The Hyperperception Model: What happens when we watch others online?

Breaking Up is Hard to Do, Especially When Facebook Won’t Let You

Erin L. Spottswood Department of Communication, Portland State University,

Portland, OR, USA [email protected]

Christopher J. Carpenter Department of Communication, Western Illinois

University, Macomb, IL, USA [email protected]

ABSTRACT Remaining connected on social network sites (SNS) with one’s former partner after the termination of a romantic relationship (breakup) is sometimes associated with negative outcomes, possi- bly due to observing former partner interactions with others on these sites. This study used the hyperperception model to explore motives for observing partner SNS interactions and to make unique predictions about what types of partner behaviors inhibit recovery from breakups. An online survey was conducted with students at two American universities (N = 214). The survey results suggest that people who want to reunite with their partners observe their partners on SNSs to determine if they have new relationships. Re- sults also suggest that recovery from breakups may be inhibited by comparing oneself to new people in the ex-partner’s SNS list of connections, and observing the ex-partner interacting with people one does not know, consistent with hyperperception model pre- dictions. These findings suggest that breakup recovery is hindered when people remain connected to their ex-partners on social media, especially if those ex-partners have many friends unknown to those people.

CCS CONCEPTS • Human-centered computing → Collaborative and social com- puting; Collaborative and social computing, Collaborative and so- cial computing theory, concepts and paradigms, Social networks; • Applied computing → Law, social and behavioral sciences.

KEYWORDS Social network sites, interpersonal communication, relationships, hyperperception

ACM Reference Format: Erin L. Spottswood and Christopher J. Carpenter. 2020. Breaking Up is Hard to Do, Especially When Facebook Won’t Let You. In International Conference on Social Media and Society (SMSociety ’20), July 22–24, 2020, Toronto, ON, Canada. ACM, New York, NY, USA, 9 pages. https://doi.org/10.1145/3400806. 3400811

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]. SMSociety ’20, July 22–24, 2020, Toronto, ON, Canada © 2020 Association for Computing Machinery. ACM ISBN 978-1-4503-7688-4/20/07.. .$15.00 https://doi.org/10.1145/3400806.3400811

1 INTRODUCTION: POST BREAKUP SURVEILLANCE

The end of a romantic relationship can be one of the most painful events a person experiences. Social Network Sites (SNSs) make it possible for someone to surveil their ex-partner on these sites after a breakup. Research suggests that remaining tied to one’s ex-romantic partner on Facebook can hinder break-up recovery [7, 15, 16]. Perhaps, this is because remaining connected gives people the opportunity to repeatedly surveil or observe their ex- partner’s online and offline activities and interactions. For some, these observations can lead to thoughts and feelings that undermine breakup recovery (e.g., “my ex-partner seems to be doing better without me” or “my ex-partner seems to be in a new romantic relationship”).

Despite the strong research that has examined how interpersonal electronic surveillance (IES) [27] undermines breakup recovery [7], it is still unclear what motivates IES and if certain motivations in particular make it especially difficult to recover from a breakup. To that end, this study uses the hyperperception model [23] to uncover the motivations that can illuminate why and how surveilling an ex-partner on SNSs can be psychologically and emotionally harm- ful. Specifically, the model explains why aspects of the channel (e.g., Facebook), sender (i.e., the ex-partner), and receiver (i.e., the person/people the ex is interacting with on the site) can make it seem that an observer’s ex-partner has already moved on, leaving a user/observer to potentially feel upset. This study will therefore help contribute to our understanding of what kinds of SNS use produce psychologically unhealthy outcomes, given that failure to recover from breakups is associated with depression [5]. This study will also continue the ongoing development and testing of the hyperperception model [23, 24].

This paper will first describe the hyperperception model and apply it to the phenomenon of people using SNSs to observe their ex-partner’s interactions. Then it will describe the methods used to assess how the motivations behind these observations are detrimen- tal to breakup recovery. Next, the results reveal a pattern consistent with the model. Finally, the discussion section will explicate how the results are consistent with key aspects of the hyperperception model and past IES research.

1.1 The Hyperperception Model The hyperperception model was developed to explain phenomena in which a SNS user makes over-attributions about other users’ in- teractional and relational intimacy according to what they observe on the site. It is an extension of the hyperpersonal model [29] which focuses on the heightened relational intimacy sometimes found between computer-mediated communication partners. The major

30

SMSociety ’20, July 22–24, 2020, Toronto, ON, Canada Erin Spottswood and Christopher Carpenter

difference between the two models is the difference in focus: instead of focusing on interaction partners like its predecessor, the new model focuses on observers of others’ interactions on SNSs as well as other communication technologies. The original model is unde- niably robust and will continue to generate a lot of innovative and compelling research. However, it was not designed to theorize and empirically explore what occurs when users observe and appraise others’ SNS interactions, which seems to dominate many people’s time spent on SNSs [11, 18]. Building off of the key components of the hyperpersonal model, the hyperperception model explicates in this context how observing an ex-romantic partner (i.e., observed sender) and another person (i.e., observed receiver) on an SNS (i.e., the communication channel) can motivate the observer to look at past, recent, and future SNS interactions (i.e., feedback loop) be- tween the dyad and develop an inflated impression of the dyad’s interactions and relationship. That inflated impression of a dyad’s intimacy is labeled a hyperperception.

This paper will focus on three of the four components of the hyperperception model: the channel, the observed sender, and the observed receiver. The channel of interest in this study is the SNS Facebook, which possess affordances and is endowed with social norms that may increase the likelihood of hyperperception effects. In terms of affordances, Facebook 1) gives users the ability to access others’ activities, posts, and interactions on the site, 2) keeps logs of that information such that those logs persist digitally on the site unless removed by a user, and 3) associates that information on the site to its users who are relatively identifiable depending on what they disclose on the site (profile, timeline, and newsfeed). In terms of social norms, Facebook is endowed with social norms that encourage users to post flattering content about themselves and ensure their public interactions are positive and affiliative [25, 30]. If an observer forgets or is not aware of these norms, they may assume that those they are observing and posting in ways that express their attraction to each other rather than merely upholding channel norms. In general, people do tend to attribute other people’s behaviors more to those people’s dispositions rather than their context [8]. These are the aspects of the channel that can encourage hyperperception.

