Can You Tell That I'm in a Relationship? Attachment and Relationship Visibility on Facebook

Abstract

People often attempt to shape others' perceptions of them, but the role of romantic relationships in this process is unknown. The present set of studies investigates relationship visibility, the centrality of relationships in the self-images that people convey to others. We propose that attachment underlies relationship visibility and test this hypothesis across three studies in the context of Facebook. Avoidant individuals showed low desire for relationship visibility, whereas anxious individuals reported high desired visibility (Studies 1 and 2); however, similar motives drove both groups' actual relationship visibility (Study 1). Moreover, both avoidant individuals and their partners were less likely to make their relationships visible (Studies 1 and 3). On a daily basis, when people felt more insecure about their partner's feelings, they tended to make their relationships visible (Study 3). These studies highlight the role of relationships in how people portray themselves to others.

Full text

http://psp.sagepub.com/
Bulletin
Personality and Social Psychology
http://psp.sagepub.com/content/early/2014/09/15/0146167214549944
The online version of this article can be found at:
DOI: 10.1177/0146167214549944
published online 17 September 2014Pers Soc Psychol Bull
Lydia F. Emery, Amy Muise, Emily L. Dix and Benjamin Le
Can You Tell That I'm in a Relationship? Attachment and Relationship Visibility on Facebook
Published by:
http://www.sagepublications.com
On behalf of:
Society for Personality and Social Psychology
can be found at:Personality and Social Psychology BulletinAdditional services and information for
http://psp.sagepub.com/cgi/alertsEmail Alerts:
http://psp.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
What is This?
- Sep 17, 2014OnlineFirst Version of Record >>
by guest on September 18, 2014psp.sagepub.comDownloaded from by guest on September 18, 2014psp.sagepub.comDownloaded from

Personality and Social
Psychology Bulletin
1 –14
© 2014 by the Society for Personality
and Social Psychology, Inc
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0146167214549944
pspb.sagepub.com
Article
Imagine two people in romantic relationships. From the
moment his relationship begins, Liam tells all his friends and
acquaintances about it, frequently bringing up his partner in
conversation. Conversely, Olivia has been in a relationship
for several months but rarely mentions her partner; she does
not want her relationship to be part of how other people see
her. What explains these divergent behaviors?
People are motivated to portray particular self-images to
others (Leary & Kowalski, 1990); yet, despite the centrality
of romantic relationships in most people’s lives, little is
known about when and how people incorporate their rela-
tionships into those self-images. As illustrated above, people
may differ radically in their desire to make their relationship
known to others, but the psychological antecedents of want-
ing one’s relationship to be visible are unstudied.
The present research investigates relationship visibility—
the centrality of relationships in the self-images conveyed to
others. We propose that relationship visibility stems from
attachment dimensions (Hazan & Shaver, 1987) and investi-
gate this process through an impression management frame-
work (Leary & Kowalski, 1990). Across three studies, we
tested the hypothesis that avoidant attachment predicts lower
desired and actual visibility, whereas anxious attachment
predicts greater desired visibility. We also expected that
impression management motivations, such as enhancing
self-esteem, gaining social status, and creating a specific
identity, might account for the association between attach-
ment and visibility.
Impression Management
People may be concerned with how others perceive them and
exert effort to sculpt these perceptions. Impression manage-
ment theory (Leary & Kowalski, 1990) posits that people
manage others’ impressions of them to fulfill different goals,
which include attaining social outcomes (e.g., enhancing sta-
tus or acceptance by others), increasing their self-esteem, or
portraying a specific identity (e.g., people’s clothing in the
workplace influences others’ perceptions of their skills and
warmth; Karl, Hall, & Peluchette, 2013). These identities
that people present can include portraying the current
549944PSPXXX10.1177/0146167214549944Personality and Social Psychology BulletinEmery et al.
research-article2014
1Haverford College, Haverford, PA, USA
2Northwestern University, Evanston, IL, USA
3University of Toronto Mississauga, Mississauga, Ontario, Canada
4University of Wisconsin–Madison, Madison, WI, USA
Corresponding Author:
Lydia F. Emery, Department of Psychology, Northwestern University,
Swift Hall 409, 2029 N. Sheridan Rd., Evanston, IL 60208, USA.
Email: lemery@u.northwestern.edu
Can You Tell That I’m in a Relationship?
Attachment and Relationship Visibility
on Facebook
Lydia F. Emery1,2, Amy Muise3, Emily L. Dix1,4,
and Benjamin Le1
Abstract
People often attempt to shape others’ perceptions of them, but the role of romantic relationships in this process is unknown.
The present set of studies investigates relationship visibility, the centrality of relationships in the self-images that people convey
to others. We propose that attachment underlies relationship visibility and test this hypothesis across three studies in the
context of Facebook. Avoidant individuals showed low desire for relationship visibility, whereas anxious individuals reported
high desired visibility (Studies 1 and 2); however, similar motives drove both groups’ actual relationship visibility (Study 1).
Moreover, both avoidant individuals and their partners were less likely to make their relationships visible (Studies 1 and 3).
On a daily basis, when people felt more insecure about their partner’s feelings, they tended to make their relationships visible
(Study 3). These studies highlight the role of relationships in how people portray themselves to others.
Keywords
attachment, close relationships, impression management, Facebook
Received July 6, 2013; revision accepted July 31, 2014
by guest on September 18, 2014psp.sagepub.comDownloaded from

