Elsevier

Computers in Human Behavior

Volume 33, April 2014, Pages 171-178
Computers in Human Behavior

Being unfriended on Facebook: An application of Expectancy Violation Theory

https://doi.org/10.1016/j.chb.2014.01.029Get rights and content

Highlights

  • Expectancy Violation Theory (EVT) is applied to Facebook unfriending.

  • Online surveys examined over 500 adults who were unfriended.

  • Being unfriended was a moderately expected and negative violation.

  • Being unfriended was a moderately-to-highly important expectancy violation.

  • The findings extend EVT to being unfriended on Facebook.

Abstract

This study considered being unfriended on Facebook as an expectancy violation that could vary in valence, importance, and expectedness according to a number of relationship and Facebook involvement characteristics. Facebook users who had been unfriended responded to a variety of quantitative scales via an online survey. Being unfriended constituted a moderately expected and negative, and moderately-to-highly important, expectancy violation. Whether ties with the unfriender were close versus weak best predicted valence and importance and the extent to which the unfriended individual used Facebook to connect with existing contacts best explained violation expectedness. Violation importance also predicted whether or not the unfriended individual contacted the former friend about being unfriended. Results supported Expectancy Violation Theory and extended knowledge about Facebook unfriending.

Introduction

The advent of social networking websites (SNSs), as well as other technologically mediated communication channels, such as texting and video chatting, has fundamentally reshaped how individuals relate to one another. For example, recent research (Caughlin & Sharabi, 2013) suggests that mediated and face-to-face communication channels are now being integrated by partners as their relationships grow, and the easier this integration is, the closer and more satisfying the relationship. In addition, text messaging, a communication channel with minimal nonverbal expression and the potential for misunderstanding, was most strongly associated with romantic partners’ positive and negative communication (Coyne, Stockdale, Busby, Iverson, & Grant, 2011). Further, the notion that relationships that are traditionally viewed as close in face-to-face contexts, such with parent–child, may be managed differently (or avoided entirely) via SNSs, is being challenged. For example, counter to expectations, being “friended” on Facebook by one’s mother or boss was met with a relatively positive response by college students (Karl & Peluchette, 2011). Indeed, Kanter, Afifi, and Robbins (2012) found that the introduction of college students’ parents on Facebook actually decreased conflict between parent and child over a two-month period.

One specific SNS-specific act that has changed the ways that individuals relate is through the act of unfriending. Unfriending (or defriending) is an option available on the Facebook social network, which is the largest SNS worldwide with 1.19 billion active monthly users as of September 2013 (Facebook Key Facts, 2013). Individuals can choose to unfriend a Facebook user with whom they had previously agreed to be friends by clicking the “unfriend” button on that individual’s Facebook page. As such, unfriending a Facebook friend is a form of purposeful online relational termination (Bevan, Pfyl, & Barclay, 2012) that is unique from similar offline behaviors in two ways. First, Facebook unfriending offers “a definite marker when the friendship is dissolved by one member” (Sibona & Walczak, 2011, p. 1). Second, the act of unfriending is unilateral in that the unfriender is consciously aware of the act of unfriending, but the individual who is unfriended never receives a formal notification of its occurrence.

The nature of Facebook relationships means that, once individuals initially accept a friend request or have their request accepted by another Facebook user, they likely expect that the resulting Facebook friendship will continue for an extended period of time. However, 63% of Facebook users had unfriended a friend in 2011, a 7% increase since 2009 (Madden, 2012). Further, many individuals were unfriended or blocked in response to their attempt to form an interpersonal relationship: 27% of SNS users reported that they unfriended or blocked someone whose flirting made them uncomfortable (Smith & Duggan, 2013). These statistics, along with the above distinct Facebook unfriending characteristics, indicates that being unfriended on Facebook could be a relational event that the unfriended individual perceives as an expectancy violation. If so, different relationship and Facebook involvement characteristics may be related to how being unfriended is interpreted as an expectancy violation. As such, this study considers a number of correlates of being unfriended on Facebook through the lens of Expectancy Violation Theory (Burgoon and Hale, 1988, Burgoon and Jones, 1976).

