know in real life. In order to initi-ate friendships in online social networks, there are factors that are important to beconsidered. There is also work [46] on finding factors that influence the responserate in social question and answering behavior, which is similar to our study in thesense that it also investigate impact of some factors on a behavior. However, thisstudy was different from ours in many ways including methodology and researchgoal. For example, we found the (friendship) factors based on our exploratorystudy results, which were grounded in our data while they formed a set of factorsfrom the literature.There is also work that discusses some friendship factors. For instance, hav-ing personal information is specified as a fundamental requirement by Parks etal. [54]. However, it is not clear how this information gathering should take place7although other work has pointed out some types of information that is shared bysocial network users such as name, email address, physical address, phone num-ber, academic classification, major, gender, etc [64]. Also, Ambady et al. believesthat visual cues play an important role as they are important for initial impressionsbetween people who are not in a friendship relation [9]. There is also work aboutusage of photos, wall posts, and friend list by Facebook users to attract other usersto their profiles. It is mentioned that these cues would provide information that cangive positive impression to other users [66, 67, 69]. McKenna et al. pointed outthat common interests are good examples of initiating connection on the internet,which confirms our finding regarding having common hobbies and interests as acue for being friend with other users [50]. Another work tried to focus on fac-tors that impact decision of users to friend or de-friend others on Facebook [1].Our results partially confirm the results of this study in the sense of the friendshipfactors that are employed by participants. For example, results of this study showthat knowing someone in real life is the top reason for befriending someone (82%).Also this study shows that users activity (as we call it UAP) and real world interac-tions (we point it out as different factors such as KRL, CMF, CRF) are importantfor befriending. There are also differences in the results. For example, this studypoints out to gender difference as an important factor that plays an important rolein making befriending decisions in the sense that men are mostly use social mediafor dating and networking while women use it for dramatically different purposessuch as getting promotions or giving positive feedback. Consequently, men usersadd friends based on attractiveness or business networks and women add friendsbased on knowing people in real life. While we did not find notable gender dif-ference in our exploratory study, we did not consider it in our second study and8also in our data analysis. Lack of finding data in terms of gender difference couldbe because of different reasons. First, it could be due to lack of representativenessin our sample. We could have found related data points, if we had collected moredata (more interviews). Second, this could be due to the change of users’ behav-ior on Facebook as the aforementioned study was done in 2011, which is ratherold in the fast changes happening in social media. Moreover, we believe that iscould be due to the difference between the goal of our study and this study as wewere mostly focused on behavior of users towards stranger’s request. There is alsoanother work, which is related to unfriending behavior (i.e. removing friendshipconnection) of users in FOSN [61]. According to the results, authors believe thatthere are 4 online factors (unimportant/frequent posts, polarizing posts, inappro-priate posts and everyday life posts) and 2 offline factors (disliked behavior andchanges in the relationship) impact users’ behavior in terms of unfriending. Al-though we faced these factors in our qualitative study, we found that among onlinefactors, users are more sensitive to inappropriate posts for removing their Face-book friends. Regarding the offline factors our qualitative data also agrees withtheir results. Although these studies discuss some friendship factors, but the waythey envisioned the problem was different as they were focusing on specific factorsand they were not including and even investigating other potential factors. Also,these studies have been conducted without considering the potential connections tothe strangers (but we did).While there is work that shows Facebook users use it as a media to maintaintheir relationships with shared social groups (e.g., classmates, high-school friends,elementary school, colleagues) [26, 44, 55, 69], there is also work showing thatusers are interested in making connection with people out of their friends circles9in real life [68], which shows the possibility of users in having connections withstrangers as they might make mistakes in evaluating friend requests that they re-ceive from others. Another study also shows that around 50% of online socialnetworks use these websites to make new friends [44]. This study also reports that47% of teen boys and 28% of teen girls have friends in their profiles that havenever met them before. Furthermore, another study shows that 15% of friend-ships in Facebook is between people who have not met each other in real life [62].This evidence shows possibility and existence of having connections with strangers(somebody who is not know in real life).On the other hand, there is previous work on privacy issues related to the usageof online social networks, which could be resulted from connection with strangers.For instance, companies collect information from social networks for marketinggoals [35], teenagers were raped by people who have been on their Facebook pro-files [10], and teenagers reveal their private information [11]. Another exampleis the work, which shows that users’ intention does not match with their privacysettings [45, 48]. Another study showed that users have difficulty in understand-ing the privacy settings and cannot configure them correctly [25]. These privacyincidents motivate our research problem to understand users’ befriending behavior.As the most related work to ours, Johnson et al. showed that the main concern isinsider’s threat rather than the outsider’s [40]. We believe that focus of our workis different as our concern is to understand user’s behavior towards friendship re-quests rather than how they manage their privacy settings. Moreover, we believethat stranger’s threat still exists as 62% of our sample reported to have at least onestranger in their friendlist.There is work on definition of privacy and digital privacy in particular, to clar-10ify what should be expected by users in terms of privacy [52]. On the other hand,it has been shown that this is not always fault of systems that results in privacy andsecurity issues and humans are a major cause of these failures [58]. Therefore,it is necessary to consider humans in designing systems. Cranor et al. proposeda framework to reason about the human in the process of designing secure sys-tems [24]. This framework was insightful during the process of qualitative dataanalysis to form our model.Finally, we found few studies that were somehow related to the connection ofsocial network users with strangers. For instance, there is a study, which focuseson the willingness of students who have Facebook profiles with their faculty. Asthe results show, at least a third of students believed that faculty should not beon Facebook. Although this result implicitly convey that this could be assumedas a connection with stranger, there are limitation regarding this study includingthe sample size, representativeness of the study as well as the outdated time ofthe research, which is back to 2006 where Facebook was used widely only by bystudents [36].11Chapter 3MethodologyIn order to understand users’ behavior in FOSNs, we chose to have two studies: aqualitative, exploratory study and a quantitative, confirmatory study. We receivedan approval for both studies from our university’s research ethics board. The restof this chapter discusses about each of these two studies with more details.3.1 Design of the qualitative studyThe goal of conducting an exploratory study was to understand users’ behaviorin FOSNs in response to friend requests, and explore the factors that are involvedin this process. Since there is no related qualitative work to support our researchquestions, we used Grounded Theory in our exploration to develop a model thatcaptures such a behavior. As it is suggested by previous work, computer engineer-ing research should focus on evidence-based discipline rather than having advo-cacy based discipline [60]. Therefore, Grounded Theory would be a good choiceas any theories or models are developed according collected evidences or data. Onthe other hand, Grounded Theory has been accounted as the appropriate approach12to answer the question of “What is going on?” or study areas that have not beenstudied [59]. In the case of our project, we wanted to understand how FOSN usersbehave and make decision when they receive a friend request. In addition, wewanted to understand what factors are involved in the process and what factorscould potentially impact users’ decision about requests coming from strangers. Inthe following section, we discuss more details about Grounded Theory and differ-ent versions of it, and justify employing this approach and also the version we usefor this project.3.1.1 What is Grounded Theory?According to the literature, Grounded Theory is an integration of a set of hypothe-ses generated to develop a theory about a substantive area in a systematic way [29].The intuition behind its name (i.e. grounded theory) is that a theory is developedin a systematic data-centered process [27, 30]. Therefore, the theory is groundedin collected data. Glaser at al. point out to the goal of Grounded Theory as anapproach to generate concepts and categories that account for a pattern of behav-ior [28]. However, as Adolph et al. mentioned, mission of Grounded Theory is togenerate a mid-level theory, which describes processes rather than providing a uni-versal truth although it might be different from notion of a theory in mathematicsand engineering.3.1.2 Different versions of Grounded TheoryThere are three versions of Grounded Theory. The first one, which is the originalversion of the method was introduced by Glaser and Strauss in their book “TheDiscovery of Grounded Theory” [27] in 1967. The second version was described13by Strauss and Corbin in the book “Basics of Qualitative Research” in 1990 [23].The latest version was introduced by Charmaz in her book “Constructing GroundedTheory” in 2006 [21]. These three versions of Grounded Theory are still usedby researchers. The choice of the most appropriate version depends on researchproblems and also the context in which research is done.In our case, we decided to choose Charmaz’s version of Grounded Theory as itis more flexible in the sense that it allows researchers to interact with participantsin the process of data creation and analysis as she mentioned in her book [20]:“Data and analysis are created through an interactive process wherebythe researcher and participant construct a shared reality.”Using the Charmaz’s version of Grounded Theory, we ended up with a modelfrom our qualitative study. We provide more details about the qualitative study inchapter 4.3.2 Design of the quantitative studyFor the quantitative part, we planned to test our findings (i.e. OLFFA model andfriendship factors in particular) from the qualitative part on a representative sam-ple and measure the fraction of users who employed those factors. Also, we wereinterested in characterizing the behaviors of FOSN users and figure out how friend-ship factors are used by users. For this aim, we developed and conducted an on-line survey using Amazon Mechanical Turk (M-Turk). There are three reasons forconducting the survey on Amazon M-Turk. First, it is widely adopted by manyresearchers in different fields of research [19, 43, 53]. The second reason is thatprevious work shows: “MTurk participants are slightly more demographically di-14verse than are standard Internet samples and are significantly more diverse thantypical American college samples” [19]. Also, previous work shows that collecteddata from Amazon M-Turk is as reliable as the data collected from traditional meth-ods [19].After defining the survey instrument (i.e. Amazon M-Turk), we designed thesurvey questionnaire. Complementing information about the survey as well as theanalyses and results are available in Chapter 5 of the thesis.15Chapter 4Exploratory StudyThe study was in the form of semi-structured interviews. In what follows, we givemore details about the study, including research questions, recruitment procedure,data collection and analysis.4.1 Grounded Theory approachWe chose Grounded Theory as the approach of this study as it is an appropriatemethod for research in areas that have not been previously explored, especiallywhen a new perspective might be beneficial [59]. Among different ways to ap-ply Grounded Theory [21, 23, 27], we chose to follow the definition proposed byCharmaz [21] because it provides a more flexible format for data analysis.4.2 Research questionsIn the exploratory study, we aimed to understand users’ befriending behaviour inresponse to friend requests, and to explore the factors that impact their decision. Byapplying the procedures of grounded theory coding, we were able to find new in-16formation, concepts, themes, and categories to develop a theoretical model, whichhelps in answering the following research questions:• RQ1: What are the factors that influence users’ when responding to friendrequests in general, and to friend requests sent by strangers in particular?• RQ2: What are the actions the users take before making a decision about afriend request?