2 Personality and Social Psychology Bulletin
self-concept, conveying a desired identity, or presenting the
self to avoid an undesired identity (Leary & Kowalski, 1990).
People in romantic relationships include aspects of their
partners in their self-concepts (Aron, Aron, Tudor, & Nelson,
1991), suggesting that relationships should be part of how
the self is presented to others. In established relationships,
people often attempt to manage others’ impressions of their
relationships and obscure negative aspects from others
(Loving & Agnew, 2001). Couples can feel pressured to
reveal the status of their relationships, because they believe it
is a social expectation, but may conceal their status when
they believe it will be perceived negatively or that telling
others will be detrimental to the relationship (Baxter &
Widenmann, 1993). Although concealing a relationship can
be exciting in the short term (Wegner, Lane, & Dimitri,
1994), it may corrode long-term relationship quality (Foster
& Campbell, 2005). Despite past research on keeping rela-
tionships secret, as well as a rich literature on impression
management, past research has not considered the centrality
of relationships in the self-images that people construct or
individual differences in the desire to make a relationship
visible. Specifically, we examine who is more likely to desire
relationship visibility, and why some people want others to
know about their relationships whereas others do not. Given
that impression motivation is especially augmented in the
public situations (House, 1980), we investigated this ques-
tion in the context of Facebook.
An immensely popular social networking website (with
more than 1 billion users; Facebook Newsroom, 2014),
Facebook allows users to construct their own profile pages,
which accurately reflect the self-concept (Back et al., 2010).
Self-presentation is a central motivation behind Facebook
use (Nadkarni & Hofmann, 2012), and Facebook enables
relationship visibility through profile pictures, posted rela-
tionship statuses (e.g., “in a relationship with . . .”), and men-
tioning one’s partner in status updates. Previous research has
linked these relational displays to relationship quality
(Carpenter & Spottswood, 2013; Papp, Danielewicz, &
Cayemberg, 2012; Saslow, Muise, Impett, & Dubin, 2013).
Although displaying a relationship on Facebook falls under
our definition of relationship visibility, previous research has
not articulated a theoretically grounded analysis of relation-
ship visibility. Moreover, past research has almost exclusively
investigated how visibility is associated with relationship
quality. In the present research, we examine individual differ-
ences as correlates of visibility; specifically, we focus on
attachment dimensions, and through our investigation of rela-
tionship visibility, we link disparate fields of research on
attachment theory and impression management theory.
Adult Attachment
Attachment shapes thoughts, emotions, and behavior in
romantic relationships (Collins & Allard, 2001), with indi-
viduals differing along two continuous dimensions of anxiety
and avoidance. Anxiety reflects the valence of self-views;
those high in anxiety view themselves as unlovable and fear
abandonment. Conversely, avoidance describes one’s views
of others; individuals high in avoidance dislike closeness and
distrust others (Bartholomew & Horowitz, 1991; Collins &
Allard, 2001). Low scores on both anxiety and avoidance
reflect attachment security, whereas high scores on either
dimension indicate insecure attachment. An extensive body
of research has shown that anxiety and avoidance predict
experiences in relationships, ranging from support seeking
(Simpson, Rholes, & Nelligan, 1992) to partners separating at
the airport (Fraley & Shaver, 1998), to relationship quality
(Etcheverry, Le, Wu, & Wei, 2013).
Attachment also influences how individuals’ partners
experience the relationship; those paired with insecure indi-
viduals tend to exhibit behavior consistent with the insecure
individual’s working models (Kirkpatrick & Davis, 1994). In
other words, insecurely attached individuals’ behavior
appears to produce the rejection that they expect. Partners of
anxious individuals are less committed and satisfied, becom-
ing emotionally detached in stressful situations. In turn, anx-
ious individuals tend to be strongly influenced by their
romantic partners (Slotter & Gardner, 2012). Similarly, the
partners of avoidant individuals are more insecure, less trust-
ing, and behave more negatively (Butzer & Campbell, 2008;
Campbell, Simpson, Kashy, & Rholes, 2001; Simpson, 1990).
Attachment describes trait-like interpersonal insecurity,
but it is also related to state-like insecurity. For instance,
those who are chronically insecure (including anxiously
attached individuals) feel less insecure about their relation-
ships when their partners overemphasize their caring for the
individual (e.g., telling their partners that they feel more
positively about them than they really do; Lemay & Dudley,
2011). Likewise, daily insecurity about the relationship may
also be related to relationship visibility, especially for indi-
viduals who are anxiously attached.
Attachment and Relationship Visibility
Although attachment has not been specifically applied to
impression management, previous research suggests a link.
People attempt to manage others’ impressions of them to
decrease a disparity between their current images and their
desired self-images (Baumeister & Jones, 1978; Leary &
Kowalski, 1990). Given that insecure individuals report
discrepancies between their actual and ideal selves
(Mikulincer, 1995), they should be especially likely to
manage others’ impressions. Anxious individuals yearn to
be accepted and loved by their partners (Collins & Allard,
2001), so we expect they will want to make their relation-
ships central to their public images. Conversely, avoidant
individuals likely do not want their relationships to appear
central to their self-concept, due to their dislike of close-
ness and desire for independence (Collins & Allard, 2001;
Mashek & Sherman, 2004).
by guest on September 18, 2014psp.sagepub.comDownloaded from