The second component of the model specifies that, in order for hyperperception to occur, the observer has to be motivated to surveil a particular sender’s posts and interactions with other users on the site. This particular sender thus becomes the observed sender in the hyperperception model. This is partly inspired by research examining interpersonal electronic surveillance (IES), which is the use of communication technology – such as SNSs – to observe other user’s online and offline behaviors and interactions [15, 27]. In fact, LeFebvre et al.’s [15] participants discussed using Facebook to see what their ex-partner was up to as well as evaluate the ex-partner’s new romantic partners. This behavior can be innocuous, or it can hinder break-up recovery.

The third component of the hyperperception model relevant to the current study is the observed receiver component. Just as the original hyperpersonal model predicts that it is easier to idealize someone that you encounter online rather than offline, the hyper- perception model argues that the observer is likely to form a height- ened perception of the interactions between the observed sender and the observed receiver when appraising their online rather than

offline interactions. In the hyperperception case, when the observer is constrained to only observing interactions on an SNS, they do not have a variety of contextualizing information. If the observer saw the observed sender and receiver interacting offline, those two might say similar kinds of things (e.g., discussing topics of common interest) but the nonverbal immediacy cues would indicate that they are not particularly close with each other. Similarly, if the observer could see the observed receiver interacting with other people on the SNS, the observer would likely see that observed receiver acting in similar ways to how the observed receiver acts with the observed sender. That would enable the observer to realize there is nothing particularly special about the interactions between the observed sender and observed receiver. This constrained access to an SNS for observation makes hyperperception effects more likely to occur.

1.2 The Effects of IES on Breakup Recovery Some studies have examined the detrimental effects of IES on breakup recovery [7, 15, 16], but this initial research has not em- pirically demonstrated why IES makes it harder for some but not all people to move on and emotionally heal after a breakup. This is in part because IES is merely the behavior. What is needed is an additional exploration of the fundamental motivations that under- mine breakup recovery when a person engages in IES. Examining the motivations behind certain types of IES would help give SNS researchers – as well as everyday users – a better understanding of when IES can hinder breakup recovery and possibly be emotionally and psychologically harmful for those doing the surveilling.

After breakup, a person may wonder if or how long their ex- partner wanted the relationship to end, and if their ex-partner wanted to start a relationship with someone else while still in a relationship with them. That would explain why someone (an ob- server) might use Facebook (the channel) to see if their ex-partner (observed sender) was interacting with or had already started a relationship with a new partner (observed receiver). The observer might even scroll through their ex-partner’s (observed sender) Face- book timeline to see how frequently and intimately their ex-partner had been interacting with a new potential partner (observed re- ceiver) on the site. The hyperperception model may be helpful in uncovering these motivations because according to the model, ob- servers are more likely to make hyperperceptions of other user’s interactional and relational intimacy when they are especially in- terested in another person’s (i.e., observed sender’s) interactions with a potential rival (i.e., observed receiver).

1.3 Motivations for IES that Undermine Breakup Recovery

Two motivations may be key in understanding the detrimental ef- fects of IES and potential hyperperceptions between an ex-partner and other SNS users: the first factor is the desire to know if an ex- partner has any new romantic connections, and the second factor is their motivation to compare themselves to the ex-partner’s po- tential new romantic connections. In terms of the first motivation, sometimes people want to know who their ex-partner is roman- tically pursuing or dating after a breakup [17, 28]. Moreover, this knowledge has been found to negatively affect those people’s ability to adjust or recover after the breakup [12]. This is consistent with

31

Breaking Up is Hard to Do, Especially When Facebook Won’t Let You SMSociety ’20, July 22–24, 2020, Toronto, ON, Canada

SNS research that shows people take advantage of these technolo- gies to see if their ex-partner has any new romantic connections [28]. The motive to know if an ex-partner has any new romantic connections might suggest that the observer is still romantically interested in their ex-partner and wants to know if the ex-partner has moved on or could be willing to reconcile. If their ex-partner’s posts and interactions with others on an SNS suggest they have moved on and might even be dating someone new, the observer may feel hurt, which in turn would negatively affect their ability to recover from the breakup.

In addition, if the observer perceives their ex-partner has es- tablished potential or actual romantic connections based on their ex-partner’s interactions with others on a SNS, they may proceed to compare themselves to the new romantic connections. Research has shown that comparing oneself to a potential romantic rival is positively related to feelings of romantic jealousy [1]. Fox and Moreland’s [6] survey of Facebook users found that people some- times compare themselves to their ex-partner’s potential and actual new romantic connections according to what they can access on the site. Relatedly, research suggests that when people believe their ex-partner has better romantic alternatives than they themselves do, they experience more distress [22]. Recall that the hyperperception model predicts that when an SNS is endowed with positivity norms (e.g., Facebook), it will lead an observed sender and an observed receiver to post and interact in ways that flatter themselves and each other. If these posts are all the observer has to make their appraisal, they may perceive their ex-partner’s potential new ro- mantic connection is more attractive and interesting than they are themselves, which will likely lead that observer to feel down about themselves and further undermine the breakup recovery process. In summary, the motivation to know if the ex-partner has any new romantic connections and the motivation to compare themselves to the ex-partner’s potential or actual romantic connections will undermine breakup recovery when these motivations are the rea- sons why someone observes/surveils their ex-partner’s Facebook interactions.