• RQ3: What are the actions the users take after making a decision about afriend request?4.3 Participant recruitmentWe posted the recruitment notices on local Craiglist and Kijiji websites. We alsodistributed flyers across our university’s campus. In the recruitment notice, weincluded a brief description of the study and a hyper-link to an existing Facebookprofile, and asked potential participants to send a personal message to that profiledescribing their interest, along with their email addresses.We asked potential participants for their email addresses so that we have a re-liable way to communicate urgent messages without depending on Facebook (e.g.,unplanned changes in the interview schedule).The owner of the profile was a graduate student in our department who wasnot affiliated with our research lab and was recruited to mediate the initial commu-nication with potential participants. The purpose of recruiting a third party (i.e.,the mediator) was to avoid any potential linkage between the user profile used forrecruitment and our study. The mediator signed a non-disclosure agreement stat-ing that all data collected through mediation would be immediately erased after17On-campusflyersParticipantOnline AdMediator ResearcherSends contact information via Facebook message Sends participant s contact information1 23 Schedules interviewFigure 4.1: Mediator rolerelaying them to us, and that all information about the study would not be sharedexternally.Overall, the mediator, denoted by M, operated under the following protocol, asillustrated in Figure 4.1:1. A potential participant P uses Facebook to send a personal message to themediator M, which contains P’s email address and interest in the study.2. M sends to the dedicated researcher an email including P’s Facebook useridentifier along with P’s email address.3. Once the researcher receives the email from M, he asks M to permanentlydelete the message that was sent by P and not to respond to any interactionsinitiated by P.Using the email addresses of potential participants, we used e-mail to schedule18Volunteer ParticipantSends a friendship request from real accountSends a friendship request from auxiliary accountInterview happens after 4 days123Figure 4.2: Volunteer roleinterviews with them. We used the mediator to avoid inaccuracies due to self-reporting, when it came to identifying which of our participants tend to acceptfriend requests from strangers. This is why we had another volunteer who sentprospective participants friend requests from two other dedicated Facebook userprofiles. The first user profile was a real account managed by another volunteer,while the second one was an auxiliary account that we created for the purpose ofthis study.1 We aimed at reducing the chances that the participants knew the realaccount. To this end, we excluded students in our department from participating inthe study.As illustrated in Figure 4.2, the volunteer controlled both accounts and sentfriend requests to potential participants according to our instructions. The volun-teer, who was a graduate student from our department but not affiliated with ourresearch lab, signed a non-disclosure agreement that prohibited him from both in-1The auxiliary account represented a male graduate student attending our university. The profileincluded a publicly available, generic picture of a man in his mid 20’s.19teracting with potential participants and sharing any collected information.To avoid any suspicion among the participants in regards to the volunteer’s ac-count, we asked the volunteer to remove Facebook friends made for the purposeof the study after the interviews were finished, rather than before the interviews.While there was a risk of two participants having a pre-existing social connection(either online or offline) and seeing that the one is a friend with the volunteers,which could have influenced the other participant, none of the interviewed partic-ipants indicated that this was the case. After each interview, we sent a debriefingmessage via Facebook to thank the participants for their interest in our study andprovided them with more details about our research.4.4 Data collectionWe began data collection when we started conducting the interviews. Our inter-views were semi-structured, which gave us the flexibility to adjust and add newquestions. We performed data analysis concurrently with the interviews in order toinform each new interview with the results obtained from the previous ones.Each interview followed roughly the interview guide reproduced in Appendix Aand had the following 6 parts:1. Overview of the project.2. Participants’ demographics (e.g., age, gender, education, occupation, lan-guage) and Facebook usage-related questions (e.g., membership time, fre-quency of usage).3. Participants’ befriending behavior in general, and their responses to friendrequests in particular. For instance, we asked questions about participant’s20friends, factors or criteria they employ to make a decisions about friend re-quests.4. Participants’ attitude towards their privacy and security.5. Participants’ attitude towards befriending strangers, and whether they hadbefriended strangers before.6. Debriefing participants and concluding the interview. During this part of theinterview, we also informed them about the friend requests that our volunteersent. We observed each participant’s reaction and asked each participant whoaccepted any of the two requests why they did so. We also asked participantsif they had any suggestions regarding the interface design that might helpthem make more informed decisions.As we mentioned earlier, interviews were in the semi-structured format and weasked questions in addition to the interview guide (Appendix A). For example, weasked questions about the potential difficulty of the user interface or obfuscation ofthe UI.As an iterative process, we analyzed the data by searching for patterns andforming concepts that were gathered into categories. We also wrote memos duringthe process of analysis to capture our understanding about the emerging categoriesand relationships among them.Thanks to the iterative data analysis performed between interviews, we wereable to detect “theoretical saturation” [28]. After 15 interviews, as Figure 4.3shows, we reached the plateau where further data collection did not add new cate-gories.210 5 10 15 200204060Number of participantsNumber of unique codesFigure 4.3: Theoretical saturation of interview dataThis is why we stopped data collection after interviewing 20 participants. Theirdemographics are summarized in Table 4.1. All interviews were conducted in per-son at our university’s campus. Each interview took about 50 minutes on average.4.5 Data analysisAs specified earlier, we employed Grounded Theory for the exploratory study. InGrounded Theory, data analysis involves searching for the concepts behind theanswers. We transcribed, anonymized, and analyzed the collected data after eachinterview with an average turn-around time of 4 days. We used a web applicationtool called Dedoose for the analysis [2]. In what follows, we describe each part ofthe analysis in detail.22Open codingAs the first step of coding, we identified, named, described, and categorized phe-nomena found in the collected data. Open coding resulted in a set of 63 uniquecodes, including both abstract (e.g., befriending behavior) and concrete labels (e.g.,Facebook frequency of use). The intuition behind having abstract labels was to helpdevelop a model. At the end, we had in total 2,620 coded excerpts, with an averageof 131 per interview. We performed triangulation by having two other coders onfour of the interview transcripts (interviews numbers 2, 6, 8, 11). The codes gen-erated by the other two coders turned out to be subsets of codes generated by themain coder. The reason was that we reached saturation after 14 interviews and wewanted to make sure about this issue. Although we did not find new themes fromthe triangulation, we kept collecting data for another 6 interviews and we stoppeddata collection as we did not find new findings during data analysis.Axial codingAfter open coding, we started to relate the generated codes to each other and endedup with 7 categories grounded in the collected data. The categories are friend-ship factors, privacy and security awareness or concerns, investigation actions,decision execution, maintenance actions, environmental factors, and interface ca-pabilities.Selective codingThe aim of selective coding was twofold: (1) to identify the main category, whichended up being decision making process for friend requests; and (2) discarded allcategories that were not related to the core category, e.g., fancy interface features.23Demographics Type Range # of Participants19-29 1130-39 6Age 40-49 250-59 060-69 1Gender Female 12Male 80-2 7Facebook Membership 2-4 9(years) 4-6 36-8 10-100 6Facebook Friends 100-500 9500-1000 5Table 4.1: Demographics of interview participantsFinally, we read the transcripts again and selectively coded any data related to thecore category.Theoretical codingDuring this stage of analysis, we applied to the data the developed theoreticalmodel. We integrated the model into related data in order to explain the corecategory. The outcome was a grounded model, or theory, about the lifecycle ofFacebook friend acceptance, which we discuss in the following section.4.6 ResultsWe now present the results of our exploratory study. First, we start by discussingthe overall model, and then continue with detailed descriptions of the model com-ponents and the relationships among them.24Privacy/ SecurityAwareness/ConcernInvestigationActionsFriendship FactorsInterface CapabilitiesDecision ExecutionMaintenance ActionsEnvironmental Factors132Decision Making ProcessFigure 4.4: Online lifecycle of Facebook friend acceptance (OLFFA) model.Shaded components on the top are the internal factors and componentswith hyphenated borders are the external factors. The middle box,which includes 3 components, represents the decision making process.The dashed arrows represent decision making flow. The solid arrowsrepresent the impact of components on each other.25The overall modelWe refer to the developed model as the Online Lifecycle of Facebook FriendAcceptance (OLFFA). It includes 7 components, as shown in Figure 4.4. Eachcomponent is derived through the coding steps that were described earlier and isrepresentative of a set of users’ behaviors.The factors that we found to have influence on the process of users’ decisionmaking can be categorized into four groups, to which we refer as components:Friendship Factors, Privacy and Security Awareness and Concerns, EnvironmentalFactors, and Interface Capabilities. Since the first two components (green shadedrectangles in Figure 4.4) are user-specific and subjective, we considered them asinternal (to the user). On the other hand, since a user does not have any directcontrol over the last two components (red rectangles with hyphenated borders), wecall them external factors. The components inside the large grey box in the middleof the figure represent the decision making process, and the numeric labels indi-cate the flow of actions associated with decisions. The rest of this section includesdiscussions about each of the components and their interactions with other com-ponents, particularly, how they impact other components or how they are impactedby other components.Friendship factorsThis is the component that was brought up and discussed by all of the participants.Friendship Factors impacts Privacy and Security Awareness and Concerns of usersin the sense that when users employ more restricted friendship factors, they becomemore sensitive about their profiles’ privacy and security.On the other hand, Friendship Factors could be impacted by Privacy and Secu-26rity Awareness and Concerns. This happens when the Friendship Factors that theusers employ change due to a an adjustment of their view on their profiles’ privacyand security:“Well, from the time my brother’s account on LinkedIn was hacked,I have always concern to have my info available on the internet. So Istarted to accept people that I feel comfortable to share my info withthem. Not like before that I was accepting almost everyone.” (P9)As the result, a user could become more conservative in making new friend-ships. A reverse change could happen as well.This component also impacts Investigation Actions and Maintenance Actions.For instance, if a user relies on the similarity of backgrounds for making friend-ships on Facebook, an investigative action could be to check out the requester’sprofile in order to see her background. Similarly, finding and removing passivefriends is another example of maintenance actions driven by friendship factors.Here is the list of Friendship Factors we have discovered:• Knowing the person in the real world (KRL): It was reported by partici-pants that they care about knowing people in real world or at least in onlinecommunities (e.g., forums), when they consider accepting friend requests onFacebook. For instance, P5 said:“If I do not know them, I do not accept them. I mean I should haveseen a person at least once to accept them as Facebook friend.”• Profile picture (PRP): The profile picture is one of the most important fac-tors for users. We encountered users who usually spend only a few seconds27to decide about friendship requests. Those users pay attention to only theprofile picture, as the fastest way to make their decision. As P4 puts it:“I can really know from pictures. If you do not have a picturethen I do not know you!”• Profile name (PRN): Similar to profile pictures, the profile name is used byusers especially for the case when they want to instantly decide about friend-ship requests. They prefer to receive requests from recognizable names, tofacilitate the process of decision making.• Common background (CBG): During the interviews, many participantsmentioned common backgrounds and interests as friendship factors. Userstend to accept friend request from people who have common backgroundwith them. These commonalities include city and country of birth or res-idence, schools and universities attended, personal interests, and hobbies,etc. When we asked for the reason, the users pointed out that these common-alities work like a trigger that helps them remember the people they have onFacebook and to know them better. For example, P17 said:“Although it is fine for me to have new friends based on my inter-ests, I would prefer to be in the same city to make closer friend-ships.”• Being active on Facebook (BAF): According to our data, the fact that thefriend requester is an active Facebook user is sometimes the most importantfactor, even more than knowing the requester. P5 expressed this by saying:28“If they send me a request, okay, I know you. I am going to acceptyour request but it has been five months and you are not postinganything. You never come to Facebook. You never post anything.Okay, I am sorry. I have to delete you because you are not addinganything.”• Gender (GEN): The gender was another factor for participants. P5 said:“I think gender is effective in terms of friend requests. You know,I am sorry to say it but put a picture of a pretty girl would gethundreds of friendship requests or even messages. I have a malefriend who was building a ‘stable’ of Facebook women. He hadabout 600 friends and they were all women. There is not a singlemale friend on the list!”• Number of mutual friends (NMF): The majority of participants confirmedthat the number of mutual friends is important, as it helps users to rememberwhether they know each other. Although it is known as a way of verificationby many users, it might fail them. P2 raised an interesting point about it:“I used number of mutual friends as a fast approach to acceptfriends but later it turned out it is not necessarily good enoughbecause I removed many friends who had large number of mu-tual connections with me. Maybe because I had a lot of friends,around 800, so I had many friends in common with people and itdid not work all the time.”• Closeness of mutual friends (CMF): Some participants highlighted that,29in addition to the number of mutual friends, it is also important to knowthe closeness of those friends. That is, even if there are a couple of mutualfriends between the receiver and the requester, it is not necessarily enoughfor users to make a decision. As P5 expressed it:“You either have to be someone I know or you have to be mutualfriends with someone I really know. Anyone else I do not takerequests anymore because I ran into some pretty weird people.”