Expectancy Violation Theory (EVT) was first proposed by Burgoon and Jones (1976) as a means for explaining how individuals perceive and interpret violations of their personal space. An expectation is a consistent pattern of predictable behavior that can be specific to an individual, context, and/or relationship (Burgoon, 1993). According to EVT, humans employ these expectations to characterize and frame their interactions with others, as well as how they perceive this interaction, process information, and subsequently behave (Burgoon, 1993, Burgoon and Hale, 1988). These expectations are violated when someone acts in a manner that deviates from a behavior that is typical or anticipated (Afifi & Metts, 1998). EVT specifies that, when expectancy violations occur, greater attention is first given to that aspect of the interaction; individuals then process and cope with these expectancy violations via an interpretation and evaluation sequence. The evaluation stage allows the individual to better understand the transgressor, the violation behavior, and how to behaviorally proceed (Bachman & Guerrero, 2006). Since its inception, the scope and utility of EVT has been successfully expanded to include both verbal and nonverbal messages, as well as a variety of relationship contexts such as romantic relationships and cross-sex friendships (e.g., Afifi and Metts, 1998, Bevan, 2003) and teacher–student relationships (Lannutti, Laliker, & Hale, 2001).

Afifi and Metts (1998) expanded the study of expectancy violations in two additional and important ways. First, they identified and categorized the relational behaviors – including ones that were both positive and negative, and major and minor in magnitude – that individuals could view as violations. Nine such expectancy violations were classified, including relationship de-escalation, which involves “behaviors that signal a lack of confidence in the viability of the relationship, cause relational trauma or lead directly to the termination of the relationship” (Afifi & Metts, 1998, p. 377). In a similar vein, Bachman and Guerrero (2006) argued that hurtful events, which are “words or actions that constitute a relational transgression by communicating a devaluation of the partner or the relationship” (p. 945) constitute a relational expectancy violation that includes a distinct disassociation through breakup category. Based upon these definitions, Facebook unfriending would represent a mediated form of the relationship de-escalation or disassociation through breakup hurtful event expectancy violations.

The second way in which Afifi and Metts (1998) expanded EVT research is by measuring multiple aspects by which expectancy violations could be interpreted by those who experience the violation, three of which are particularly relevant to this investigation: (1) valence, which was specified in the original EVT theory and examines the extent to which the violation is interpreted as positive (i.e., exceeds expectations) versus negative (i.e., falls short of expectations; Bachman & Guerrero, 2006); (2) expectedness, which assesses the degree to which the violation deviates from the range of possible anticipated behaviors; and (3) importance, which focuses on how much of an impact the violation will have on the relationship.

Afifi and Metts (1998) found that relationship de-escalation was the second most relationally important expectancy violation, but was one of the three most expected violations across their three studies. Further, in two of their three studies, this expectancy violation was significantly more negative than the scale midpoint. Bachman and Guerrero (2006) similarly found that the disassociation through breakup hurtful event was viewed as a more negative expectancy violation than seven of the 10 hurtful event categories they examined. These scholars also determined that negative valence was the primary variable to distinguish between dating partners who stayed together or not as a result of the hurtful event. Thus, being unfriended on Facebook is likely to be a mediated expectancy violation that could vary in terms of how negatively-valenced, expected, and important it is.

Expectancy Violation Theory has been expanded to computer-mediated communication (CMC) contexts to understand modality switching and response latency as potential expectancy violations. For example, Ramirez and Wang (2008) found that information shared during modality switching between online and face-to-face (FtF) interactions was viewed as an expectancy violation. Further, Jin (2012) found that viewing a partner’s avatar behavior as an expectancy violation mediated the association between self-disclosure and trusting the virtual other when switching modalities. Further, hypothetical situations where high reward individuals who responded to an email two weeks later or never were significantly less expected than one-day latency responses. However, response latency did not differ in expectancy levels for low reward individuals (Kalman & Rafaeli, 2010).