1.4 Hypotheses 1.4.1 Relationship Status. This study seeks to develop the observed sender component of the hyperperception model by uncovering what motivates an observer to take advantage of Facebook to ob- serve their ex-partner and assess their interactions with others on the site. The observed sender component specifies that such a motivation needs to exist for hyperperception to occur, but it does not yet specify all the potential motives. An important individual variable that may encourage IES is whether the observer is single. Being single in general is associated with feelings of loneliness after the end of a romantic relationship [19]. People who find new part- ners may recover from the relationship dissolution quicker [13] and thus be less motivated to observe their former partners or assess their partner’s attempts at forming new relationships. Moreover, as a result of being single, the observer may have more time to peruse Facebook and in turn be more likely to encounter their ex-partner’s posts and interactions on the site (fodder for hyperperception pro- cesses to take place). In light of these factors, this study predicts that:

H1: Relative to people who are in a new relationship, people who are single a) will engage in more IES of their ex-partner, b) report a stronger motive to surveil their ex-partner to determine if their ex-partner is developing a new romantic relationship, and c) report a stronger motive to compare themselves to their ex-partner’s FB friends of the same sex as themselves.

1.4.2 Desire for Reconciliation. Another individual variable that can lead an observer to take interest in their ex-partner’s FB interac- tions is whether or not the observer wants to reunite with their ex. It is not uncommon for people to want to get back together with an ex-partner after a breakup [14]. It is also not uncommon for people to resume and repair a romantic relationship after a break-up [3, 4]. If the observer wants to get back together with their ex-partner, they are more likely to want to know who their ex-partner has been recently interacting with to get a sense of whether or not the potential of a reconciliation is likely or possible. In particular, they may wish to know if a new romantic partner stands between themselves and reconciliation and if so, determine the merits of this new person relative to themselves. Therefore, this study predicts that:

H2: People’s desire to reunite with their ex-romantic partner is positively related to a) IES of their ex-partner, b) motivation to surveil their ex-partner to determine if their ex-partner is devel- oping a new romantic relationship, and c) motivation to compare themselves to their ex-partner’s FB friends of the same sex as them- selves.

1.4.3 Poor Recovery from Relationship Dissolution. Surveilling an ex-partner on Facebook can hinder break-up recovery and incline an observer to continue to surveil and assess their ex-partner’s interactions with others on the site [7]. This pattern occurs because the notion that an ex-partner is moving on generally, as well as romantically, can lead the observer to wonder how their ex-partner seems to be moving on so quickly and easily. These ruminations can in turn make it difficult for the observer to recover and could lead them to feel that their partner did not love them as much as they thought or that their relationship was not as deep as they felt [21]. These thoughts and feelings may incline the observer to continue to surveil and assess their ex-partner’s interactions with others on the site, further hindering their recovery from the breakup. As described above, this pattern of surveillance is likely to be motivated by interest in new people the ex-partner may be developing new romantic relationships with. In light of these possibilities, this study predicts that:

H3: Breakup recovery is negatively correlated with a) IES of their ex-partner, b) motivation to surveil their ex-partner to determine if their ex-partner is developing a new romantic relationship, and c) motivation to compare themselves to their ex-partner’s FB friends of the same sex as themselves.

1.4.4 Constrained Access. As alluded to before, sometimes surveilling an ex-partner on Facebook may indicate that the ex- partner is interacting with potential new partners. What may make this especially worrisome is if the observer is not connected to those potential new partners offline, as this makes it also less likely they will be connected to them on Facebook [31]. The hyperper- ception model’s receiver component predicts that these constraints

32

SMSociety ’20, July 22–24, 2020, Toronto, ON, Canada Erin Spottswood and Christopher Carpenter

against accessing the potential new partner’s on-site and off-site interactions may make it difficult for the observer to contextualize their ex-partner’s interactions with the potential new partner on the site. The receiver assumption of the hyperperception model indicates that this lack of access will hinder the observer’s ability to accurately determine if the unknown person is possibly pursuing or developing a romantic relationship with their ex-partner (and vice versa). Additionally, not knowing the people one’s ex-partner interacts with prevents the observer from seeing if a potential new partner is very affiliative in all of their interactions on the site and is in fact not paying their ex-partner any special attention. The hyperperception model emphasizes that not being able to access these kinds of information may leave the observer wondering and worrying if their ex-partner is pursuing or being pursued by a po- tential new romantic partner. This concern should make it even more difficult for the observer to recover from the break-up. There- fore, this study predicts that constrained access could play a key role in hindering recovery from a breakup:

H4: Breakup recovery has a stronger negative correlation with the extent to which their ex-partner interacts with potential new partners that the subject does not know offline than interaction with potential new partners that the subject does know offline.

2 METHOD 2.1 Participants Participants (n = 101) were recruited from a Communication De- partment subject pool at a regional, Midwestern state university in the United States, and they received extra credit for their par- ticipation. More subjects (n = 103) were recruited by emailing the students at the same university an invitation to participate for a chance to win $30 in a raffle. An additional sample was recruited from a regional state university in the Pacific Northwest (n = 10). To participate, they were required to have formerly been involved in a romantic relationship such that they have not reunited with their former partner, to currently use Facebook, and to be Facebook friends with their ex-partner. There were 214 participants (64 iden- tified as male, 147 female, 2 transgendered, and 1 preferred not to answer) with ages from 18-46 (M = 21.66, SD = 3.54).

2.2 Procedure All participants were given a link to an online survey. First, the in- formed consent was displayed. Then all participants were screened with a series of yes or no questions to make sure they met the study qualifications. The measures followed in the order described, with some additional measures not relevant to the current study omitted. Participants were then given a link to register for the raffle or record their participation for the subject pool.

2.3 Measures A confirmatory factor analysis was conducted using the centroid algorithm in the lessR statistical package for R statistics [9]. A four- factor model that included the multi-item measures was tested. The data were consistent with the model after a few changes described below as indicated by the low average error (average absolute resid- ual = .04, a measure similar to the standardized root mean residual,

SRMR). All means, standard deviations, and alpha reliability esti- mates can be found in Table 1.