• User’s activity pattern (UAP): Another friendship factor was user’s activitypattern, including what kind of information is shared (i.e., either relevant orirrelevant) and how often the content is shared. For instance, P1 said:“I do care about what they post. If they post, like, things that Iwould find disturbing for me, ding!! I would delete them.”Furthermore, our participants disliked being friends with those who justmonitor others’ posts, and possibly report to mutual contacts:“My aunt turned out was watching my page and then reported myactivities to my mom. And that did not go over well and I justblocked them. I would never befriend anybody who just monitorsothers.” (P6)Given this dislike for passive users, it was interesting to discover that someof our participants had changed their activity on Facebook over the years.They undergone a shift from active to passive users, who just read others’posts, without regularly adding any content. According to our participants,30an active user is the one who is willing to have a lot of Facebook friends andperforms a variety of activities, such as sharing photos, notes, and videos, aswell as posting their status, etc.• Closeness and quality of friendship in real life (CFR): We found in theinterview data that it is important for users to make sure how good of afriend they might become with the requester and if they might get along. Forinstance, P6 reported:“If I know them then, it takes a little bit longer because then Ihave to decide because my half-brothers and their daughters haverequested to be my friends. And yes, I know them but, no I donot want them on my page. Because the girls I do not get alongwith when they come for Christmas dinner. We only see them atChristmas time and I do not get along with those girls. My half-brothers, the one I do not – I have only met this past summer forthe first time, so I do not know him and I am not interested!”Another participant, P5, expressed similar concerns:“I found this quite upsetting but there is a woman on my site whoI worked with. We were quite close at work but I did not like anumber of things that she did, and you know I did not accept herrequest.”• Application-based friendship (APF): There was another factor raised byour participants where users tend to make friendships with others for thesake of receiving bonuses from some applications such as games. As a result,31such users would send and accept more friendship requests.Privacy and security concerns and awareness:As described earlier, this component is influenced by and impacts Friendship Fac-tors. Maintenance Actions also impacts this component. This might happen as amaintenance activity, for example, when a user monitors a friend’s profile and sheends up facing surprisingly irrelevant content posted by this friend. This observa-tion would cause them to be aware of fake or hijacked accounts posing as closefriends:“I remember that I found that there were two accounts for a friend ofmine and I thought he had created another one. When I asked, it turnedout that the first one was a fake account and he had already deactivatedhis previous account. So, somebody had created an account similar tohis first account. I did not know that. I even checked my name to seeif there is any fake account for me as well as other friends.” (P17).Another source of influence on this component is Environmental Factors ingeneral and media in particular. Some participants noted that their awareness ofprivacy and security on Facebook were affected by media reports. For example, P7shared:“Previously, I would just add like a lot of random people and acceptrequests. Later, I became more conservative, as I heard from mediaabout leakage of users’ information.”P1 also believed that there were security incidents reported by media that influ-enced her behavior:32“Because there are a lot of issues with Facebook, like pictures, as therewas the recent one about the girl who committed suicide and how herphoto was used for some porn website so things like that. So for thepictures that I post on Facebook, they are never of my face.”P3 had similar concern describing his experience:“I used to post a lot of photos on Facebook but then there are issueswith security. The more you post, the more you cannot take backbecause I read in a blog that even if you post a photo on Facebook andget rid of it from your account, just delete an album, you are still goingto be on Facebook. So because of that I stopped posting photos on myaccount.”We also found an interesting point about the effect of security and privacyincidents in other online services, which results in change of behavior on Facebook.P10 said:“I had profiles on LinkedIn and Evernote but then I removed it be-cause of some security leak in passwords. I got sensitive in terms ofdisclosing information on my accounts.”Interface capabilities:Our participants reported a set of issues related to capabilities of the interface—e.g., lack of required information, device-specific design, and frequent changes ofprivacy settings—that would impact Investigation Actions and Maintenance Ac-tions.33Some of the participants could not easily find desired information in order tomake decisions about friendship requests. As a result, they preferred sometimes tothink about requests, rather than looking for additional information on Facebookabout the requesters. This raises the issue of information visibility in the interface.For instance, P3 provided the following suggestions:“Definitely need to have what/where they are from, what they have, ifit is in academic backgrounds, then what they studied and where. Andif it is just maybe a few interests that they have, [it] could never hurt, Ithink. Just because you look at a person and you think they are inter-ested in photography I do not think it could actually hurt anyone. Sojust something along those lines that can give you more information.”Regarding the issues related to device-specific design, P8 shared her experienceas follows:“In terms of an interface, maybe a bigger button, I think just becausesometimes all those buttons look very similar and you tend to clickone. If you are using your phone and looking at someone who youare not a friend of, but you want to (this has happened to me before),you want to message that person instead before you add as a friendand then by mistake because the buttons are right next to each other Iwould press add a friend, send a friend request, or add a friend insteadof message. So when that goes out that is it. They receive it and thenyou cannot really retract that.”P13 mentioned another issue in this regard:34“It really depends if I use my phone or my desktop when I accept orreject a request. Using the desktop, I spend way more time while thisis not the case with my iPhone. So you would be lucky to have me ondesktop when receiving your request. On iPhone, I would make mydecision very quickly. If I do not remember, I would just reject.”This issue shows the gap between usability of device-specific designs of interfacesfor accepting/rejecting requests.The last issue about the interface was frequent changes made to the interface,the privacy settings in particular. Participants found it difficult to catch up withthese changes.Investigation actions:Before making their mind in regards to friendship requests, some of our partici-pants took one or more of the following actions:• Sending personal message: Specified by many participants, sending per-sonal message is a common technique for obtaining additional informationabout the requesting user, especially when he is not known to the receiver.As P7 explains:“I would personally ask them on private messaging and say that Ido not know you or asking some questions like ‘have I met you?’”• Checking out photos: It was also common among the participants to go tothe profile and, if possible, check out photos of the requester. They reported35to be helpful to recognize the requester, to either make decide about therequest or start communicating with the requester via messaging.• Looking for commonalities: Another action taken by our participants wasto explore for commonalities in terms of background, friends, interests, etc.,as P5 illustrated:“Do we have common interests? Do you know some friends ofmine? We have something in common maybe?”This action seemed to be done by those participants who had new friends,in order to help them know people better, as well as those who wanted tohave limited list of friends, in order to help them verify requesters, in casethe profile picture or name were not recognized.• Checking mutual friends profiles: Some of our participants reported that,although it was important to know if there were any mutual friends, it alsotook time to check out the mutual friends’ profiles for evaluating the close-ness of the relationship. Although it was important to some of our partici-pants, some other participants said that they would skip this step because itwas too time-consuming and required somewhat high cognitive load:“I really want to know more than just number of our mutualfriends and see if those are close friends but I check that whenit does not take me a long time. Like less than 5 minutes other-wise I won’t do that.” (P13).36Decision execution:We found three types of behavior for decision execution. (1) Some participantswould make their decisions immediately after they received requests. If they couldfind information they needed to make the decision, then they would easily makeit right away. There were other participants who would accept friend requestsright away, although for different purpose. They would do so in order to find outmore about the requester (after becoming friends) and then decide if they wantedto unfriend her or not.(2) Otherwise, they would reduce their set of decision criteria, in order to ex-pedite the process. In such cases, participants with less concerns about privacy andsecurity would most likely accept friend requests:“If I get a friend request that we share mutual friends but I do not knowthem, I am always hoping that I can check their profile. Sometimes itis restricted so you cannot. So I accept the friend request.” (P5)(3) On the other hand, some users would leave requests as they are, and post-pone further investigations.Maintenance actionsThe interview data revealed three types of Maintenance Actions that our partici-pants took after accepting friend requests.One of the common maintenance actions was to remove friends after a while,due to a number of different reasons. For examples, those friends that had beenadded in order to play face boo games, would be removed when there was no needto be friends with them. Another common reason was finding content shared by37to-be-removed users irrelevant. As a result of these actions, users may adjust theirPrivacy and Security Awareness and Concerns, which would eventually impacttheir Friendship Factors.One other type of maintenance actions was to define different levels of accessfor friends. This usually happened in two ways. One was to define separate groupsof friends and then specify visibility of the posts using these groups. The other waywas to deny specific users the ability to see a post or any desired content on-the-fly.This means that participants sometimes set the access level manually to avoid agroup of friends accessing the post. As an example, P7 said:“If it is for family pictures, I would just change the privacy setting torelatives. Then, I do not have to remember every one of those friends.Sometimes I do not even have to create a group for relatives though. Ican remember who are my relatives.”The third type of actions was for our participants to update the privacy settingsof their profiles. However, some of our participants, who were sensitive abouttheir privacy, complained about frequent changes that Facebook privacy settingsundergo:“It changes a lot, but from time to time I try to go back and look at it,but that could be like once a year or so.” (P3)On the other hand, we found that some participants were not even aware ofprivacy settings in the interface. When we asked about the possibility of access toinformation of their profiles, some of them did not even know if it were possible.P2 said:38“I guess so, because I have not seen that at all. But, now that you havetalked about that, to me that means there are thousands of people thatcan check who I am. Some groups are pretty big. I have not thoughtof it.”This issue with frequent changes in Facebook privacy settings illustrates the re-lationship between Maintenance Actions and Interface Capabilities, in which thelatter impacts the former.Environmental factorsAnalysis of interview data revealed that there are three environmental factors thatinfluence Investigation Actions and Privacy and Security Awareness and Concerns,as discussed before.First, the participants referred to the lack of time, as a factor that influencedtheir decisions about friend requests. For instance, P17 said“I have always problem with the lack of time during break times. Ihave to check updates, requests, messages, etc. in just 15 minutes. Ionce accepted a friend by mistake, as the requester had just same nameas a friend of mine and I had not checked his profile to get more infoabout him.”The second factor is the lack of concentration, while checking out Facebook:“On the way to university, I usually check out my profile on the bus. Ionce accepted a request when I was on the bus and that was a wrongdecision. I guess I was distracted by stops and also other passengersso that I forgot to send a message to the requester.” (P20)39The third environmental factor was the effect of media. As described earlier,the Privacy and Security Awareness and Concerns of our participants were im-pacted by media reports about security and privacy incidents.4.7 DiscussionIn order to answer the research questions, we decided to go one step back andenvision the problem as part of a bigger context. Therefore, we managed to comeup with a model which discusses users’ behavior when they want to accept/reject afriend request. This idea was supported with the fact that there is no previous studyfocused on this aspect of users behavior. Armed with such a model, we would beable to uncover behavior of users towards strangers since this scenario would bea specific case of the model. We define stranger as a person who is not familiarin real life or online communities. In this regard, we indirectly asked participantsabout their interaction with strangers so that we can reveal more details about thisscenario.Befriending strangers:As described in Section 4.3, before each participant was interviewed, the partici-pant received two friend requests, one from a Facebook profile of a real user, andthe other from an auxiliary profile made up for the purpose of the study. Five par-ticipants accepted at least one request from one of these accounts, and one of themaccepted requests from both accounts. When we reached in our interviews the de-briefing part, in which we informed the participants that these requests were fromour research team, their reactions varied.The participant who had accepted both requests said that it was okay with him40and he did not care about strangers among his Facebook friends, since he did nothave any idea that anybody could make any use of his profile data. The other fourparticipants who had accepted requests from either real or auxiliary accounts ofthe researchers had different attitudes. After hearing the scenario, they got nervousand one of them said:“I would not have accepted the request if I knew more. I saw the guyis from UBC and is a graduate student. I thought that it should nothurt.”Another participant, most of whose profile was accessible publicly, had similarlynervous reaction, especially when we explained the possibility of any user access-ing his profile information. He commented that in the future, he would pay moreattention regarding friend requests.In addition, we found evidence in interview data suggesting that some OSNusers don’t pay attention to possible threats, when it comes to making friendshipconnections:“I seem to be a million times more strict than most people. I knowsome friends who accept anybody that requests. Well, I mean a lot ofpeople do. They do take it too easy. How can you have 2,000 friends?”