Introduction

Online social networking sites, such as Facebook, have been responsible for facilitating communication between friends and acquaintances, renewing old friendships, and providing information about the activities, interests, and opinions of people’s friends and acquaintances (Ellison et al. 2007). These positive outcomes of social networking sites, however, come with a host of social issues, including concerns about users’ privacy (Boyd and Ellison 2007). User profiles from many social networking sites are now available through search engines like Google and Yahoo (Young 2007). Users, therefore, face a number of threats due to the vast amount of personal information that can be accessed through such a profile.

Facebook, in particular, has spawned its own jargon referring to the “profile browsing” people engage in while using the site. It is not uncommon, in reference to Facebook, to hear or read phrases referring to “profile stalking” (Sarno 2007), “Facebook cyber-stalking” (Harkin 2007), “Facestalking”(Sng 2007), “Stalkbook,” “status creeping,” and just plain “stalking” (Caven 2007).

Thankfully, instances of criminal stalking behavior resulting from online social networking sites are few. But, it appears that these sites do provide an avenue for individuals to access information about and make contact with other individuals. While such behaviors may not constitute stalking in a legal sense, they may constitute a lesser form of termed ‘obsessive relational intrusion” (Haugaard and Seri 2004; Sinclair and Frieze 2005; Spitzberg and Cupach 2003).

To date there are very few empirical studies dedicated to understanding the use and consequences of these new social networking sites. Most of these studies have been restricted to examinations of frequency of visitation by members and its benefits (Ellison et al. 2007), differences between users and non-users (Hargittal 2008), and the uncertain meaning of online ‘friend’ status (Tong et al. 2008). Our question is, do these online social networking sites, in particular Facebook, encourage obsessive relational intrusive behaviors and if so, what general types of these behaviors occur?

We begin by reviewing the literature on relational intrusive behaviors and stalking. Then, we examine online social networking sites, specifically Facebook, in an attempt to draw a parallel between the functions of these sites and behaviors commonly depicted as relational intrusion in offline relationships. Next, we present the methodology and results of a frequency analysis of these online social network behaviors. We conclude with a discussion of the results and the implications of these findings for consumers, marketers, and the developers of such websites.

Literature Review

Relational Intrusion, Stalking, and the Like

Stalking is an issue that has received a growing amount of attention (Meloy 2007). In general, the legal definitions of stalking contain three components: 1) a repeated pattern of intrusive behavior exhibited by the perpetrator that is unwanted by the victim; 2) a threat, implicit or explicit, made by the perpetrator; and 3) a resulting feeling of fear in the victim (Meloy 2007; Sheridan and Davies 2001). Further definitions submit that the perpetrator’s behavior must be judged as intentional (Spitzberg and Cupach 2003). The behaviors associated with stalking have a wide range, and can be grouped in a number of ways. Nadkarni and Grubin (2000) grouped the types of behaviors into three broad categories. The first type, surveillance, includes such behaviors as orchestrating accidental run-ins with the victim, and appearing frequently at places such as the victim’s work. The second type, communicating, includes phoning, emailing, sending letters, or sending gifts. The last type includes behaviors that have escalated to aggression or violence where the victim’s physical being or property are threatened or harmed (Nadkarni and Grubin 2000).

The conceptualization of stalking includes victims that may or may not be related to the perpetrator; however, there exists another form of pursuit where prior acquaintance (either real or delusional) of some degree is assumed by the pursuer. Such pursuit has been termed obsessive relational intrusion (ORI) and is defined “as repeated, unwanted pursuit and invasion of one’s sense of physical or symbolic privacy by another person, either stranger or acquaintance, who desires and/or presumes an intimate relationship” (Spitzberg and Cupach 2003, p. 34). While ORI is related to stalking it does “not meet the legal standard for stalking” (Dutton and Winstead 2006). The various terms, relational intrusion and stalking, may be viewed on a continuum with some behaviors classified as “mild” forms of intrusion (e.g., leaving messages) while others at more intimidating levels (e.g., obsessive following/pursuit) are considered stalking. While relational intrusion may initially consist of somewhat benign behaviors, they can and often do escalate into more aggressive forms of relational pursuit or even stalking (Nadkarni and Grubin 2000; Ravensberg and Miller 2003).

Our focus is on obsessive relational intrusion, specifically online obssessive relational intrusion (o-ORI). It has been suggested that increases in technology, especially the new social networking phenomenon, have made it easier for potential stalkers to access their victims (Spitzberg and Hoobler 2002). The purpose of this research is to determine whether or not online social networks provide an environment in which relational intrusive behavior can occur and, if so, to examine the form in which this o-ORI occurs.