The CMC study that is most informative to the current consideration of Facebook unfriending as an expectancy violation is by McLaughlin and Vitak (2011). They identified a number of norm and expectancy violations specific to the Facebook context. Negative Facebook violations included posting too many status updates, being too emotional in status updates and wall posts, engaging in conflicts, heated interactions, or name-calling, and posting/tagging inappropriate photos of the individual. These negative norm violations are consistent with the online reasons for unfriending (i.e., posting too frequently about everyday events, posting unimportant information, being polarizing, and making inappropriate comments) identified by Sibona and Walczak (2011), suggesting a logical association between expectancy violations and specific online behaviors such as being unfriended. Further, unfriending the transgressor was one way that Facebook users responded to another’s Facebook norm or expectancy violation (McLaughlin & Vitak, 2011).

Facebook unfriending is a distinct form of relationship termination that occurs only in an online context (Bevan et al., 2012), and is an increasingly common option that is exercised by Facebook users (Madden, 2012). Clicking the unfriend button can be an impulsive decision or the result of a great deal of thought for a Facebook user, and it is likely that one consideration when deciding to unfriend someone, even in a fleeting moment of impulse, is how they would respond to being unfriended if they discover that they are no longer friends with Facebook user. Previous EVT research suggests that being unfriended is akin to de-escalating a relationship or disassociating from someone; however, the unique aspects of being unfriended (i.e., there is a definitive marker, and it is a unilateral act) means that we must explore the extent to which being unfriended on Facebook constitutes an expectancy violation. Thus, we investigate how individuals who are unfriended on Facebook would characterize such a relational event according to Afifi and Metts’s (1998) three expectancy violation aspects in RQ1: To what extent is being unfriended on Facebook viewed as an expectancy violation according to (a) valence, (b) expectedness, and (c) importance?

Facebook users can be “friends” with a broad spectrum of individuals, ranging from relatively weak ties such as those they have never met face-to-face, to friends of friends, acquaintances, and more distal family members such as second cousins, to the very closest ties we share, such as romantic partners, siblings, parents, and close friends. However, on Facebook, unless specific privacy or viewing settings are adjusted for individual friends, all of a user’s Facebook friends will see information that is posted by that user, thus “‘levelling’ the online relationship playing field” regarding how individuals of different levels of closeness can access information and interact with one another on Facebook (Bevan et al., 2012, p. 1462). Perhaps reflecting the unique nature of Facebook interactions, Bevan et al. (2012) found that being unfriended by close relationship partners was met with greater rumination by unfriended individuals than when distant partners unfriended them, but that levels of negative emotional response did not differ according to relationship closeness type.

From the EVT perspective, Bevan (2003) observed variations in the extent to which sexual resistance was viewed as a more negative, unexpected, and important expectancy violation according to whether the resistor was a dating partner or a cross-sex friend. From these findings, Bevan (2003) proposed relationship type as a relationship characteristic that can influence the perception, interpretation, and evaluation of expectancy violations. In a similar vein, McLaughlin and Vitak (2011) determined that the basic closeness level of Facebook friends can be associated with how mildly versus strongly a Facebook expectancy violation is perceived to be, as well as how the Facebook user would respond to the violation. Namely, more neutrally or negatively valenced Facebook friends (i.e., weak ties) who committed a negative violation were more likely to be unfriended than positively valenced friends (i.e., close ties). Further, Cohen (2010) found that individuals would be less tolerant of expectancy violations committed by media figures than those enacted by friends.