(P5)Another participant had a set of “friends” from accessory shops (she did not knowthem) while they had access to the profile information e.g., other friends in herprofile. Some participants seemed to have no criterion for making friendship. Theywould just add anybody, as P11 explained:41“I am always nice to requests on Facebook, as I cannot remember thatI have rejected a request.”Attitudes Towards Strangers: These observations made us more curiousabout users’ perception of Facebook users they do not know in real life. Our anal-ysis suggests that, when it comes to one’s attitude towards strangers on Facebook,our participants can be roughly divided into three groups.We found that one group of participants had a “take it easy” attitude towardsaccepting friend requests from strangers. As P1 justified:“I have spent some time with them on Facebook and they do not seemsomebody who would cause me pain!”As P1 mentioned, it is enough to have a feeling that a person is not going to makeany trouble for them. The other reason for accepting their requests is that havingless commonality might be even an advantage, as P16 illustrated:“I know some people in real life who have common things with melike our neighbor’s kids that we lived in the same neighborhood, wewent to the same school. But I do not want him to be on my Facebookprofile. I prefer to have more of these unknown guys instead of ourneighbor’s son, as some of them post cool stuff and I don’t need to beworried about my posts, because none of them would tell my dad whatI am doing!”On the other hand, for some other participants, only knowing a requester inreal life did not necessarily mean that this was a right person to be friends with onFacebook. P2 illustrated this point with the following example:42“I have like friends from primary school who ask me to be [Facebook]friends. But, in primary school you are friends with all your classroomso then it will be like your real friends. And that has not been done for15 years. So now I do not accept them anymore if I see that we are inreally different world and everything. It is my private life and I am anew person now.”P1 explains this attitude further:“If you have not kept in contact or you have not actually tried to stayin contact, I feel like there is no point. Long ago in the past, I do notgo back there.”Users who have this attitude are less vulnerable to the threat of accepting a stranger’srequest.The third group’s attitude was not as clear cut as for the first two groups. Asa result, participants from this group were influenced by the various factors speci-fied in our model. This group would be also vulnerable to the threat of acceptingstrangers’ requests, as participants from this group reported issues in recognizingpeople in real life or online communities.These groups are not necessarily mutually exclusive, i.e., the same user canexhibit in the majority of cases the behaviour of one group, and yet handle some ofthe requests following the pattern of another group.Accepting While Not Intending: Our analysis revealed that some of our par-ticipants would make inconsistent decisions. For instance, they would accept friendrequests although they didn’t have intention to be Facebook friends with the re-questers, as an example of P11 illustrates:43“Some requests are from people that I had a quick chat with them orsomehow I remember them but honestly I don’t want to be friends withthem. However, I will accept if they send me request.”These participants seem to find it socially awkward to reject friend requests. P18made it explicit.“I always have this problem with some of people I know but I don’thave a really good relationship with them that I cannot say no to theirrequest. I don’t know why but I think it’s better to accept rather thanreject them.” (P18)Usage Differences: We discovered differences in the way our participants usedFacebook, and these differences seem to correlate with they way they treated friendrequests. Although it has been previously shown that users tend to use OSNs (in-cluding Facebook) to make connections and share different kinds of data, we foundthree “flavours” of users:• Contributors: These are traditional users who both consume and contributenew content. They make friendships, share photos, share personal informa-tion, post updates, and interact with others by commenting and favoring theirshared content. From the point of view of this group, the aim of FOSNs isto make an environment in which people feel free to share information withothers and receive feedback. While they are willing to have more friends,they are also conscious about their profile privacy and friendship manage-ment, as P16 illustrated.“I really enjoy using Facebook when I share posts or comment on44a post and receive likes. But this is because I know my friendsand feel comfortable with them”• Observers: On the other end of the spectrum, there are users that avoidhaving social interaction and prefer to passively observe others. They havedifferent reasons for this behavior including lack of time, security concerns,difficulty to use the interface. As the result, they do not share any informationand they are willing to make connection with as many users as possible.“I like Facebook as it gives me the chance to read my friends’posts and watch their photos, read news and many other things.Of course I don’t share anything as I use my phone and it’s reallydifficult to type a lot. Moreover it takes a lot of time.” (P13)• Conscious Contributors: In addition to these two extremes of the spectrum,there are advanced contributors who are more sensitive about the audienceof their posts and other shared content. This third group of people reportsmore issues regarding friendship management, as P15 illustrates:“What I am looking for on Facebook is to interact with othersand share my info as well as see their posts. I am spending a lotof time to manage my profile and I have this difficulty to put myfriends in different groups as I want to have them but I don’t liketo share my personal photos or posts with all of them.”To summarize, our observation indicates that we can categorize users of FOSNsinto three groups, with Contributors and Conscious Contributors being more likelyto have issues in terms of privacy and security of their profiles. This sheds light45on the point that privacy and security would have different meanings for usersaccording to the type of their FOSN usage. Consequently, this may impact user’sattitude towards friend requests.Our Online Lifecycle of Accepting Friends model could be helpful for FOSNdesigners, when it comes to supporting users in deciding about friend requests. Themodel could aid in considering various factors that impact user decisions.46Chapter 5Quantitative StudyWhile the exploratory study allowed us to identify possible factors that have a rolein users’ decisions about friendship requests, we wanted to test these factors on arepresentative sample and measure the fraction of users who employ those factors.Also, we wanted to characterize FOSN users’ behavior and find factors that areemployed significantly more than other factors for making decision about friendrequests sent from other users, in particular strangers. Therefore, we decided toconduct an online survey that would allow us to collect quantitative data from arepresentative sample.5.1 Research questionAs previously mentioned, considering the results from the exploratory qualitativestudy, we chose to have another study to test the factors that were reported byparticipants. In particular, we were interested in characterizing users’ behaviorwhen it comes to decide about a friend request. Therefore, we aimed to answer thefollowing research question:47• Which of the factors identified in study #1 (exploratory study) influenceusers in deciding to accept a friend request?While the aforementioned question is in the position of a generic research ques-tion, we also focused on details in the sense that we considered four differentgroups that could potentially happen to FOSN users. These four groups includereceiving friend requests from strangers and whether they are accepted (group #1)or rejected (group #2), as well as the situation that a request is sent by a knownperson (in real life) and if it is accepted (group #3) or rejected (group #4). Havingthese groups, we aimed to investigate the key factors in each group and comparethem together.5.2 Survey executionFor each of the friendship factors identified from the interviews, the survey had atleast one statement (e.g., “If I recognize someone’s picture, I would accept his/herfriendship request on Facebook.”) and asked participants to indicate their agree-ment on Likert scale of 1-5. For those factors that had more than one statement,we used the mean score. For testing data quality, we have included contradictingstatements. For example, “I would accept a friendship request from a Facebookapplication.” and “I don’t tend to accept friendship requests sent by Facebook ap-plications.” All questions from the survey can be found in Appendix B.We recruited 425 M-Turk participants from USA and Canada. Each USA par-ticipant received $0.50 and Canadian $0.75. It took 16 minutes on average for ourparticipants to finish the survey. We removed 28 participants because of contradic-tions in their answers, which left us with responses from 397 participants.485.3 Sample representativenessWe compared our sample demographics to Facebook demographics in order todiscuss the representativeness of our sample. This comparison could give insighton limitation of our sample, which is potentially helpful for future work.5.3.1 Age comparisonAs Figure 5.1 shows, our sample is younger than Facebook users. We got moreyounger participants (18-24: 31% vs 23.2% and 24-34: 39% vs 24.4%) and fewerparticipants in higher age ranges (35-54: 25% vs 31.1% and 55 and above: 5% vs15.6 %). We did not have any preference to recruit participants from younger agerange and as mentioned earlier, we recruited participants from Amazon M-Turk.However, previous work shows that the turkers are relatively young with about80% in 18 to 40 years old age range (Average = 31, Minimum = 18, Maximum= 71, Median = 27) [53], which could be the reason for having a younger samplerather than Facebook demographics. It is also worth mentioning that we did nothave any participants in the age range of 13 to 18, as we chose to recruit participantswho were at least 19 years old (due to the official rules of British Columbia).5.3.2 Gender comparisonIn terms of gender and as Figure 5.2 shows our sample consists of more maleparticipants rather than female ones (58% vs 42%) while 53.3% of Facebook usersare females and 45.7% are males.49Percentage of participants (%)Figure 5.1: Comparison of our sample to Facebook population in terms ofage.Percentage of participants (%)Figure 5.2: Comparison of our sample to Facebook population in terms ofgender.50Figure 5.3: Distribution of the sample based on the levels of education.Figure 5.4: Distribution of the sample based on different employment sta-tuses.51Figure 5.5: Distribution of the sample based on length of membership onFacebook.Figure 5.6: Distribution of the sample based on frequency of Facebook usage.525.4 ResultsIn this section, we are presenting the results from the quantitative study. First, weprovide statistics related to participants demographics, then descriptive statisticsregarding employment of the friendship factors, and finally we discuss four groupsthat could happen to FOSN users when they receive a friend request. Analysesof these four groups would be insightful to reveal the factors that impact users’decision towards friend requests.Participants DemographicsDemographics of our participants show diversity of the sample. In terms of age,we had participants from 19 years old to 65 and more. Gender-wise our partici-pants were fairly evenly distributed. Participants also had diverse education levels(26% with high school or lower degree, 59% with undergraduate degree, 10% withgraduate degree, and 5% had other education levels). The employment status ofour participants varied, too: 56% employed, 22% students, 16% unemployed, 2%unemployed and 4% had other employment status.We also asked our participants general questions about their Facebook usageand experience. The majority (94%) were Facebook users for more than 2 years.In terms of usage frequency, 92% reported that they login into Facebook at leastonce a month, while 80% login several times a week. They were also asked to go totheir Facebook profile and enter the exact number of their friends. Our participantshad wide range of friendship circles, with minimum of 10 and maximum 3,000(mean 328, median 203). This shows that collected data came from users withdifferent befriending patterns. Majority (64%) of participants receive at least onefriend request in a month and only 7% receive friend requests less than once a year.53Number of Facebook friendsNumber of participantsFigure 5.7: Distribution of number of Facebook friends among participants.Friendship FactorsFigure 5.9 summarizes results of the survey on the friendship factors. The barsshow the percentage of all participants who reported employing each of the factors,i.e., they agreed with the corresponding statement(s).Starting from the most popular factors, requester’s profile picture (84%) andname (82%), participants accept friendship requests if they recognize the requesters.Seventy seven percent agreed with statement “I tend to accept friendship requestsfrom people I know in real life or online communities.”Another factor was “common background” (CBG). While 74% of participants54Figure 5.8: Distribution of frequency for receiving friend request.agreed that it is important to know requester’s background, the survey results showthat the participants were not specifically interested in a single type of backgroundinformation. And the importance varied among participants. For