Social Networking Sites—Issues of Privacy and Relational Intrusion

Facebook is described as “a social utility that connects people with friends and others who work, study, and live around them. People use Facebook to keep up with friends… and learn more about the people they meet”(Skiba 2007). Lampe et al. (2006), found that one of the main reasons people use Facebook is to engage in “social searching”: that is, to investigate those in their offline community.

The amount of personal information that Facebook allows users to publish online is working to alter the idea of personal privacy (George 2006). With more than half of Facebook’s users logging on every day, it may be no surprise that Facebook has popularized stalker-like behavior (Govani and Pashley 2005). However, the question remains as to whether these “stalker-like behaviors” are indeed representative of the legalized conceptualizations of stalking, or whether they are more analogous to the different and lesser form of deviant behavior, known as obsessive relational intrusion (ORI) (Spitzberg and Cupach 2003).

The primary goal of most social networking sites is to connect people who already have an offline relationship with one another (Lampe et al. 2006). And while information has become privy to an audience beyond that which the author intended, there also exists potential for the user’s information to be exploited by people they know. In other words, there exists the potential for o-ORI. Not only are the types of behaviors inherent in offline ORI comparable to those found on Facebook, but so are the contexts in which they occur. Studies show that in cases of offline ORI, the pursuer was often identified as an ex-partner, a friend, or an acquaintance (Spitzberg and Rhea 1999). These seem directly related to the relationships which Facebook works to maintain through its network structure. Spitzberg and Hoobler (2002) coined the term cyber-obsessional pursuit for the online version of ORI. In their research, they focused on the way in which aspects of stalking and ORI are translated back and forth from ‘real life’ to the electronic world. However, as with much of the research on social networking, the direction of relationship formation was notably biased; in the questionnaire developed in their study, many of the activities were precluded by an online association, and then followed by offline behaviors. In contrast, prior ‘real-life’ familiarity with the person being ‘stalked’ would be more analogous to the type of online connections Facebook fosters. Furthermore, their study did not explore cyber-stalking in the context of a social network.

Online (Obsessive) Relational Intrusion

Current research in the area of stalking and obsessive relational intrusion (ORI) has identified a number of ‘tactics’ (i.e., behaviors) that are indicative of the phenomenon. Spitzberg and Cupach (2003) developed a 63-item self-report measure to assess the extent to which individuals were victims or perpetrators of these various tactics. Their development and refinement of this measure has resulted in a 23-item assessment consisting of a variety of tactics that range from simple attempts to make contact (e.g., sending messages) and escalate to surveillance and finally to physical threats.

To extend this work to a social media application such as Facebook, we began by identifying similar ‘tactics’ that are available to users within the functionality of Facebook. We identified 38 different Facebook activities (e.g., sending gifts, posting on walls, inviting to events) that were similar in nature to the ORI tactics identified by Spitzberg and Cupach (2003). For example, the Facebook application of sending gifts such as virtual flowers or other similar gifts is conceptually similar to the ORI tactic of “leaving unwanted gifts.” Within the 38 Facebook behaviors we found correspondence with 12 of the ORI tactics and present these in Table 1. This similar functionality, however, begs the question of the extent to which these behaviors are cause for concern. To examine this, we selected a sample Facebook behavior from each of the 12 ORI tactics and presented these to an expert panel consisting of three researchers (1 male, 2 female) well published in aggression and violence research. They were presented with each behavior and asked to classify each as either a potentially dangerous behavior, a definitely dangerous behavior, or a benign behavior. The panel approached the task from the perspective of a Facebook user with an ex-intimate as a Facebook friend. There was good agreement between the panel (inter-rater reliability = 71%; Cohen’s Kappa = .42. see Table 1) with the majority of disagreements occurring between the behaviors that were classified as being either potentially or definitely dangerous. Two of the 12 tactics were identified as benign behaviors while the remaining were either potentially dangerous (7) or definitely dangerous (3).

Table 1 Mapping ORI ‘tactics’ to facebook functionality
Full size table

In summary, the functionality of Facebook provides a platform for relational intrusive behaviors, many of which have been identified as potentially dangerous. If actual users of Facebook, and by extension users of online social networking sites engage in these behaviors, then this provides evidence that the sites themselves facilitate a form of ORI. To examine this, we conducted a frequency analysis of user behaviors on Facebook. This is detailed next.

Methodology and Results

Subjects and Procedures

To conduct a frequency analysis of user behaviors in a social media context, we distributed an online survey to 1022 respondents recruited from the email lists of an undergraduate business faculty at a large Canadian university. We received 230 respondents that included an even distribution of males and females, the majority of whom were between the ages of 18 and 25. The survey was voluntary and respondents were entered into a draw for cash prizes.