The surveillance measure included the 7 items from the Fox & Tokunaga [7] post-breakup IES such that the items were phrased in the present tense to indicate ongoing IES. This measure contained items such as “I notice when my ex-partner updates his/her Facebook page.” 1-7 Likert response options were used such that higher scores indicated more frequent IES of the ex-partner on Facebook. The distribution was somewhat positively skewed.

Next, the two motives to surveil the ex-partner were measured with a 1-7 Likert scale with higher scores indicating greater de- gree of that motive. First was the four-item Tong [28] measure of participants’ motives to try to use Facebook to determine if their ex-partner has started a new relationship. Four new items mea- sured the motive to surveil their partners’ interactions to compare themselves to potential new partners. The items were phrased to refer to people their partner interacts with that were the same sex as themselves because most participant’s ex-partners are, at least, sexually attracted to people of the same sex as the participant. This new measure included items such as “It is important for me to know if I am as attractive as the other people of the same sex as I am that my ex knows on Facebook” and “I sometimes look at what other people who are the same sex as I am are saying to my ex on Facebook so I can see if they are as interesting as I am”. The distributions for both motives were positively skewed.

Then there were two items measuring partner interaction with potential new partners. Both used frequency response scales rang- ing from “1-Never” to “6-Constantly.” The first asked “How often do you see your ex interact with non-family members on Facebook who are of the same sex as you are, that you have personally met offline?” and the second changed the final phrase to “you have NOT person- ally met offline?” The phrasing was chosen to focus on potential new partners without suggesting that directly.

Then the participants were asked to indicate their relationship status as “Single” (43.67%), “Romantically involved with someone else” (55.7%), or “Other” (one participant chose “Other” and typed “Engaged”). The “Engaged” participant was recoded as romantically involved.

Next participants were asked about their desire to reunite with their ex-partner using the item from Tan, Agnew, VanderDrift, and Harvey [26], which asked, “On a scale from 0 to 10, how much would you like to reunite with your ex in a romantic relationship at any point in the future?’ (0 = absolutely not and 10 = definitely).” The distribution was substantially positively skewed.

Finally, the participants were asked to respond to the Sbarra and Emery [20] acceptance of relationship termination measure to measure the extent to which the participants had successfully recovered from the termination of their former relationship. Three items (3rd, 7th, and 8th) were dropped to improve factorial valid- ity as the fit diagnostics provided by lessR indicated there was a substantial amount of unique variance in these items not shared with the factor. The response scale ranged from “1-not at all my feelings” to “5-very much my feelings.” The measure was coded so that higher scores indicated greater recovery. The distribution was negatively skewed.

33

Breaking Up is Hard to Do, Especially When Facebook Won’t Let You SMSociety ’20, July 22–24, 2020, Toronto, ON, Canada

Table 1: Descriptive Statistics

Variable M SD Alpha Surveillance 3.28 1.58 0.92 Motive: New 3.38 1.55 0.82 Motive: Compare 2.73 1.69 0.92 Interacting Known 2.56 1.3 - Interacting Unknown 2.71 1.38 - Want to Reunite 2.47 2.99 - Recovery 6.2 0.88 0.88

3 RESULTS The first set of hypotheses predicted that, relative to people who are in a new relationship, those who are single will engage in more IES of their ex-partner on Facebook and report higher levels of both motives. A series of t-tests were conducted to test this hypothesis. First, the surveillance scores were higher for single people (M = 3.64, SD = 1.64) versus those in new relationships (M = 3.03, SD = 1.49), t (212)=2.84, p= .005, r = .24 (H1a). Second, the scores for the motive to see if the partner was with someone new were higher for single people (M = 3.71, SD = 1.56) than those in new relationships (M = 3.16, SD = 1.51), t (212) = 2.58, p = .01, r = .17 (H1b). Finally, the scores on the measure of motive to compare themselves with potential new partners of their ex were higher for single people (M = 3.12, SD = 1.85) versus those in new relationships (M = 2.46, SD = 1.52), t (212) = 2.84, p = .005, r = .18 (H1c). These results were consistent with H1.

The second set of hypotheses predicted that the participants’ desire to reunite with their ex-partner would be positively related to surveillance and both motives. Examination of the correlation matrix in Table 2 shows that the predicted correlations with the desire to reunite with their ex were all statistically significant and in the predicted direction, though the correlation with surveillance was the strongest. In order to determine which of the two motives was more strongly associated the desire to reunite with the ex, that desire was regressed on the motive to see if the ex-partner was with someone new and the motive to compare oneself to the ex-partner’s potential new partners seen on Facebook. The squared multiple correlation was statistically significant, R2 = .08, F (2, 211) = 8.63, p < .001. The motive to determine if there was a new partner predictor was statistically significant b = .39, β = .20, p

= .03, whereas the motive to compare predictor was smaller and not statistically significant b = .17, β = .10, p = .29. The data were consistent with the hypothesis for surveillance (H2a) and the motive to see if the partner was with someone new (H2b), but less so the motive to compare oneself (H2c).

The third set of hypotheses predicted that breakup recovery would be negatively correlated with surveillance of the ex-partner on Facebook and both motives. Again, the correlation matrix showed that all three variables had substantial and statistically significant negative relationships with recovery. Recovery was re- gressed onto the two motive variables. The squared multiple corre- lation was statistically significant, R2 = .23, F (2, 211) = 30.73, p < .001. The motive to determine if there was a new partner predictor was weak and not statistically significant b = -.05, β = -.09, p = .29, whereas the motive to compare predictor was substantial in size and statistically significant b = -.21, β = -.41, p < .001. The data were consistent with the hypothesis for surveillance (H3a) and the motive to compare oneself (H3c) but less so the motive to determine if the partner was with someone new (H3b).