instance, only15% would accept friend requests from users who were born in the same city asthey were. Similarly, only 18% would accept friendship requests from users wholive in the same city as they do. On the other hand, 27% would be interested in hav-ing Facebook friends from the same school/university. The most popular type was“common interests/hobbies,” with 35% relying on this background information intheir decisions about friend requests. This particular result was corroborated in theinterviews, with participants reporting interest in new FOSN friendships with thosewho share interests or hobbies.Another factor that we tested was activeness of friends, with 72% reportinginterest in accepting friend requests from active users. In terms of gender (GEN),55Fraction of participants employing each of the friendship factors.App-based friendship = APFNumber of mutual friends = NMFUser s activity pattern = UAPGender = GENCloseness of friendship relationship = CFRCloseness of mutual friends = CMFBeing active = BAFCommon background = CBGSchools attended = SchoolCity of living = CityLProfile picture = PRPCity of birth = CityB Profile name = PRNKnowing in real life = KRLCommon hobbies = HOBFigure 5.9: Employment of friendship factors by participants.5639% of participants confirmed they consider it during decision making for friend-ship requests. The “number of mutual friends” (NMF), which is currently shownin the Facebook’s friendship request dialog, was only used by 31% of participantsfor making their decisions. On the other hand, the majority of participants (63%)do care about “closeness of mutual friends” (CMF) to them. Regarding the impactof “user activity pattern” (UAP), we found that 38% of participants were reluctantto accept a friend request if they saw irrelevant posts shared by the requester. Thiswas expected, as our interviews showed that although people like to have access tothe posts of requester, they usually do not have this level of access. The results alsoshow that “closeness and quality of friendship in real life” (CFR) was important for60% of participants. We also measured the number of participants who would ac-cept “requests from Facebook applications”. Results show that 22% of participantstook APF into consideration, as a factor in deciding about friend requests.Characterizing users’ behaviorRegarding our interest in focusing on users behavior towards friend requests andthe way they employ different friendship factors, we considered four groups inwhich FOSN users might be situated when they receive friend requests. Thesegroups originate from the point that we envision the process of making decisionabout friend requests to have friendship factors as the dependent variables (DVs)and decision of either accepting or rejecting friend requests as the independentvariable (IV). Examples of dependent variables could be profile picture (PRP),number of mutual friends (NMF), etc. Therefore, we consider four groups in whichusers accept/reject friend requests from known/strangers. In the following section,we explain each group.57Figure 5.10: Four groups discussed in the analyses.Group #1 (G1): In this group, users could accept a friend request from strangers.This group is highly of our interest as it is interesting for us to know that whatwould be friendship factors dominantly employed by users.Group #2 (G2): The second group is the situation that users would reject friendrequest from strangers. Although the group itself might not be far from expectation,the comparison of this to other groups, in particular group #1, would be interestingand reveal useful findings.Group #3 (G3): The third group is when participants would accept friendrequests from non-stranger (i.e., know people in real life). Similar to the group #2,although this group itself might not result in unexpected results and observations,we are more interested in finding significant differences when it is compared toother groups.58Group #4 (G4): The last group is the case that users reject friend requestsfrom non-strangers. This group could reveal important information regarding em-ployment of friendship factors. Moreover, it could be interesting to compare theresults from this group to group #2 in the sense that what differences are betweenintuitions behind rejecting strangers and non-strangers.For the purpose of statistical analysis and comparison between the groups, weused Mann-Whitney’S U test [49], which is a non-parametric version of t test. Thistest is useful when dependent variable is ordinal. The point of a Mann-WWhitney’sU test is that it treats the data as ordinal data. Therefore, we can order the data butthe difference between any of the two values is not consistent. What a Mann-Whitney U test does is to calculate the rank for each value instead of using thevalues as is. As the null hypotheses, no difference in the ranks between each twogroups was expected in terms of employment of friendship factors. The follow-ing part includes results and discussion from statistical analyses and comparisonsperformed on the four groups.Group #1 versus Group #2: As mentioned before, regarding our interest infocusing on users behavior towards strangers’ requests, we aimed to compare be-havior of participants who already reported to have strangers on their friend list(i.e., group #1) to those who did not (i.e., group #2). First, we found that 62%of users reported to have at least one stranger among their Facebook friends. Asthe result, we divided our data-set into two parts labeled as G1 (group #1) and G2(group #2). This was done by analyzing the answers to one of the survey questions,which explicitly asked participants if they have any strangers (at least one) amongtheir Facebook friends. We did not impose any definition for a stranger and par-ticipants answered based on their own definition. Then, we compared the results59P<0.05Fraction of participants employing each of the friendship factors for Group #1 and #2 (%)Application-based friendship =Number of mutual friends =User s activity pattern =Gender =Closeness of friendship relationship =Closeness of mutual friends =Being active =Common background =Schools attended =City of living =Profile picture =P<0.05 P<0.05City of birth = Profile name =Knowing in real life =Common hobbies =P<0.05Figure 5.11: Comparison of friendship factors employment between G1 andG2.60corresponding to each of the friendship factors to investigate similarities and dif-ferences. In the following, we describe the results of our comparison for each ofthe factors (Figure 5.11).We found that while only 68% of participants in G1 consider the knowledge ofthe requester in real life (KRL) in their decision process, this number jumps to 91%for G2, with the difference being statistically significant (Mann-Whitney’s test: p= 0.0003 < 0.05). We interpret this result as an indicator for the level of awarenessin these two groups.For profile name (PRN), although we did not see much difference between thegroups, participants in G1 reported more interest than those in G2 (80% vs 87%)for using profile name as a factor.For common background, we looked at four types of background information,including city of birth (CityB), city of Living (CityL), schools/universities attended(School), and common hobbies/interests (HOB). For the first three factors, wecould not find statistically significant difference between participants in G1 andG2. However, G2 participants were slightly more interested in them (CityB: 19%vs 12%, CityL: 21% vs 15%, School: 29% vs 25%). The difference was significantwhen it came to “common hobbies/interests” (HOB). While 40% of participantsfrom G1 employed this as a friendship factor, there were only 25% in G2 who didso (Mann-Whitney’s test: p = 0.03 < 0.05). This result could be leveraged as acue by socialbots to customize profile information in order to increase the chanceof getting their friend requests accepted. “Being active” (BAF) was also morepopular among G1 (76%) members rather than G2 members (64%), although thedifference was not statistically significant.Regarding the “number of mutual friends” (NMF), we saw significantly more61members in G1 (37%) than G2 (19%) employing it as a factor in their decisions(Mann-Whitney’s test: p = 0.01 < 0.05). Also, comparison of G1 and G2 in termsof “closeness of mutual friends” (CMF) indicated that more participants in G2(77%) cared about it than in G1 (57%) (Mann-Whitney’s test: p = 0.03 < 0.05).The results of comparison for NMF and CMF suggest that informing users aboutthe closeness of the requester with the mutual friends would be more effective thanonly showing the number of such friends (available in current interface).For user’s activity pattern, we found that participants from G2 were slightlymore interested in UAP than from G1. We suspect that the absence of statisticallysignificant results in regards to UAP is due to the difficulty of finding a pattern,as we had this feedback in exploratory study. Regarding closeness of friendshiprelationship, we did not find statistically significant difference between G1 and G2.This result is expected, as it more relates to groups in which friendship requestsare sent from known users, according to our interview data. Finally, we couldnot find statistically significant difference between participants in G1 (20%) andG2 (25%) regarding application-based friendship (APF), although we expected toobserve significantly more participants in G1 who rely on this factor. This mightbe because of the shortage in the number of participants who have received thistype of friendship requests.Group #1 versus Group #3: Similar to the previous section, we performedcomparison between group #1 and group #3 in order to understand possible dif-ferences between users’ behavior in these two groups. For G3, we consideredparticipants who tend to be friend on Facebook with people they know in real liferegardless of their behavior on Facebook. In the following, we describe the resultsof our comparison for each of the factors (Figure 5.12).62P<0.05Fraction of participants employing each of the friendship factors for Group #1 and #3 (%)Application-based friendship =Number of mutual friends =User s activity pattern =Gender =Closeness of friendship relationship =Closeness of mutual friends =Being active =Common background =Knowing in real life =City of living =City of birth =Profile picture =Profile Name =Schools attended =Common hobbies =Figure 5.12: Comparison of friendship factors employment between G1 andG3.63Among the friendship factors, we found significant difference between thesetwo groups only in UAP (i.e., user’s activity pattern)(Mann-Whitney’s test: p =0.008 < 0.05). As the results show, users in G1 are more interested in the activitypattern. We interpret this difference as an indicator that shows in G1 users aremore interested to have connection with people who have activities close to theirpreferences. There are also other observations that are insightful although theyare not statistically significant. For instance, NMF and HOB seem to be moreimportant in G1 rather than G3. Expectantly, KRL and CMF were employed morein G3 as the group itself is happened when a request is accepted from a knownperson.Group #1 versus Group #4: From comparison of group #1 and group #4 (Fig-ure 5.13), we found three factors to be employed significantly different in these twogroups. The analysis indicates that “School” factor (i.e., having attended commonschools/universities) is more employed in G1 (Mann-Whitney’s test: p = 0.005 <0.05). This actually makes sense as in G4 users are not easygoing even with peo-ple they know in real life. The next friendship factor is UAP that has significantdifference in terms of employment by users in G1 and G4 (Mann-Whitney’s test: p= 1.005e-07 < 0.05). According to the analysis, 67% of participants who reportedto have experience in G4 believed that user’s activity pattern is important. On theother hand, only 34% in G1 reported UAP as a factor taken into consideration.This result also makes sense as it means that users in G4 care much more aboutactivity pattern. The next factor that analysis shows significant difference is CFR(Mann-Whitney’s test: p = 0.002 < 0.05). While 78% in G4 believe that closenessof friendship in real life is an important factor to be considered, 55% has the sameopinion in G1. This could be interpreted by the fact that G1 is happened when a64P<0.05Fraction of participants employing each of the friendship factors for Group #1 and #4 (%)Application-based friendship =Number of mutual friends =User s activity pattern =Gender =Closeness of friendship relationship =Closeness of mutual friends =Being active =Common background =Knowing in real life =City of living =City of birth =Profile picture =Profile Name =Schools attended =P<0.05P<0.05Common hobbies =Figure 5.13: Comparison of friendship factors employment between G1 andG4.65stranger’s request is accepted. We also found that NMF is employed more in G1rather than G4 although there is no significant difference reported by the analysis.Group #2 versus Group #3: According to the analysis (Figure 5.14), thereare four friendship factors that are differently employed by users in G2 and G3.Starting from the first one, CMF is reported more to be used in G2 with 77%while 64% is reported in G3 (Mann-Whitney’s test: p = 0.04 < 0.05). KRL has asimilar story with 91% in G2 versus 76% in G3 (Mann-Whitney’s test: p = 0.01< 0.05). We also found that activity pattern (UAP) is more popular in G2 (41%versus 25%) rather than G3 (Mann-Whitney’s test: p = 0.0001 < 0.05). CFR is thelast factor with significant differences in G2 and G3. We found that CFR is alsomore employed in G2 in comparison to G3 (Mann-Whitney’s test: p = 0.05 <=0.05). Reviewing the aforementioned factors, we can observe that factors, whichhelp users recognize users’ identities are paid more attention in G2. We also foundthat HOB and NMF are more popular factors in G3 while the difference is notstatistically significant.Group #2 versus Group #4: Analysis of G2 and G4 results in finding fourfriendship factors that are differently employed in these two groups (Figure 5.15).CMF was reported by 77% who had experienced G2 and by 67% in G4 (Mann-Whitney’s test: p = 0.04 < 0.05). For KRL, the story was almost the same as 91%reported to consider it in G2 versus 76% in G4. Both show that making sure thatrequester is known has higher importance in G2 rather than G4. However, resultsabout CFR reveals another interesting point. By comparing the fractions related toCFR, it shows that CFR is paid more attention in G4 (78%) than G2 (64%)(Mann-Whitney’s test: p = 0.05 <= 0.05). This could be interpreted such that users in G4care more about the quality of friendship in real life rather than only investigating66P<0.05Fraction of participants employing each of the friendship factors for Group #2 and #3 (%)Application-based friendship =Number of mutual friends =User s activity pattern =Gender =Closeness of friendship relationship =Closeness of mutual friends =Being active =Common background =Knowing in real life =City of living =City of birth =Profile picture =Profile Name =Schools attended =P<0.05P<0.05P<0.05Common hobbies =Figure 5.14: Comparison