Each respondent was randomly assigned to one of six groups based on two variables. We assigned respondents to report either their behaviors (i.e., as “doers” of the o-ORI tactics) or the behaviors of others towards them (i.e., as “receivers” of the o-ORI tactics). Further, respondents were asked to report these behaviors with reference to either an ex-intimate (i.e., someone with whom they had previously had an intimate relationship), a close friend, or an acquaintance. Descriptive statistics of the sample are provided in Table 2.

Table 2 Description of sample
Full size table

Measurement

To assess the extent to which respondents engaged in or were receivers of o-ORI tactics, we used the same list of 38 behaviors that were identified given the functionality of Facebook and that fit within the 12 behaviors identified as forms of ORI given in Table 1. The items (i.e., behaviors) were presented in an order from relatively low intrusion (e.g., sending messages) to relatively moderate intrusion (e.g., keeping tabs on the person through Facebook). The response scale was one of usage frequency: 0 = never, 1 = only once; 2 = 2 to3 times, 3 = 4 to 5 times; 4 = more than five times; consistent with the original ORI form.

Results

We present results of our frequency analysis in Appendix A. Frequencies represent percentage of respondents that reported either having engaged in this behavior (“doers”) or perceived that they had been targets of this behavior (“receivers”). The frequencies represent the percentage of those that had reported that behavior at least once.Footnote 1

For both groups (“doers” and “receivers”), frequency of relational intrusive behaviors was highest for close friends. There were few differences between the two with respect to frequencies of behaviors save for some behaviors that were unknown to victims (e.g., monitoring-like behaviors such as reading one’s wall or obtaining information about the person). Respondents that were assigned to the ex-intimate or acquaintance groups reported significantly higher frequencies (p < .05Footnote 2) of offending behaviors (i.e., they had done the behavior) versus being targets of the behaviors. Again, these differences appear for monitoring-like behaviors such as visiting the groups that their partners were members of, reading their mini-feeds, or checking the person’s feeds for updates.

We compared frequencies of behaviors across groups within the “doers” category. For the majority of behaviors, we found that the frequencies were highest for the closest friends category and that these were significantly more prevalent (i.e., p < .05) than those of the ex-intimates or acquaintances categories. Nearly half of the behaviors occurred equally frequently between the ex-intimate group and the acquaintance group.

Since research in the area of relational intrusion and stalking has classified various ORI tactics into several types (e.g., communication, surveillance), we endeavored to do the same. To do so, we employed exploratory factor analysis using principal axis factoring and oblique rotation. The general criteria for factor definition were: 1) eigenvalues greater than one; 2) factors prior to the leveling of the scree plot, 3) minimal secondary loadings; and 4) a minimum of three items per factor. We ran separate factor analyses for the two groups of “doers” and “receivers.”

For the “doers” group, eight factors produced eigenvalues greater than one with a scree suggesting leveling between six and seven factors (KMO = .80). Successive extraction and rotation produced a five-factor solution (see Table 3). Poorly loaded items (i.e., items with secondary loadings of greater than .40 or without any loadings of .50 or higher) were removed. The remaining five-factor solution accounted for 70% of the common variance. We identified the five factors as primary contact (e.g., behaviors aimed at making face-to-face contact with the person), monitoring (e.g., looking at the person’s information and profiles), secondary contact (e.g., contacting others within the person’s Facebook network), expressions (e.g., sending the person intimate messages, virtual flowers, etc.), and invitations (e.g., sending the person invitations to groups, events, communities).

Table 3 Oblique-rotation pattern matrix for principle axis factor analysis of the o-ORI items for offenders
Full size table

For the “receivers” group, eight factors produced eigenvalues greater than one with a scree suggesting leveling between four and five factors (KMO = .88). Successive extraction and rotation produced a three-factor solution (see Table 4). The three-factor solution accounted for 75% of the common variance. We identified the three factors as primary contact (e.g., sending the person invites), expressions (e.g., sending Facebook gifts, sending kisses/hugs); and secondary contact (e.g., posting on the walls of the person’s friends/family/coworkers).

Table 4 Oblique-rotation pattern matrix for principle axis factor analysis of the o-ORI items for targets
Full size table

A general discussion of these results follows.

Discussion

The goal of this research was to determine whether or not Facebook facilitated relational intrusion-like behaviors. We find evidence of behaviors identified in the research on stalking and relational intrusion and find that many of them are facilitated by the Facebook application.

Examination of frequency of these behaviors in the ex-intimate groups reveals some interesting data. While respondents in the other groups reported relatively high frequencies of each of the behaviors, there were some behaviors reported by the ex-intimate groups at the same level of frequency. Because obsessive relational intrusion, and ultimately stalking, is most likely to occur between ex-intimates (Spitzberg and Cupach 2006; Spitzberg and Rhea 1999), there may be cause for concern the extent to which this may occur through online social networks. For example, more than half responded that they had used Facebook to make contact with ex-intimates. Further, almost a third responded that they used Facebook to “keep tabs” on the person and to obtain information about the person’s activities by reading their wall and profile. These types of behaviors represent a form of monitoring and surveillance which is consistent with lower forms of obsessive relational intrusion (Spitzberg and Rhea 1999). Of concern is the fact that Facebook allows this behavior to occur in relative anonymity—i.e., it is near to impossible to determine who has been visiting one’s space on Facebook and how often. Furthermore, research in this area reports that between 5% and 40% of college students in the United States have experienced obsessive relational intrusion (Spitzberg and Cupach 2006).