These findings suggest that the type of relationship in which a violation occurs is used by individuals as an important framing device for perceiving and interpreting the expectancy violation. Thus, based on these EVT findings, we predict in H1 that this will also be the case when an individual is unfriended on Facebook: Those who are unfriended by an individual with whom they have close ties (defined as current or former relationship types that are familial, romantic, and friend in nature) will view this event as a more (a) negative, (b) unexpected, and (c) important expectancy violation than those who are unfriended by an individual with whom they have weak ties (i.e., co-worker, acquaintance, friend of a friend, and/or someone only known online).

The second relationship characteristic that could be relevant to how unfriending as an expectancy violation is interpreted is the length of the Facebook friendship. It is likely that the longer the pair has been Facebook friends, the more they have interacted online – and possibly offline – and the more they have viewed one another’s Facebook behaviors and learned what to expect of that person. Indeed, relationship length was a direct, negative predictor of Facebook monitoring behaviors, suggesting that individuals use Facebook to learn more about their friends as a means of developing relationships (Darvell, Walsh, & White, 2011). Further, online romantic relationship length was positively associated with attitude similarity, intimacy, trust, and perceived level of certainty about the relationship (Anderson & Emmers-Sommer, 2006).

Based on these findings, online relationship length could also then be a relationship characteristic that assists in interpreting being unfriended on Facebook as an expectancy violation. As no known research has examined whether relationship length is associated with how an expectancy violation is perceived and interpreted, examining this potential association could inform EVT. Following this logic, H2 states: The longer the individuals are Facebook friends, the more being unfriended will be viewed as a (a) negative, (b) unexpected, and (c) important expectancy violation.

Two forms of how involved individuals are with Facebook are considered here in relation to the three aspects of expectancy violations. Vitak, Ellison, and Steinfeld (2011) define the first, Facebook intensity, as the level of users’ psychological investment in Facebook. Indeed, the more intense Facebook users were, the more they responded to being unfriended by ruminating and experiencing negative emotion (Bevan et al., 2012). As such, H3 states: The more individuals are intense Facebook users, the more being unfriended will be viewed as a (a) negative, (b) unexpected, and (c) important expectancy violation.

The second form of Facebook involvement examined here is use of Facebook to maintain existing social connections, which reflects that fact that most individuals primarily use Facebook to preserve their existing associations rather than for initiating new ones (e.g., Ellison et al., 2007, Foon Hew, 2011, Lewis and West, 2009). These individuals focus more on interacting with their Facebook friends and are likely interested in “keeping the door open” for similar interactions in the future. Bevan et al. (2012) found that the more unfriended individuals used Facebook to maintain their existing connections, the more they ruminated about being unfriended, though there was not a significant relationship with the experience of negative emotion. Thus, according to H4: The more individuals use Facebook to connect with existing social contacts, the more being unfriended will be viewed as a (a) negative, (b) unexpected, and (c) important expectancy violation.

To understand which of these four relationship and Facebook characteristics are the best predictors of the unfriending expectancy violation, they are examined in RQ2 in relation to the three expectancy violation aspects: Which relationship and Facebook usage characteristics are the best predictors of expectancy violation (a) valence, (b) expectedness, and (c) importance in response to being unfriended?

A final consideration regards the EVT evaluation process about being unfriended on Facebook: whether individuals respond to being unfriended by contacting their former Facebook friends about the unfriending event. Such a response would be consistent with McLaughlin and Vitak (2011), who found that confrontation was one response to a Facebook norm or expectancy violation. In addition, two communicative responses to dating partners’ hurtful event expectancy violations involved direct communication with the partner: integrative and distributive communication (Bachman & Guerrero, 2006). Further, the more negative the violation, the more likely the distributive response was employed (Bachman & Guerrero, 2006). H5 similarly predicts: Individuals who view being unfriended as a more (a) negative, (b) unexpected, and (c) important expectancy violation will report subsequently contacting the individual about being unfriended.