of friendship factors employment between G2 andG3.67P<0.05Fraction of participants employing each of the friendship factors for Group #2 and #4 (%)Application-based friendship =Number of mutual friends =User s activity pattern =Gender =Closeness of friendship relationship =Closeness of mutual friends =Being active =Common background =Knowing in real life =City of living =City of birth =Profile picture =Profile Name =Schools attended =P<0.05P<0.05Common hobbies =Figure 5.15: Comparison of friendship factors employment between G2 andG4.68about knowing people in real life. In addition, we found UAP to be given morevalue in G4 rather than G2, which is aligned with our expectation considering thedefinition of groups.Group #3 versus Group #4: There are three friendship factors that are differ-ently employed by users in G3 and G4. Regarding our analysis, “School” factor ismore popular in G3 rather than G4 while UAP and CFR are more employed in G3in comparison to G4. Interpreting the results is similar to the previous section asusers in G4 are more interested to have connection with people who are close tothem and also have acceptable behaviors in their point of view. On the other hand,being friend with people who are know in real life is appreciated and as the resultpopularity of “School” factor in G3 could be an example of that.We discuss these results as cues to propose our suggestions for interface designimprovements in the next section.5.5 DiscussionConsidering the first goal defined for the survey, we analyzed the data related toeach of the factors to investigate how much they are used. As the result, exceptfor UAP and APF, all other friendship factors were employed by at least more than50% of participants, which shows the validity of friendship factors inferred fromthe exploratory study. In addition, we asked survey participants to share with usother friendship factors if they have any. Analysis of answers to this question didnot add to the factors themselves. The participants who answered this question,mostly suggested features that could be added to the friend request decision di-alogues. As mentioned earlier, since having access to user’s wall is usually notpossible, people may not consider UAP as a friendship factor that could be eas-69P<0.05Fraction of participants employing each of the friendship factors for Group #3 and #4 (%)Application-based friendship =Number of mutual friends =User s activity pattern =Gender =Closeness of friendship relationship =Closeness of mutual friends =Being active =Common background =Knowing in real life =City of living =City of birth =Profile picture =Profile Name =Schools attended =P<0.05Common hobbies =Figure 5.16: Comparison of friendship factors employment between G3 andG4.70ily investigated although they are interested in it. However, according to the ex-ploratory study, participants prefer to have information about the activity patternsof requesters. For APF, a low percentage was expected from the interview study,in which only few participants reported receiving friendship requests from appli-cations.For the second goal, the idea of focusing on four different groups and per-forming statistical analysis and comparison helped us to uncover the impact ofthe friendship factors. As the results show, we found a set of friendship factorsthat could play a notable role and affect user’s decision towards friend requests.As previously mentioned, our findings show that in G2 and G4 that are groups inwhich users reject friend requests from strangers and known people, four factorsincluding CMF, KRL, UAP, and CFR are paid more attention and users care moreabout them while NMF, HOB and School are significantly less employed. By re-viewing these factors, it is implied that privacy and quality of relationship are takeninto consideration. For instance, closeness of mutual friend is more important thannumber of mutual friend. Also knowing in real life or closeness of friendship inreal life as well as activity pattern could be interpreted as an indicator for quality ofa relationship. By focusing more on these groups, we can also find out some dif-ferences between G2 and G4. In this regard, analysis show that CMF and KRL aremore employed in G2 as they are more about investigating the identity of requester.On the other hand we found that CFR is more used in G4 as it helps to make sureFacebook friends are of those that are close friends in real life as well. This re-sult could be leveraged for improving the interface design so that user makes moreinformed decision.71Chapter 6DiscussionIn this thesis, we aimed for a better understanding of user’s befriending behavior inFOSNs, and also what makes them to accept or decline friendship requests. This isan important problem as accepting friend requests from strangers is still known tobe a vulnerable behavior. Our work contributes to provide socio-technical solutionsto help FOSN users be aware of their decisions towards friendship requests sentfrom strangers. We performed two studies in order to tackle our research problem.First, we conducted interviews with 20 active Facebook users to better un-derstand user’ behavior for making decision about friend requests. According toour analysis, we showed there are three factors that impact users decisions includ-ing internal factors (Friendship Factors, Privacy/Security Awareness and Concern),external factors (Environmental Factors, Interface Capabilities) as well as a 3-stepprocess of decision making (investigation, decision execution, maintenance). Webelieve that this model is helpful to improve part of interface related to scenarioof receiving friendship requests. Having this model, it is possible to understandthe process of friend request acceptance and improve the interface design. This72model also helps us to answer our research questions (factors and actions taken byFacebook users) as it introduces the friendship factors employed by FOSN users.Moreover, it gives insight to find out what are the pre-actions and post-actionsthat are taken by users before and after the decision making process about friendrequests.Second, we aimed to find possible factors that have a role in users’ decisionsabout friendship requests in order to address our research question, which wasto characterize users’ behavior when it comes to decide about a friend request.Therefore, we tested the friendship factors revealed by the exploratory study andmeasured the fraction of users who employed those factors. Then, we consideredfour different groups in which users are situated when they receive friend requests.We characterized users behavior in these groups by performing statistical analysisand comparing employment of friendship factors in each group. Characterizing theusers behavior helped us to find key factors that influence users to take a decisionabout a friend request. It also allowed us to collect quantitative data from a rep-resentative sample (397 participants using Amazon Mechanical Turk). Accordingto our results, accepting stranger’s request is still a threat as having a stranger infriend list was reported by 62% of our participants. We also found interesting re-sults from the analysis of the groups. We introduced 4 friendship factors (Knowingin the real world, common hobbies/interests, number of mutual friends, closenessof mutual friends) that can significantly impact users’ decision towards stranger’srequest. We also found friendship factors that are employed significantly differentin the groups based on our pairwise comparison of the groups. Also, our resultsshow that majority of the factors that are employed significantly by users are notcurrently available in the interface design of FOSNs (e.g., User’s activity pattern,73Closeness of friendship relationship, Closeness of mutual friends). Therefore, theinterface design could be improved by providing features that include these friend-ship factors.As the third part, we proposed 4 guidelines to improve the interface such thatusers can make more informed decisions when they receive friendship requestsfrom strangers.Interface Design Recommendations: As discussed before, the results fromthe analysis of our survey data revealed interesting points about friendship fac-tors that could be used for improving the Facebook interface. Therefore, we offerthe following suggestions for designing user interfaces for accepting friendshiprequests:• The interface should convey the importance of making accurate decisionsabout friendship requests and encourage users to make informed decisions.For instance, users could be notified by a pop-up window (similar to currentdesign) asking users to go to another page in order to make an informeddecision, using useful information or a check list. Having such a featurein the interface is supported by the OLFFA model since it helps users toappreciate the importance of these decisions.• The interface could contain a message box so that requesters can briefly spec-ify how they know the user. Another suggestion is to give access to photosselected by each user to better recognize the requester. We had reports fromparticipants of both studies complaining about unclear small photos. Thiskind of improvement would facilitate the investigation/maintenance actions(in the decision making process of OLFFA model) for users.74• It could be helpful if user had access to statistics (number of likes, numberof comments, number of personal messages, number of common photos)about interaction with his/her friends. In this case, it is easier to investigatecloseness of mutual friends, which was shown to be more useful than onlythe number of mutual friends. In other words, this feature would facilitatethe Investigation Actions in the OLFFA model for finding out closeness ofmutual friends.• The interface could encourage the user to specify the access level for newfriends at the time the user accepts a friend request. We suggest this becauseour analysis showed that 31% of participants in S1 did not define any accesslevel for their friends while 9% in S2 reported similar behavior. Therefore,this could be helpful (at least for users who accept stranger’s requests) asa facilitator for performing maintenance actions and help users to be morecautious about the level of access they grant to their Facebook friends. Alter-natively, it could be helpful to suggest different friend types such as acquain-tances, co-workers, close friends to users when they accept a new friendrequest.It is worth mentioning that although we believe these recommendations could behelpful for the Facebook interface improvement, they are currently hypotheses tobe tested.As previously mentioned, offering these recommendations is motivated by ourresearch question and the goal to prevent or at least limit the security implicationsof large-scale infiltration attacks shown by previous work. These recommendationscould be helpful to improve the part of the interface related to receiving friend75requests.6.1 Future workThere are several directions for future work. One direction could be to focus oneach component of the model (external and internal factors) and investigate po-tential impact of them on deciding about stranger’s request. Another directionis to perform structural model testing on the proposed model Structural EquationModeling (SEM) to test the significance of impact of model components on eachother. Although the model is already verified by the Grounded Theory, it could beinteresting to figure out how significant is the relations between each of the compo-nents. Finally, another direction is to conduct a user study and investigate impactof modifying the interface using the proposed recommendations. It could be doneby running user studies on controlled groups . For instance, one group can be askedto use the real Facebook interface and the other group can be asked to use a sys-tem, which has already the prototype installed on it. Then, these groups could becompared in terms of performance criteria, which is the rate of request acceptancefrom strangers. It could be also interesting to perform evaluation by comparing theresults to other solutions such as fake account prediction algorithms.6.2 LimitationsOur work has several limitations. In the exploratory part, it would be better to havemore diversity in terms of age so that the model could be representative of a widerrange of Facebook users. On the other hand, although we reach saturation in datacollection, we had five participants who accepted friendship requests from the vol-unteer. Having more participants from this group could result in more interesting76observations and a more accurate model as it could include cultural aspect of users’behavior. Another limitation that might exist in our study is the Hawthorne effect[4]. However, as we used a qualitative technique in this study (semi-structuredinterviews), we tried to elicit the participants’ thinking during the study. Also, ac-cording to previous work [47], the Hawthorne effect, as it is usually understood,is nothing more than a popular myth that should not be assumed as a reliable basisto question the validity of any experimental study. Also, the Hawthorne effect isvery controversial and there are different interpretations for this effect and it is notnecessary valid and safe to criticize based on this effect [47].In the survey, we asked participants to report their activities, which might in-clude some inaccuracies similar to all self-report studies. As an alternative, it couldbe done by providing them with different scenarios and then asking them questions.We refrained from doing this due to the time limits of our survey. Also, our sampleis not representative of all Facebook users, as we recruited participants only fromUSA and Canada. Having participants from other countries could reveal more in-teresting points about users befriending behavior. Moreover, we did not focus ongender differences in our study design, which has been shown to be important byprevious work [1] as men and women have different social media activities and dif-ferent reasons for making their connections in social media including online socialnetworks (Facebook in particular). Also, while we briefly discussed the obser-vations in which participants reported the behavior transfer from their experiencein other social networks such as LinkedIn to Facebook (e.g., in terms of securityawareness) in the exploratory results, we could deeply focus on behavior of par-ticipants on other social networks such as Flickr, LinkedIn, tumblr, Twitter, andGoogle+. We could also consider personal social behavior of participants in the77real word by capturing information about personal rating of shyness, introversionand extroversion, and number of friends they have in the outside world.78Chapter 7ConclusionOur work contributes to provide socio-technical solutions to help them be aware oftheir decisions towards friendship requests sent from strangers. First, we aimed tobetter understand their behaviors by conducting an exploratory study. Accordingto analysis of interviews, we developed a model, which we call Online Lifecycleof Facebook Friend Acceptance (OLFFA). In this model, we showed that there arethree factors that impact users decisions