To date, there has been relatively little empirical research on the phenomenon of online relational intrusion. This study is an initial foray into this field and we find evidence of five types of behaviors facilitated by Facebook as reported by offenders. We find that offenders use Facebook to facilitate primary contact by providing information about where a target might be (e.g., at specific events advertised on Facebook, or showing up at locations mentioned by the target in their profile). Because many students update their status regularly on Facebook, it provides others with information about where they may have face-to-face contact with the person of interest. Should these others include those with pursuit goals, this may facilitate unwanted interaction between the two. Facebook also permits offenders to send expressions of affection to other people using a host of virtual gifts such as flowers or even virtual kisses, hugs, or caresses. This functionality allows one to move from simple contact through messaging features and monitoring to more active pursuit of a relational interest. Further, offenders reported making secondary contact with targets through members of the target’s network.

Three of these behaviors—primary and secondary contact as well as expressions, are also reported by targets. Monitoring behavior occurs anonymously, thus most targets were unable to report these types of behaviors occurring towards them. Invitations were subsumed under the factor ‘primary contact.’ Of concern to users of Facebook should be that offenders report significant monitoring behaviors (akin to lurking) that are not noticed by the target.

More than half of the sample reported being on Facebook several times a day. According to the results of this study, it appears that much of the time on Facebook is spent monitoring the activities of others and making incidental contact with others. Further, many of the sample reported having in excess of 200 ‘friends’ in their Facebook network. This suggests that the majority of these relationships on these sites are mostly superficial and that users seem to use little-to-no criteria in terms of who is added to their friends list.

This research has led to a number of new avenues for inquiry into the effect of social networking sites on society in general. The limitations of our study are many. Our research did not specifically ask about ‘stalking’ or “unwanted relational intrusion,” as we wished simply to report the frequencies of behaviors on Facebook. Our conclusions are limited to the fact that behaviors like o-ORI or stalking “may,” or certainly “can” occur using Facebook as a medium. We did not examine the extent to which people feel threatened by others’ use of the Facebook medium. Further research should examine the extent to which students have felt threatened or disturbed by behaviors on these sites. While we have evidence of lower forms of relational intrusion occurring through Facebook, further research should examine the extent to which these lower forms escalate into higher forms of relational intrusion over time.

Prescriptions for User Privacy

Our findings suggest that social networking sites can be avenues for obsessive relational pursuit behaviors. These findings have a number of implications for users of these sites and for the developers of these sites. Users of social networking sites are well-advised to be selective in their criteria for adding ‘friends’ and be wary of individuals that begin to exhibit higher forms of relational intrusion behaviors (e.g., sending unwanted messages of affection, unwanted virtual gifts). In addition, users should make themselves aware of the various privacy settings that are available through the site. For example, Facebook allows one to restrict the majority of their profile to friends only.

Developers of social networking sites should be cognizant that despite the many healthy, interpersonal connections that they facilitate through technology, their sites do provide a means by which one may essentially ‘stalk’ a target of interest in relative anonymity. Developers should put warnings on the sites, develop functionality so that users may see who is visiting their profiles and how often. In addition, Facebook should develop additional functionality that allows users to distinguish between levels of friendship. For example, those identified as acquaintances may only have access to some portions of another’s profile. Those identified as close friends may have higher access. Currently, Facebook does allow users to block some individuals and to limit profile viewing to others. Users should be educated on these privacy settings. One user commented on privacy settings as follows:

Ever since I had Facebook my privacy settings have been on the highest setting. Unless I have accepted the person as my friend they cannot see my profile. I only accept friend requests from people I know to a degree higher than simply knowing their name. Unfortunately, people are not as careful with their privacy as I am.

In this age of ubiquitous social networking sites such as MySpace and Facebook, casual "stalking" of friends and acquaintances is quite common. Many feel that these sites implicitly encourage this type of behavior. Young Internet users can access more personal information on friends and acquaintances than past generations could ever imagine. One respondent commented, “I recently removed my Facebook profile, after [I] was continually being contacted by people from my past that I preffered (sic) not to be contacted by. It was causing way too much drama in my life.” This comment and our findings are indicative of what we term “online relational intrusive behaviors” facilitated by these social networking sites. We hope that this initial inquiry into this phenomenon activates further research in this area.