The fourth hypothesis predicted that breakup recovery would have a stronger negative relationship with the extent to which the ex-partners were perceived to be interacting with potential new partners that the participant did not know than those the partici- pant did know. Recovery was regressed onto observed ex-partner interaction with knowns and unknowns. The squared multiple cor- relation was statistically significant, R2 = .06, F (2, 211) = 6.19, p = .002. Interactions with known others was not statistically signifi- cant b = -.03, β = -.05, p = .61, whereas interactions with unknown others was moderate in size and statistically significant b = -.13, β = -.20, p = .02. These results were consistent with H4

Table 2: Correlation Matrix

1 2 3 4 5 6 1. Surveillance 2. Motive: New 0.75* 3. Motive: Compare 0.69* 0.67* 4. Interacting Known 0.38* 0.36* 0.34* 5. Interacting Unknown 0.44* 0.43* 0.35* 0.66* 6. Want to Reunite 0.34* 0.27* 0.23* 0.07 0.14* 7. Recovery -0.48* -0.36* -0.47* -0.18* -0.23* -0.63*

* p < .05

34

SMSociety ’20, July 22–24, 2020, Toronto, ON, Canada Erin Spottswood and Christopher Carpenter

4 DISCUSSION This study set out to develop the observed sender component and test the observed receiver component of the hyperperception model by examining the motivations that affect the likelihood someone would engage in IES of their ex-partner on an SNS like Facebook, as well as what variables predict poor break-up recovery. The model starts from the assumption that people will sometimes be motivated to observe someone on an SNS. But this study expanded the model by empirically examining some potential motives in the ex-partner context. In addition, this study tested the hyperperception model’s unique prediction in the observed receiver component that con- strained access could negatively affect recovery from a break-up. Some of the results were not only consistent with the hypotheses, but also hint at possible hyperperception processes that could occur when one surveils or observes their ex-partner’s Facebook activ- ity. The results that were inconsistent with some of this study’s hypotheses do not necessarily invalidate the likelihood that hyper- perception processes may play a role in IES and breakup recovery, but do suggest further avenues for research. The theoretical and methodological implications readers can take away from the results of this study are discussed next.

4.1 The Importance of Motivations and Potential for Hyperperceptions During IES

The results concerning single people versus people in new relation- ships were consistent with predictions; not only did they replicate previous findings from IES research [28], they also uncover some of the fundamental motivations behind IES phenomenon. In addition, these results also help develop a key component of the hyperper- ception model: that some observers are especially motivated to observe and appraise a particular person’s (i.e., observed sender’s) SNS interactions. In this case, the “still single” observer may be in- terested in determining if their ex-partner (the observed sender) has romantically moved on to a new relationship and also interested in assessing how they (the observer) compare to one or several of the ex-partner’s potential new romantic connections (observed receivers) according to who they see their ex-partner interact with on Facebook.

The results also indicated the desire to reunite with ex-partners was associated with surveillance, in particular to see if the ex is with someone new. These results not only show how understanding the motivations behind IES can help explain why some – but not all – engage in IES after a breakup, they also point towards potential hyperperception processes. Applying the results of this study to the hyperperception model, imagine the participants are the ob- servers and the ex-partner is the sender the observer is primarily interested in observing and appraising their Facebook interactions with other potential romantic connections. It makes sense that wanting to reunite with an ex-partner would increase an observer’s IES of their ex-partner, as well as increase their motivation to see if the ex-partner has a new romantic connection. However, there was no substantial relationship between wanting to reunite with an ex-partner and comparing oneself to an ex-partner’s potential new romantic connections. Perhaps the comparison motive was an insubstantial predictor of a desire to reunite with an ex-partner be- cause some of the people who want to reunite with their ex-partner

try to avoid making such upward comparisons in order to protect their ego. Ignorance may indeed be bliss or at least beneficial when preserving hope for a reconciliation. We suggest additional research to explore that possibility.

We also predicted that relative to people who were struggling post-breakup, people who were recovering well after a breakup would (a) engage in less surveillance of their ex-partner, (b) be less motivated to surveil their ex-partner to determine if their ex-partner is developing a new romantic relationship, and (c) be less motivated to compare themselves to their ex-partner’s Facebook connections. Results show that understanding the motivations behind IES can help explain why some but not all engage in IES after a breakup, and they also point towards potential hyperperception processes. This study found that people who were recovering better from their breakup did less surveilling of their romantic partner on Facebook. This suggests that they were not motivated to surveil and appraise their ex-partner’s interactions with others that seem like new or potential partners that their ex is developing or pursuing. This finding is consistent with the hyperperception model; if an observer felt they had recovered from the breakup, they are likely spending more of their time on Facebook engaging with their own social connections rather than worrying about and spending their time looking to see what their ex-partner is doing on Facebook. Similarly, people who were recovering from their breakup were less likely to be motivated to compare themselves to those they saw interacting with their ex-partner on the site. This makes sense; if an observer has moved on, they will likely not be inclined to compare themselves to an ex-partner’s new romantic connections.

However, breakup recovery did not impact whether or not an observer was motivated to see if their ex-partner was in a new relationship. While this seems inconsistent with the results of the other hypotheses concerning breakup recovery, it makes sense why the extent to which someone had recovered from a breakup would not consistently be associated with wanting to use Facebook to see if their ex-partner had a new romantic connection. For those who had been recovering well after the breakup, they might still be friends with their ex-partner (it is not uncommon for romantic relationships to transition to platonic ones [2]), and as such might want to know and even be happy to learn if their ex-partner was dating someone new. One participant mentioned in an open-ended comments box at the end of the survey that they engaged in surveillance in the hopes that their ex-partner was doing well. On the other hand, an observer who is still grieving the end of a relationship and wanting their ex-partner back (which is also common [10]) might also be inclined to see if their ex-partner is with someone new to see if a reconciliation is possible. There may be several reasons why being more or less recovered from a breakup would not affect an observer’s motivation to see if their ex-partner is with someone new. All in all, break-up recovery impacts the motivations behind IES phenomenon, which sheds new light on this vein of research.