including internal factors (Friendship Fac-tors, Privacy/Security Awareness and Concern), external factors (EnvironmentalFactors, Interface Capabilities) as well as a 3-step process of decision making (in-vestigation, decision execution, maintenance). We believe that this model is helpfulto improve part of interface related to scenario of receiving friendship requests asusers’ behavior could be interpreted by this model. We also showed that acceptingstranger’s request is still a threat as having at least one stranger in friend list wasreported by 62% of our participants. As the second step, we chose to have anotherstudy, which was a survey on Amazon Mechanical Turk. We wanted to test thefriendship factors revealed from the exploratory study. Moreover, we focused on79different situations that FOSN users may face when they receive friend requests.Therefore, we considered four groups including acceptance of strangers, rejectionof strangers, acceptance of known users, and rejection of know users. We statis-tically analyzed the differences between employment of friendship factors usingpairwise comparisons between the aforementioned groups. Results reveal differentfriendship factors that are employed significantly more in each of the groups. Theinteresting finding about the results is that majority of friendship factors that weredominantly employed (specially for identifying requesters) are not currently pro-vided by the Facebook interface. For instance, User’s activity pattern, Closeness offriendship relationship, Closeness of mutual friends are not currently available forFacebook users while these are desired by users who want to make sure about theidentity of requesters or make sure about the requester’s activity pattern. Regard-ing our interest in finding the reason(s) behind acceptance of strangers from FOSNusers (i.e. group #1), we focused on the results of from the comparison of group#1 and group #2. We also introduced 4 friendship factors (Knowing in the realworld, common hobbies/interests, number of mutual friends, closeness of mutualfriends) that can significantly impact users’ decision towards stranger’s request.Then we leveraged these results in order to come up with interface design recom-mendations that could be used to improve the interface design such that users canmake informed decisions towards requests sent from strangers.To summarize, this work has the following contributions:1. We developed a model for online lifecycle of Facebook friendship accep-tance, which explains the factors that influence users’ behavior in responseto friend requests.802. We characterized such factors and analyzed their impact on users’ decisionwith regards to friend requests. We also identified four factors that signifi-cantly impact users’ befriending decisions.3. Based on both qualitative and quantitative results, we suggest design guide-lines for FOSN interfaces that we expect can help users make informed de-cisions about friend requests.81Bibliography[1] →pages 8, 77[2] → pages 22[3] → pages 7[4] → pages 77[5] → pages 7[6] → pages 7[7] → pages 7[8] → pages 7[9] N. Ambady and R. Rosenthal. Thin slices of expressive behavior aspredictors of interpersonal consequences: A meta-analysis. Psychologicalbulletin, 111(2):256, 1992. → pages 8[10] R. Antone. Another isle man allegedly baits teen victim on myspace.Honolulu Star Bulletin, 9, 2006. → pages 10[11] T. Bahrampour and L. Aratani. Teens’ bold blogs alarm area schools.Washington Post, 17, 2006. → pages 10[12] J. Bates. Sniffing out socialbots: The combustive potential of socialmedia-based algorithms.,December 2011. → pages 282[13] L. Bilge, T. Strufe, D. Balzarotti, and E. Kirda. In WWW ’09: Proceedingsof the 18th International Conference on World Wide Web, pages 551–560,New York, NY, USA. ACM. ISBN 978-1-60558-487-4.doi: → pages 1[14] P. M. Blau. Exchange and power in social life. Transaction Publishers,1964. → pages 5, 6[15] J. Bollen, H. Mao, and X. Zeng. Twitter mood predicts the stock market.Journal of Computational Science, 2(1):1 – 8, 2011. → pages 2[16] Y. Boshmaf, I. Muslukhov, K. Beznosov, and M. Ripeanu. The socialbotnetwork: when bots socialize for fame and money. In Proceedings of the27th Annual Computer Security Applications Conference, ACSAC ’11,pages 93–102, New York, NY, USA, 2011. (Best Paper Award), ACM. →pages 1[17] Y. Boshmaf, I. Muslukhov, K. Beznosov, and M. Ripeanu. Design andanalysis of a social botnet. Computer Networks, pages 1–22, 2012.doi:10.1016/j.comnet.2012.06.006. URL →pages 1[18] Y. Boshmaf, I. Muslukhov, K. Beznosov, and M. Ripeanu. Key challenges indefending against malicious socialbots. In Proceedings of the 5th USENIXconference on Large-scale exploits and emergent threats, LEET’12,Berkeley, CA, USA, 2012. USENIX Association. → pages 3[19] M. Buhrmester, T. Kwang, and S. D. Gosling. Amazon’s mechanical turk anew source of inexpensive, yet high-quality, data? Perspectives onPsychological Science, 6(1):3–5, 2011. → pages 14, 15[20] K. Charmaz. Grounded theory. Strategies of qualitative inquiry, 2:249,2003. → pages 14[21] K. Charmaz. Constructing Grounded Theory. SAGE publications, 2006. →pages 3, 14, 16[22] E. Chung. Facebook easily infiltrated, mined for personal info., November 2011. → pages 183[23] J. Corbin and A. Strauss. Basics of Qualitative Research: Grounded TheoryProcedures and Techniques. Sage, Newbury Park, CA, 1990. → pages 14,16[24] L. F. Cranor. A framework for reasoning about the human in the loop. InUPSEC’08: Proceedings of the 1st Conference on Usability, Psychology,and Security, pages 1–15, Berkeley, CA, USA, 2008. USENIX Association.→ pages 11[25] S. Egelman, A. Oates, and S. Krishnamurthi. Oops, i did it again: mitigatingrepeated access control errors on facebook. In D. S. Tan, S. Amershi,B. Begole, W. A. Kellogg, and M. Tungare, editors, CHI, pages 2295–2304.ACM, 2011. ISBN 978-1-4503-0228-9. → pages 10[26] N. B. Ellison, C. Steinfield, and C. Lampe. The benefits of facebookâA˘IJfriends:âA˘I˙ social capital and college studentsâA˘Z´ use of online socialnetwork sites. Journal of Computer-Mediated Communication, 12(4):1143–1168, 2007. → pages 6, 9[27] B. Glaser and A. L. Strauss. The Discovery of Grounded Theory, Strategiesfor Qualitative Research. Aldine Publishing Company, Chicago, Illinois,1967. → pages 13, 16[28] B. G. Glaser. Theoretical sensitivity : advances in the methodology ofgrounded theory. Sociology Press, Mill Valley, CA, 1978. → pages 13, 21[29] B. G. Glaser and J. Holton. Remodeling grounded theory. In Forum:Qualitative social research, volume 5, 2004. → pages 13[30] B. G. Glaser and A. L. Strauss. Discovery of substantive theory: A basicstrategy underlying qualitative research. American Behavioral Scientist, 8(6):5–12, 1965. → pages 13[31] A. W. Gouldner. The norm of reciprocity: A preliminary statement.American sociological review, pages 161–178, 1960. → pages 5[32] R. Gross and A. Acquisti. Information revelation and privacy in onlinesocial networks. In Proceedings of the 2005 ACM workshop on Privacy inthe electronic society, pages 71–80. ACM, 2005. → pages 6[33] M. T. Hallinan. The process of friendship formation. Social Networks, 1(2):193–210, 1979. → pages 4, 584[34] D. H. Hargreaves. Interpersonal relations and education. Routledge and K.Paul, 1972. → pages 6[35] J. Hempel and P. Lehman. The myspace generation. Business Week, 3963:88–93, 2005. → pages 10[36] A. Hewitt and A. Forte. Crossing boundaries: Identity management andstudent/faculty relationships on the facebook. Poster presented at CSCW,Banff, Alberta, pages 1–2, 2006. → pages 11[37] G. C. Homans. Social behavior: Its elementary forms. 1961. → pages 5[38] M. Huber, S. Kowalski, M. Nohlberg, and S. Tjoa. Towards automatingsocial engineering using social networking sites. Computational Science andEngineering, IEEE International Conference on, 3:117–124, 2009.doi: → pages 3[39] T. N. Jagatic, N. A. Johnson, M. Jakobsson, and F. Menczer. Socialphishing. Commun. ACM, 50(10):94–100, 2007. → pages 2[40] M. Johnson, S. Egelman, and S. M. Bellovin. Facebook and privacy: it’scomplicated. In Proceedings of the Eighth Symposium on Usable Privacyand Security, page 9. ACM, 2012. → pages 10[41] A. N. Joinson. Looking at, looking up or keeping up with people?: motivesand use of facebook. In Proceedings of the SIGCHI conference on HumanFactors in Computing Systems, pages 1027–1036. ACM, 2008. → pages 6[42] J. E. Katz and R. E. Rice. Social consequences of Internet use: Access,involvement, and interaction. MIT press, 2002. → pages 7[43] A. Kittur, E. H. Chi, and B. Suh. Crowdsourcing user studies withmechanical turk. In Proceedings of the SIGCHI conference on humanfactors in computing systems, pages 453–456. ACM, 2008. → pages 14[44] A. Lenhart and M. Madden. Social networking websites and teens: Anoverview. Pew/Internet, 2007. → pages 9, 10[45] Y. Liu, K. P. Gummadi, B. Krishnamurthy, and A. Mislove. Analyzingfacebook privacy settings: user expectations vs. reality. In Proceedings ofthe 2011 ACM SIGCOMM conference on Internet measurement conference,IMC ’11, pages 61–70, New York, NY, USA, 2011. ACM. ISBN978-1-4503-1013-0. doi: → pages 1085[46] Z. Liu and B. J. Jansen. Factors influencing the response rate in socialquestion and answering behavior. In Proceedings of the 2013 Conference onComputer Supported Cooperative Work, CSCW ’13, pages 1263–1274, NewYork, NY, USA, 2013. ACM. ISBN 978-1-4503-1331-5.doi:10.1145/2441776.2441918. URL → pages 7[47] R. Macefield. Usability studies and the hawthorne effect, 2007. → pages 77[48] M. Madejski, M. Johnson, and S. Bellovin. A study of privacy settingserrors in an online social network. In Pervasive Computing andCommunications Workshops (PERCOM Workshops), 2012 IEEEInternational Conference on, pages 340–345, March 2012.doi:10.1109/PerComW.2012.6197507. → pages 10[49] H. B. Mann and D. R. Whitney. On a test of whether one of two randomvariables is stochastically larger than the other. The Annals of MathematicalStatistics, 18(1):50–60, 03 1947. doi:10.1214/aoms/1177730491. URL → pages 59[50] K. Y. McKenna, A. S. Green, and M. E. Gleason. Relationship formation onthe internet: What’s the big attraction? Journal of social issues, 58(1):9–31,2002. → pages 7, 8[51] F. Nagle and L. Singh. Can friends be trusted? exploring privacy in onlinesocial networks. In Proceedings of the 2009 International Conference onAdvances in Social Network Analysis and Mining, pages 312–315,Washington, DC, USA, 2009. IEEE Computer Society. ISBN978-0-7695-3689-7. doi:10.1109/ASONAM.2009.61. URL → pages 1[52] L. Palen and P. Dourish. Unpacking privacy for a networked world. InProceedings of the SIGCHI conference on Human factors in computingsystems, pages 129–136. ACM, 2003. → pages 11[53] G. Paolacci, J. Chandler, and P. G. Ipeirotis. Running experiments onamazon mechanical turk. Judgment and Decision making, 5(5):411–419,2010. → pages 14, 49[54] M. R. Parks and K. Floyd. Making friends in cyberspace. Journal ofComputer-Mediated Communication, 1(4):0–0, 1996. → pages 786[55] J. Peluchette and K. Karl. Social networking profiles: An examination ofstudent attitudes regarding use and appropriateness of content.CyberPsychology & Behavior, 11(1):95–97, 2008. → pages 6, 9[56] J. Ratkiewicz, M. Conover, M. Meiss, B. Gonçalves, S. Patil, A. Flammini,and F. Menczer. Truthy: mapping the spread of astroturf in microblogstreams. In Proceedings of the 20th international conference companion onWorld wide web, WWW ’11, pages 249–252, New York, NY, USA, 2011.ACM. → pages 2[57] H. Rheingold. A slice of my life in my virtual community. High noon on theelectronic frontier: Conceptual issues in cyberspace, pages 413–36, 1996.→ pages 7[58] B. Schneier. Secrets and lies. Wiley, Indianapolis, Ind., 2000. ISBN0471253111. → pages 11[59] R. Schreiber and P. Stern. Using Grounded Theory In Nursing. SpringerSeries. Springer Publishing Company, 2001. ISBN 9780826114068. URL → pages 13, 16[60] M. Shepperd. Empirically-based software engineering. Upgrade-TheInternational Journal for the Informatics Professional, 4(4):37–41, 2003. →pages 12[61] C. Sibona and S. Walczak. Unfriending on facebook: Friend request andonline/offline behavior analysis. In System Sciences (HICSS), 2011 44thHawaii International Conference on, pages 1–10. IEEE, 2011. → pages 9[62] M. A. Stefanone, D. Lackaff, and D. Rosen. We’re all stars now: realitytelevision, web 2.0, and mediated identities. In Proceedings of thenineteenth ACM conference on Hypertext and hypermedia, pages 107–112.ACM, 2008. → pages 10[63] T. Stein, E. Chen, and K. Mangla. Facebook immune system. InProceedings of the 4th Workshop on Social Network Systems, SNS ’11,pages 8:1–8:8, New York, NY, USA, 2011. ACM. ISBN978-1-4503-0728-4. doi: → pages 3[64] F. Stutzman. An evaluation of identity-sharing behavior in social networkcommunities. Journal of the International Digital Media and ArtsAssociation, 3(1):10–18, 2006. → pages 6, 887[65] J. W. Thibaut and H. H. Kelley. The social psychology of groups. 1959. →pages 5[66] S. T. Tong, B. Van Der Heide, L. Langwell, and J. B. Walther. Too much of agood thing? the relationship between number of friends and interpersonalimpressions on facebook. Journal of Computer-Mediated Communication,13(3):531–549, 2008. → pages 8[67] J. B. Walther, B. Van Der Heide, S.-Y. Kim, D. Westerman, and S. T. Tong.The role of friends’ appearance and behavior on evaluations of individualson facebook: Are we known by the company we keep? HumanCommunication Research, 34(1):28–49, 2008. → pages 8[68] Y. C. Yuan and G. Gay. Homophily of network ties and bonding andbridging social capital in computer-mediated distributed teams. Journal ofComputer-Mediated Communication, 11(4):1062–1084, 2006. → pages 10[69] S. Zhao, S. Grasmuck, and J. Martin. Identity construction on facebook:Digital empowerment in anchored relationships. Computers in humanbehavior, 24(5):1816–1836, 2008. → pages 8, 988Appendix AInterview MaterialsAll the documents related to the exploratory study are attached in this section.89 1 .  I NT ERVIEW  Note: At the beginning of the interview, w e will not inform the interviewees the potential threats of accepting a stranger s friendship request in Facebook.   O b jectives: ŽůůĞĐƚŝŶƚĞƌǀŝĞǁĞĞƐ ƌĞƐƉŽŶƐĞƐ to investigate users  behaviors towards friendship requ ests sent from users and strangers in particular.  Sample:  Active users on Facebook ( login to account at least once a week) .  Data collection should be done till we reach the theoretical saturation.   Agenda:   1 .  'ŝǀĞĂŶŽǀĞƌǀŝĞǁŽĨƚŚĞƉƌŽũĞĐƚ͗͞dŚĞƉƵƌƉŽƐĞŽĨƚŚĞƐƚƵĚLJŝƐƚŽŝŶǀĞƐƚŝŐĂƚĞthe factors users employ when making a decision to befriend other users͘͟ 2 .  Introduce second interviewer and specify his role.   Part1:   A.  EXPLORATORY STUDY DOCUMENTS ͶINTERVIEW  A.1 General Questions :  a.  What  is your age?  b.  What is your gender?  c.  What  is your highest level of education?  d.  What is your major or occupation?   e.  How long have you own a Facebook account?  f.  How often do you use Facebook?  g.  What is your first language?    A.2 The befriending behavior of users with strangers:    a.  How many friends do you have on Facebook?  b.  How often do you receive friend requests?  c.  