Emery et al. 3
Moreover, anxious and avoidant individuals may have
different goals tied to impression management, which may
explain their discrepancies in desired relationship visibility.
People impression-manage in pursuit of social or material
outcomes, self-esteem, or a certain self-image (Leary &
Kowalski, 1990). For social or material outcomes, we
expected that anxious individuals might believe that their
relationships bring them higher social status than do avoid-
ant individuals. In general, people are motivated to achieve
social approval for their relationships (e.g., Leslie, Huston,
& Johnson, 1986). We are not suggesting that avoidant indi-
viduals do not desire relationships, but given that anxious
individuals seek high closeness in their relationships, they
might be more likely to view displaying a relationship as
socially desirable. In terms of self-esteem goals, anxious
individuals tend to report negative self-views (Collins &
Read, 1990), so relationship visibility may be one means of
restoring their self-worth. In terms of self-image goals,
avoidant individuals may be motivated to make their rela-
tionships less visible to be perceived as independent from
their partners, given that they value independence (Collins &
Allard, 2001). However, anxious people likely do not desire
perceptions of independence. Self-image goals may also
encompass relationship quality. Both anxious and avoidant
individuals have poor relationship quality (Etcheverry et al.,
2013), so their beliefs about other people’s perceptions of
their relationship quality may drive their decisions to make
their relationships more or less visible to others.
The Current Studies
Romantic relationships shape the self-concept (e.g., Aron et
al., 1991), so considering the role of relationships in self-
presentation is essential to understanding impression man-
agement (Leary & Kowalski, 1990). The present research
aimed to determine whether attachment (Hazan & Shaver,
1987) underlies the decision to make a relationship more,
versus less, visible. Specifically, we expected that anxious
individuals would desire to make their relationships a central
part of their self-image on Facebook, whereas avoidant indi-
viduals would not (Hypothesis 1a). We expected this pattern
of effects would also emerge for actual relationship visibility
(Hypothesis 1b), particularly for avoidant individuals.
However, anxious individuals’ desired visibility might not
translate into reality, given that their partners influence them
(Slotter & Gardner, 2012). We also anticipated that individu-
als’ attachment avoidance would predict their partners’ actual
relationship visibility (Hypothesis 2), but that attachment
anxiety would not.
For Hypotheses 1 and 2, we tested relatively permanent
forms of relationship visibility (a dyadic profile picture and a
dyadic relationship status). We were also interested, how-
ever, in more transient forms of relationship visibility, and
we anticipated that on days when people felt more insecure
about their partner’s feelings for them, they would be more
likely to post about the relationship or their partner
(Hypothesis 3a). However, we predicted that this effect
would be moderated by attachment, such that anxious indi-
viduals experiencing insecurity about their partner’s feelings