Section snippets

Participants and general procedures

Adults (i.e., age 18 or older) who were aware that they had been unfriended by at least one Facebook friend (N = 547) participated in an online questionnaire that originated from researchers at a small, private university in the western U.S. The majority of participants were female (n = 364, male n = 145) and white/Caucasian (n = 390, Asian n = 44, Hispanic/Latino n = 21, other n = 18, Black/African American n = 10, American Indian or Alaska native n = 6, native Hawaiian or Pacific Islander n = 11). Respondents’

Preliminary analyses

Madden (2012) determined that younger individuals and females were more likely to unfriend on Facebook and Cohen (2010) found that females anticipated less closeness than males in response to experiencing expectancy violations. Further, it is likely that having more Facebook friends may decrease the likelihood of viewing being unfriended as an expectancy violation. Thus, age and number of Facebook friends were considered via bivariate, one-tailed correlation analyses as potential covariates in

Discussion

This study, one of only a handful that has empirically examined the act of Facebook unfriending, employed Expectancy Violation Theory to understand being unfriended as an expectancy violation. Results indicate that being unfriended is viewed by the unfriended individual as a particularly relationally important mediated expectancy violation. Further, whether the ties with the former friend were close or weak, Facebook friendship length, Facebook intensity, using Facebook to maintain social

Acknowledgement

The authors acknowledge the contributions of Brett Barclay, Lauren Hochleutner, Sarah Hudani, Alexandra Kiesselbach, Cristina Kutzbach, Shea Ledbetter, Jacqueline Lynch, Michelle Kermani, William Miller, Rafe Olson, Jeanette Pfyl, Carson Simms, and Victoria Young to this project.

References (30)

  • J.L. Bevan et al.

    Negative emotional and cognitive responses to being unfriended on Facebook: An exploratory study

    Computers in Human Behavior

    (2012)
  • S.A. Jin

    The virtual malleable self and the virtual identity discrepancy model: Investigative frameworks for virtual possible selves and others in avatar-based identity construction and social interaction

    Computers in Human Behavior

    (2012)
  • W.A. Afifi et al.

    Characteristics and consequences of expectation violations in close relationships

    Journal of Social and Personal Relationships

    (1998)
  • T.L. Anderson et al.

    Predictors of relationship satisfaction in online romantic relationships

    Communication Studies

    (2006)
  • G.F. Bachman et al.

    Relational quality and communicative responses following hurtful events in dating relationships: An expectancy violations analysis

    Journal of Social and Personal Relationships

    (2006)
  • J.L. Bevan

    Expectancy Violation Theory and sexual resistance in close, cross-sex relationships

    Communication Monographs

    (2003)
  • J.K. Burgoon

    Interpersonal expectations, expectation violations, and emotional communication

    Journal of Language and Social Psychology

    (1993)
  • J.K. Burgoon et al.

    Nonverbal expectancy violations: Model elaboration and application to immediacy behaviors

    Communication Monographs

    (1988)
  • J.K. Burgoon et al.

    Toward a theory of personal space expectations and their violations

    Human Communication Research

    (1976)
  • J.P. Caughlin et al.

    A communicative interdependence perspective of close relationships: The connections between mediated and unmediated interactions matter

    Journal of Communication

    (2013)
  • E.L. Cohen

    Expectancy violations in relationships with friends and media figures

    Communication Research Reports

    (2010)
  • S.M. Coyne et al.

    “I luv u:)!”: A descriptive study of the media use of individuals in romantic relationships

    Family Relations

    (2011)
  • M.J. Darvell et al.

    Facebook tells me so: Applying the theory of planned behavior to understanding partner-monitoring behavior on Facebook

    Cyberpsychology, Behavior, and Social Networking

    (2011)
  • N.B. Ellison et al.

    The benefits of Facebook “friends”: Social capital and college students’ use of online social network sites

    Journal of Computer-Mediated Communication

    (2007)
  • Facebook Key Facts (2013). <http://newsroom.fb.com/Key-Facts>. Retrieved...
  • Cited by (0)

    An earlier version of this paper was presented at the 2012 biennial meeting of the International Association for Relationship Research in Chicago, IL.

    View full text