Have you ever accepted a  friendship request from a stranger you do not know in real-life or have not met before online or offline?  d.  What kind of factors do you rely on when you decide to accept a friendship request from a stranger?  ( For any factor users ask, we need to dig into mo re details by asking questions)  (Gender, Friends, Mutual Friends, Profile, Picture, Wall Ͷshow the activity in Facebook)  i)  ( T he interviewee mention ed  gender . )  Will you accept a friendship request from a homosexual stranger or a heterosexual one?  ii)  ( T he interviewee mentioned friends. )  How many friends does the stranger have that you will accept his/her friendship request?  iii)  (The interviewee mentioned mutual friends. )  How many mutual friends does the stranger have that you will accept his/her friendship request?  iv)  (The interviewee mentioned profile .)  Same/different hometown  Same/different schools  90 Same/different age  v)  (The interviewee mentioned wall .)  Active/quiet person  ͘ϯhƐĞƌƐ ĂƚƚŝƚƵĚĞƐƚŽǁĂƌĚƐƚŚĞŝƌprivacy security:  a.  Have you ever set your privacy setting?  i)  (If yes)  How did you modify your privacy setting?  b.  Have you assigned different privacy setting to your friends?  i)  (If yes)  How did you modify your privacy setting fo r different friends?  c.  Have you had reported any security incident before in your online activities on Facebook, email, etc. ?  d.  Have you realized that if you accept a friendship request from a stranger, he/she will have the access to your personal information?  i)  (If yes)  What kind of information do you think will be exposed to the strangers?  e.  Do you mind your private data being exposed to the strangers?  i)  (If yes)  What kind of information do you mind being accessed to the strangers?  ͘ϰhƐĞƌƐ ĂƉƉĞĂůŽĨƐƚƌĂŶŐĞƌƐ͗ a.  Ho w do you describe your connection with the stranger that you have accepted his/her friendship request?  b.  Are you emotionally attached with the strangers?   3 .  At the very end, do mention that the request will be removed.   Debriefing happens here!   Part 2:   B.1: What would be your suggestion if you want to design the window for friendship requests?  B.2: Will you change your behavior towards friendship requests? (If participant had accepted the request)  B.3: Do you have anything else related to this study that  you want to share with us?                91Recruitment,	  Version	  1.2,	  06/09/2013	  1	  of	  1	  Understanding Users’ Befriending Behavior In Online Social Networks  The Electrical and Computer Engineer department at the University of British Columbia invites participants for a study on “users’ befriending behavior in online social networks”. Each participants will receive 25$ for their participation. We require volunteers to participate in one-hour interview session. We will ask questions related to their occasional befriending behavior, and the factors they use when deciding to accept a friendship request. Participants requires the following criteria: 1. Being active Facebook user, which means checking the profile for at least once a week. 2. Having the ability to speak English 3. At least 19 years old  If you would like to participate in this study, please login to Facebook and send a personal message that includes your email address to this user profile:  	  	  	  The	  University	  of	  British	  Columbia	   	  92	   Consent	  Form,	  Version	  1.2	   06/09/2013	  1	  of	  2	  Consent Form for a User Study  Principal Investigator: The principal investigator of this research is Dr. Konstantin Beznosov from the Department of Electrical and Computer Engineering (ECE) at the University of British Columbia (UBC). You can contact him at or (604) 822-9181 Co-Investigators: Yazan Boshmaf, Ph.D. Candidate Hootan Rashtian, M.A.Sc Student  All co-investigators are from the ECE department at UBC. You can contact them at (604) 827-3410 Purpose: The purpose of the study is to investigate the behavior of Facebook users. We aim to investigate the factors users employ when making a decision to befriend other users. We also aim to understand how users manage the sharing of their data with their new friends.  Study Procedures:  You will be interviewed about your befriending behavior in Facebook. We will ask you questions about the factors and cues you employ to decide whether to accept a friendship request, and the techniques you use to mange your new friendship. The whole session will be approximately one hour long and is audio-recorded. You will be observed during the session by one co-investigator.  Confidentiality: The identities of all participants will remain anonymous and will be kept confidential. Identifiable data and audiotapes will be stored securely in a locked cabinet or in a password protected computer account.  Contact for information about the study: If you have any questions or require further information about the project you may contact Prof. Konstantin Beznosov at (604) 822-9181, Hootan Rashtian at (778) 834-4327. Contact for concerns about the rights of research subjects: If you have any concerns about your treatment or right as a research subject, you may contact the Research Subject Information Line in the UBC Office of Research Services at (604) 822-8598 or if long distance e-mail     T H E  U N I V E R S I T Y  O F  B R I T I S H  C O L U M B I A        93	   Consent	  Form,	  Version	  1.2	   06/09/2013	  2	  of	  2	  Consent: Your participation in this study is entirely voluntary and you may refuse to participate or withdraw from the study at any time. Note that you should be at least 19 years old in order to participation. There is also a $25 award of honorarium as appreciation of your participation. Your signature below indicates that you have received a copy of this consent form for your own records and indicates that you consent to participate in this study.  Participant Signature  Date   Printed Name of the Participant Signing Above   Researcher Signature  Date   Printed Name of the Researcher Signing Above 94Confidentiality Agreement, Version 1.1, 03/09/2013 	  1	  of	  1	    Mediator Confidentiality Agreement  I, __________________________________, student of _______________ with ID _____ enrolled in program _____ at University of British Columbia, agree that my employment by Dr. Konstantin Beznosov shall be strictly on the following terms and conditions:   1. I acknowledge that I have completed the TCPS tutorial. 2. I acknowledge that I have been advised that all information and documents that I may have knowledge of or access to through my employment about the Facebook befriending study are strictly confidential. 3. I undertake and agree at all times to treat as confidential all information acquired through my employment with Facebook befriending study including email addresses, Facebook IDs, etc., and not to disclose for any purpose. I acknowledge that such information is not to be altered, copied, interfered with or destroyed, except upon authorization from principal investigator of the study. I will not discuss such information with any party, nor will I participate in or permit the release, publication or disclosure of such information, nor will I copy, distribute, or disseminate such information, except as authorized in the course of my employment or by law. I understand that this agreement and undertaking includes:  a. Never discussing the personality of a participants, his or her file or any details about them;  b. Avoiding the use of names of participants in conversations with other clients, friends or relatives;  c. Ensuring that disclosures of information are made only to persons entitled to that information;  SIGNED at __________________________, British Columbia, this __________ day of ______________________________, 2013.  __________________________________     _____________________________ Signature          Witness    T H E  U N I V E R S I T Y  O F  B R I T I S H  C O L U M B I A         95Confidentiality Agreement, Version 1.1, 03/09/2013 	  1	  of	  1	    Volunteer Confidentiality Agreement  I, __________________________________, student of _______________ with ID _____ enrolled in program _____ at University of British Columbia, agree that my employment by Dr. Konstantin Beznosov shall be strictly on the following terms and conditions:   1. I acknowledge that I have completed the TCPS tutorial. 2. I acknowledge that I have been advised that all information and documents that I may have knowledge of or access to through my employment about the Facebook befriending study are strictly confidential. 3. I undertake and agree at all times to treat as confidential all information acquired through my employment with Facebook befriending study including email addresses, Facebook IDs, etc., and not to disclose for any purpose. I acknowledge that such information is not to be altered, copied, interfered with or destroyed, except upon authorization from principal investigator of the study. I will not discuss such information with any party, nor will I participate in or permit the release, publication or disclosure of such information, nor will I copy, distribute, or disseminate such information, except as authorized in the course of my employment or by law. I understand that this agreement and undertaking includes:  a. Never discussing the personality of a participants, his or her file or any details about them;  b. Avoiding the use of names of participants in conversations with other clients, friends or relatives;  c. Ensuring that disclosures of information are made only to persons entitled to that information;  SIGNED at __________________________, British Columbia, this __________ day of ______________________________, 2013.  __________________________________     _____________________________ Signature          Witness    T H E  U N I V E R S I T Y  O F  B R I T I S H  C O L U M B I A         96Debriefing,	  Version	  1.1,	  06/09/2013	  	  1	  of	  1	  	  Debriefing	  Our	  main	  objective	  during	  this	  study	  was	  to	  understand	  why	  users	  of	  online	  social	  networks,	  such	  as	  Facebook,	  befriend	  stranger.	  To	  improve	  the	  quality	  one	  of	  the	  team	  member	  sent	  you	  a	  friendship	  request	  that	  you	  have	  accepted.	  We	  inform	  you	  that	  after	  this	  study	  you	  are	  free	  to	  defriend	  our	  team	  member.	  Note,	  that	  we	  have	  not	  collected	  any	  of	  your	  private	  information	  that	  you	  are	  not	  aware	  of.	  The	  friendship	  request	  was	  sent	  to	  you	  so	  that	  we	  can	  be	  more	  specific	  with	  questions	  about	  your	  online	  practices.	  	  Thank	  you,	  OSN	  Befriending	  Study	  Team	     T H E  U N I V E R S I T Y  O F  B R I T I S H  C O L U M B I A         97Appendix BSurvey QuestionsThanks a lot for participating in this survey. In this survey, there are questions aboutyour activities on Facebook. It will take you about 15 to 20 minutes to answer thequestions. For the likert-scale questions, please choose one number from 1 to 5,where 1 means “strongly disagree” and 5 means “strongly agree”.1. What is your age?• 19 to 25• 26 to 30• 31 to 35• 36 to 40• 41 to 45• 46 to 50• 50 to 55• 56 to 6098• 61 to 65• 61 and more2. What is your gender?• Female• Male3. What is your highest level of education completed?• High school• Undergraduate• M.Sc• PhD• Other:4. What is your employment status?• Employed• Student• Retired• Unemployed• Other:5. How long have you owned a Facebook account?• Less than a year• 1 to 2 years99• 2 to 3 years• 3 to 4 years• 4 to 5 years• More than 6 years6. How often do you login into Facebook?• Every hour• Several times a day• Once a day• Several times a week• Once a week• Several times a month• Once a month• I have my account de-activated• Other:7. Please go to your Facebook profile. How many friends do you have on yourFacebook profile?• Answer:8. How often do you receive friendship request?• Everyday• At least once in 2-3 days100• At least once a week• At least once a month• At least once every 6 months• At least once a year• At least once in every two week• Other:9. Have you ever accepted a friendship request from somebody who you do notknow in real life or online communities?• Yes• No10. Check all groups that you would likely befriend on Facebook:• Parents• Siblings• Relatives• Close friends• Friends• Acquaintance• Colleagues• Other:11. If I distinguish the person from the picture, I would accept the friendshiprequest.101• 1• 2• 3• 4• 512. I usually become friends with:• Only females• Only males• I do not care about the gender13. Knowing the number of mututal friends is enough for me to accept a friend-ship request.• 1• 2• 3• 4• 514. If I have mutual friends with the person who sent me a friendship request, Iwould look at the closeness of those mutual friends to me in addition to justthe number of mutual friends.• 1• 2102• 3• 4• 515. If I know somebody in real world or online communities, I would accepther/his friendship request on Facebook.• 1• 2• 3• 4• 516. If I recognize someone’s name, I would accept her/his friendship requests onFacebook.• 1• 2• 3• 4• 517. ( ) of my friends actively share content on Facebook (1: a few, 5: almost all)• 1 (a few)• 2• 3103• 4• 5 (almost all)18. I tend to accept friendship request from everybody, who was born in the sIame city as I.• 1• 2• 3• 4• 519. I tend to accept friendship request from everybody, who lives in the samecity as I do.• 1• 2• 3• 4• 520. I tend to accept friendship request from everybody, who have attended thesame school/university as I do.• 1• 2• 3104• 4• 521. Similarity in personal interests or hobbies is sufficient for me to accept friend-ship requests.• 1• 2• 3• 4• 522. I mostly accept friendship requests from people who share a lot of contenton Facebook.• 1• 2• 3• 4• 523. Users who passively monitor others’ posts on Facebook does’nt motivate meto post less content on Facebook.• 1• 2• 3105• 4• 524. I limit my activities on Facebook because I know my friends are not inter-ested in the content that I post.• 1• 2• 3• 4• 525. I don’t tend to accept friendship requests sent from Facebook applications.• 1• 2• 3• 4• 526. I used to share more content since I felt more comfortable to share contentwith my Facebook friends.• 1• 2• 3• 4106• 527. If my friends shared content irrelevant to me, I would remove them from myfriends list.• 1• 2• 3• 4• 528. I don’t accept a friendship request if I have just common interests or hobbieswith the person who sent me friendship request.• 1• 2• 3• 4• 529. I would accept friendship requests sent from a Facebook application (forexample a game) on behalf of others.• 1• 2• 3• 4107• 530. Who is a Facebook user that you do not want to have a friendship connectionwith on Facebook?• Any body who seems to be annoying (sending weird message, irrele-vant post, etc.) regardless of being known in real life or not. 308• Any body except people that are known to some extent• Any body except for those that have strong connections in real life31. How would you define different levels of access for Facebook friends?• Creating separate lists with different access levels• Using manual exemption feature for each shared content• I do not define different levels of access108


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items