4.2 Constrained Access & Possible Hyperperception Effects

This study not only wanted to expand the hyperperception model by investigating the motivations that are crucial to understanding why an observer would hone in on a specific or observed sender’s

35

Breaking Up is Hard to Do, Especially When Facebook Won’t Let You SMSociety ’20, July 22–24, 2020, Toronto, ON, Canada

SNS activity and interactions, it also wanted to uncover how be- ing limited in their ability to contextualize an observed receiver’s interactions with others besides the observed sender (i.e., the ob- server’s ex-partner) may lead to potential hyperperceptions of the ex-partner’s and unknown person’s relationship. This is because being constrained in their ability to access the observed receiver’s interactions with others besides their ex-partner on Facebook may lead them to make assumptions about their ex-partner’s dating life, and undermine their ability to recover from the breakup. We predicted that when the observer sees their ex-partner interacting with people that they personally do not know offline, they will have a harder time recovering from the break-up than seeing their ex-partner interact with people they do know offline. This builds upon the logic that people tend to not be connected to people they do not know offline on Facebook [31], thus making it difficult for an observer to access an observed receiver’s Facebook interactions with others besides their ex-partner on the site and in face-to-face contexts. This situation would undermine the observer’s ability to contextualize their ex-partner’s interactions with an unknown observed receiver on the site. The results were consistent with this hypothesis and suggest that perhaps the participants who were having a difficult time recovering from the breakup were observing their ex-partner interact with an unknown person on Facebook and their interactions hinted at the possibility that their ex-partner was in or was developing a new romantic relationship with this unknown person.

In addition, hyperperceptions are thought to be more likely to occur if the observed receiver and observed sender are interacting on an SNS endowed with positive and affiliative posting and inter- action norms (e.g., Facebook [27, 30]). Applying this to the present study, imagine the observer trying to contextualize the interactions they see between their ex-partner and an observed receiver they do not know offline, meaning that the observer will likely be con- strained in how much they can access the observed receiver’s public interactions with others on the site. This in turn will make it difficult for them to contextualize their ex-partner’s and observed receiver’s interactions on Facebook. It is possible that being constrained in their ability to access the observed receiver’s Facebook interactions undermined break-up recovery for those who participated in this study because a) the participants (i.e., observers) perceived that the observed receiver was a potential new romantic connection and b) both the ex-partner and the potential new romantic connection were interacting in ways that held up the site’s positive and affilia- tive posting and interaction norms. This would lead the observer to perceive that their ex-partner had already moved on which could be experienced as hurtful and confusing. Future research may want to retrieve the SNS interactions that are exchanged between the ob- served sender and observed receiver to see if they exude positivity and affiliation. It could then be uncovered what types of messages trigger hyperperceptions, which in turn may undermine breakup recovery.

The finding that being constrained in their access to the observed receiver’s Facebook interactions made it harder for an observer to recover from a breakup suggests that perhaps some of this study’s participants were honing in on observed receivers who they per- ceived as their ex-partner’s potential new romantic connections. Constraint is key here: the less an observer knows the observed

receiver, the less sure they can be of the nature of their ex-partner’s relationship with said unknown observed receiver, as well as be hindered in their ability to ascertain if their ex-partner’s interac- tions with this unknown person is indicative of the quality of their relationship. If the observer perceives that the ex-partner’s inter- actions with an unknown observed receiver means they are in or are developing a new romantic relationship with that person, they may question how much their ex-partner ever cared for them or if their ex-partner was ever happy in their relationship. All these thoughts and feelings could continue to undermine their (the ob- server’s) ability to recover from the breakup. However, given that this study did not more precisely measure which observed receivers the observer was focusing on, nor their perception of the level of intimacy expressed in the ex-partner’s and observed receiver(s’) interactions, this is a potential but not definite finding that would suggest hyperperceptions play a detrimental role in the breakup re- covery process. Future research will need to focus on the observer’s perception of their ex-partner’s interactions with potential new romantic connections more specifically to better uncover how con- strained access can lead to hyperperceptions which may undermine breakup recovery.

4.3 Limitations There are a few limitations worth noting. First, the sample was moderate in size and drawn from an undergraduate university population. This study still demonstrates that the results warrant further testing of how hyperperceptions can explain the detrimen- tal effects of IES on breakup recovery but bigger and more diverse samples are needed to determine if and how hyperperceptions affect IES and breakup recovery. Furthermore, this study would have been well served by including some individual and situational variables that may explain why an observer would be motivated to observe if their ex-partner is developing a new romantic rela- tionship (e.g., traumatic or sudden break-up, low self-esteem, etc.). Future research would benefit from including the aforementioned and other variables that may influence and potentially increase IES behaviors of ex-partners, as well as hyperperceptions of that ex-partner’s interactions and relationships with others on SNSs.

4.4 Conclusion Recovering after the end of a romantic relationship is difficult, and SNSs may make it more difficult for some people if they use SNSs to observe and appraise their ex-partner’s interactions with others who they consider to be their ex-partner’s new potential romantic connections. This study set out to uncover the motivations behind IES that in turn undermine breakup recovery, and used the hyperperception model to justify which motivations were explored in this study. It also used the hyperperception model to predict that when an observer is constrained in their ability to contextualize the Facebook interactions between their ex-partner and a potential new romantic connection the observer may have a more difficult time recovering from the breakup. The results suggest uncovering the motivations behind IES helps explain this phenomenon, and also imply that the hyperperception model can be used to pinpoint key individual motivations and situational factors that explain why IES undermines breakup recovery for some but not all SNS users.

36

SMSociety ’20, July 22–24, 2020, Toronto, ON, Canada Erin Spottswood and Christopher Carpenter

REFERENCES [1] Philip Broemer. and Michael Diehl. 2004. Romantic jealousy as a social compari-

son outcome: When similarity stings. Journal of Experimental Social Psychology, 40, 3, 393-400. DOI: https://doi.org/10.1016/j.jesp.2003.08.002

[2] Amy L. Busboom, Dawn M. Collins, Michael D Givertz, and Lauren A. Levin. 2002. Can we still be friends? Resources and barriers to friendship quality after romantic relationship dissolution. Personal Relationships, 9, 2 215-223. DOI: https: //doi.org/10.1111/1475-6811.00014

[3] William R. Cupach, Brian H. Spitzberg, Colleen M. Bolingbroke, and Bobbi S. Tellitocci, 2011. Persistence of attempts to reconcile a terminated romantic rela- tionship: A partial test of relational goal pursuit theory. Communication Reports, 24, 99-115. DOI: 10.1080/08934215.2011.613737

[4] Megan R. Dillow, Christopher R. Morse, and Walid A. Afifi. 2008. Romantic reconciliation as influenced by implicit theories of relationships and information seeking style. Human Communication, 11, 3, 349-358.

[5] Tiffany Field, Miguel Diego, Martha Pelaez, Osvelia Deeds, and Jeannette Delgado. 2009. Breakup distress in university students. Adolescence, 44, 176, 705-727.

[6] Jesse Fox and Jennifer J. Moreland. 2015. The dark side of social networking sites: An exploration of the relational and psychological stressors Associated with Facebook use and affordances. Computers in Human Behavior, 45, 168-176. DOI: https://doi.org/10.1016/j.chb.2014.11.083

[7] Jesse Fox and Robert S. Tokunaga. 2015. Romantic partner monitoring after breakups: Attachment, dependence, distress, and post-dissolution online surveil- lance via social networking sites. Cyberpsychology, Behavior, and Social Network- ing, 18,9, 491-498. DOI: https://doi.org/10.1089/cyber.2015.0123

[8] Bertram Gawronski. 2004. Theory-based bias correction in dispositional inference: The fundamental attribution error is dead, long live the correspondence bias. Euro- pean Review of Social Psychology, 15, 1, 183-217. DOI: 10.1080/10463280440000026

[9] David W. Gerbing. 2014. R data analysis without programming. Routledge, New York, NY.

[10] Sarah Halpern-Meekin, Wendy D. Manning, Peggy C. Giordano, and Monica A. Longmore. 2013. Relationship churning in emerging adulthood: On/off rela- tionships and sex with an ex. Journal of Adolescent Research, 28, 2: 166-188. DOI: 10.1177/0743558412464524

[11] Adam Joinson. 2008. Looking at, looking up, or keeping up with people? Proceed- ings of the twenty-sixth annual CHI Conference on Human Factors in Computing Sys- tems. ACM Press. Boston, MA, USA. DOI: https://doi.org/10.1145/1357054.1357213

[12] Jody K. Kellas, Dawn Bean, Cherakah Cunningham, and Ka Y. Cheng. 2008. The ex-files: Trajectories, turning points, and adjustment in the development of post- dissolutional relationships. Journal of Social and Personal Relationships, 25, 1, 23-50. DOI: https://doi.org/10.1177/0265407507086804

[13] David Knox, Marty E. Zusman, Melissa Kaluzny, and Chris Cooper. 2000. College student recovery from a broken heart. College Student Journal, 34, 3: 322-324.

[14] Pamela J. Lannutti and Kenzie A. Cameron 2002. Beyond the breakup: Heterosex- ual and homosexual post-dissolutional relationships. Communication Quarterly, 50,2, 153-170. DOI: https://doi.org/10.1080/01463370209385654

[15] Leah LeFebvre, Kate Blackburn, and Nicholas Brody. 2015. Navigating romantic relationships on Facebook: Extending the relationship dissolution model to social networking environments. Journal of Social and Personal Relationships, 32, 1, 78-98. DOI: https://doi.org/10.1177/0265407514524848

[16] Veronica Lukacs and Anabel Quan-Haase. 2015. Romantic breakups on Facebook: New scales for studying post-breakup behaviors, digital distress, and surveillance. Information, Communication & Society, 18, 5, 492-508. DOI: https://doi.org/10. 1080/1369118x.2015.1008540

[17] Amy Lyndon, Jennifer Bonds-Raacke, and Alyssa D. Cratty. 2011. College stu- dents’ Facebook stalking of ex-partners. Cyberpsychology Behavior, and Social Networking, 14, 12, 711-716. DOI: https://doi.org/10.1089/cyber.2010.0588

[18] Alice E. Marwick. 2012. The public domain: Social surveillance in everyday life. Surveillance & Society, 9, 4, 378-393. DOI: https://doi.org/10.24908/ss.v9i4.4342

[19] Ami Rokach. 1989. Antecedents of loneliness: A factorial analysis. The Journal of Psychology, 123, 4, 369-384. DOI: https://doi.org/10.1080/00223980.1989.10542992

[20] Davod A. Sbarra and Robert E. Emery. 2005. The emotional sequelae of nonmarital relationship dissolution: Analysis of change and intraindividual variability over time. Personal Relationships, 12, 2, 213-232. DOI: https://doi.org/10.1111/j.1350- 4126.2005.00112.x

[21] Susan Sprecher. 1994. Two sides to the breakup of dating relationships. Personal Relationships, 1, 3: 199- 222. DOI: https://doi.org/10.1111/j.1475-6811.1994.tb00062. x

[22] Susan Sprecher, Diane Felmlee, Sandra Metts, Beverley Fehr, and Debra Venni. 1998. Factors associated with distress following the breakup of a close relationship. Journal of Social and Personal Relationships, 15, 6, 791-809.

[23] Erin L. Spottswood and Christopher J. Carpenter. In press. The hyperpercep- tion model: How observing others on social media can affect people in close relationships. In Rebecca Lind (Ed.) Produsing Theory Vol.3. Peter Lang, New York.

[24] Erin L. Spottswood and Christopher J. Carpenter. 2020. The hyperception model: What happens when you do not know your significant other’s friends offline. International Journal of Interactive Communication Systems and Technologies 10,

1, 1-12 DOI: 10.4018/ijicst.202001010 [25] Erin L. Spottswood and Jeffrey T. Hancock. 2017. Should I share that? Prompting

social norms that influence privacy behaviors on a social networking site. Journal of Computer-Mediated Communication, 22, 2, 55-70. DOI: https://doi.org/10.1177/ 0265407598156005

[26] Kenneth Tan, Christopher R. Agnew, Laura E. VanderDrift, and S. Marie Harvey. 2015. Committed to us: Predicting relationship closeness following nonmarital romantic relationship breakup. Journal of Social and Personal Relationships, 32, 4, 456-471. DOI: https://doi.org/10.1177/0265407514536293

[27] Robert S. Tokunaga. 2011. Social networking site or social surveillance site? Understanding the use of interpersonal electronic surveillance in romantic rela- tionships. Computers in Human Behavior, 27, 2, 705-713. DOI: https://doi.org/10. 1016/j.chb.2010.08.014

[28] Stephanie T. Tong. 2013. Facebook use during relationship termination: Uncer- tainty reduction and surveillance. Cyberpsychology, Behavior, and Social Network- ing, 16, 11, 788-793. DOI: https://doi.org/10.1089/cyber.2012.0549

[29] Joseph B. Walther. 1996. Computer-mediated communication: Impersonal, inter- personal, and hyperpersonal interaction. Communication Research, 23, 1, 3-43. DOI: https://doi.org/10.1177/009365096023001001

[30] Sonja Utz. 2015. The function of self-disclosure on social network sites: Not only intimate, but also positive and entertaining self-disclosures increase the feeling of connection. Computers in Human Behavior, 45, 1- 10. DOI: https://doi.org/10. 1016/j.chb.2014.11.076

[31] Alyson L Young and Anabel Quan-Haase. 2013. Privacy protection strategies on Facebook: The internet privacy paradox revisited. Information, Communication & Society, 16, 4, 479-500. DOI: https://doi.org/10.1080/1369118x.2013.777757

A APPENDIX 1. ONLINE SURVEY Electronic surveillance of their ex-partner

Fox & Tokunaga (2015) post-breakup interpersonal electronic surveillance

[modified to be in present tense] 7-point Likert 1. When visiting my ex-partner’s Facebook page, I read the

new posts of his/her friends. 2. I often spend time looking through my ex-partner’s Facebook

pictures. 3. I pay particularly close attention to news feeds and posts

that mention my ex-partner. 4. I notice when my ex-partner updates his/her Facebook page. 5. If there are messages on my ex-partner’s wall or pictures I

don’t understand, I try to investigate them. 6. I am generally aware of my ex-partner’s Facebook activities. 7. I explore my ex-partner’s Facebook page to see if there is

anything new or exciting Motive to surveil their ex-partner to determine if their ex-partner is developing a new romantic relationshipTong (2013) Detecting new romantic partner subscale of Facebook information-seeking foci [4th item extended to focus on new exes] 7-point Likert response

1. I look at my ex’s relationship status on Facebook to see if she/he has begun a new relationship with anyone.

2. Facebook can reveal if my ex has a new romantic interest. 3. I look through my ex’s Facebook photos to see if she/he has

a new romantic partner. 4. Using Facebook lets me keep track of my ex’s online com-

munications with people my ex might be interested in Motivation to compare themselves to their ex-partner’s FB friends of the same sex as themselves

1. When I see my ex interacting with someone of the same sex as me on FB, I look at the profile and compare myself to that person.

37

Breaking Up is Hard to Do, Especially When Facebook Won’t Let You SMSociety ’20, July 22–24, 2020, Toronto, ON, Canada

2. It is important for me to know if I am as attractive as the other people of the same sex as I am that my ex knows on Facebook.

3. I sometimes look at what other people who are the same sex as I am are saying to my ex on Facebook so I can see if they are as interesting as I am.

4. I use Facebook to decide if my ex has Facebook friends who are more attractive than I am.

Please respond to the following 2 questions using this scale: • Never • Occasionally • Sometimes • Regularly • Often • Constantly

How often do you see your ex interact with non-family members on Facebook who are of the same sex as you are, that you have personally met offline?

How often do you see your ex interact with non-family members on Facebook who are of the same sex as you are, that you have NOT personally met offline?

Subject relationship status: What is your romantic relationship status? • Single • Romantically involved with my ex • Romantically involved with someone else

Partner relationship status: What is your ex’s romantic relationship status? • Single • Romantically involved with me • Romantically involved with someone else • I don’t know

How long were you together before your broke up? • Days • Weeks • Months • Years

How long has it been since your relationship with your ex ended? • Days • Weeks • Months • Years

Who initiated the breakup? Fox & Tokunaga (2015) • 1- I did • 2- Mostly me

• 3- We both did • 4- Mostly by partner • 5- My partner did

Desire to reunite with their ex Tan, Agnew, VanderDrift, & Harvey 2015 ‘On a scale from 0 to 10, how much would you like to reunite

with your ex in a romantic relationship at any point in the future?” (0 = absolutely not and 10 = definitely).

Breakup recovery Sbarra & Emery (2005) Acceptance of Relationship Termination

adaptation of Kitson (1982) Acceptance of Martial Termination scale with some wording adjustments to focus on the breakup with ex-partner

For each statement, would you please pick the number which best reflects your feelings.

1, ’not at all my feelings’ to 5, ’very much my feelings.’ 1. Since we brokeup, everything I do seems like an effort. 2. I find myself spending a lot of time thinking about my ex-

partner. 3. I’m feeling like myself again. 4. Sometimes I can’t believe that we brokeup. 5. I find myself wondering what my ex-partner is doing. 6. Since we brokeup, I have no interest in anything. 7. I’m angry at my ex-partner. 8. I do not feel any guilt about the breakup. 9. I feel I will never get over the breakup. How often do you use Facebook (including on a com-

puter/smartphone/tablet/ etc.? • Never • Rarely • Occasionally • Sometimes • Often • Constantly

Age What is your age? Please only type a two digit number. Sex For the purposes of this study, how would you like to be identi-

fied? • Female • Male • Trans* non-binary, two-spirit, genderqueer, other • Prefer to not answer

Please leave any comments you wish to make here: • [open-ended]

38