would be particularly likely to post about their relationships
on Facebook (Hypothesis 3b).
We also examined the self-presentational goals underly-
ing desired visibility and actual visibility. We adopted an
exploratory approach to these analyses and did not advance
specific hypotheses, given the lack of previous research on
attachment and impression management. However, we gen-
erally expected that anxious individuals would be motivated
to have more visible relationships to enhance their social sta-
tus and self-esteem. Both anxious and avoidant individuals
might be motivated by beliefs about others’ perceptions of
their poor relationship quality (Etcheverry et al., 2013), but
this might have differential effects on their desired and actual
relationship visibility. We also expected avoidant individuals
to be motivated to have less visible relationships due to the
belief that others perceive them as being independent from
their partners.
We tested our hypotheses across three studies (the online
supplemental material is available at http://pspb.sagepub.
com/supplemental). In Study 1, we examined the associa-
tions between attachment and both desired and actual rela-
tionship visibility (Hypotheses 1a and 1b), as well as
self-presentational goals underlying these associations. In
Study 2, we tested the causal effects of attachment on desired
visibility (Hypothesis 1a). In Study 3, we investigated the
association between attachment and partners’ actual relation-
ship visibility (Hypothesis 2). We also examined whether
daily variations in insecurity about a partner’s feelings are
associated with more relationship visibility (Hypothesis 3a)
and whether attachment anxiety moderates this association
(Hypothesis 3b).
Study 1
Study 1 aimed to test whether anxiety and avoidance are
associated with relationship visibility. We also examined the
interaction between anxiety and avoidance to assess attach-
ment security. Furthermore, Study 1 explored motivations
underlying relationship visibility, based on self-presentation
goals from the impression management literature (social and
material outcomes, self-esteem, and identity development;
Leary & Kowalski, 1990).
In this and subsequent studies, we decided to include only
heterosexual participants in the analyses. For heterosexual
individuals, displaying a relationship primarily conveys to
others the knowledge that an individual is in a relationship
and that it is central to the self-image being portrayed.
However, for non-heterosexual individuals, displaying one’s
relationship also discloses one’s minority sexual orientation.
As such, this decision may be unrelated to whether the rela-
tionship itself is central to one’s self-image (see Bogaert &
by guest on September 18, 2014psp.sagepub.comDownloaded from

4 Personality and Social Psychology Bulletin
Hafer, 2009), especially given that being part of a socially
marginalized relationship can harm well-being (Lehmiller,
2012). As a result, relationship visibility is likely more com-
plicated for individuals who do not identify as heterosexual.
Investigating the correlates of disclosing sexual orientation
on Facebook falls beyond the scope of the current research,
so we excluded non-heterosexual individuals from the cur-
rent analyses.
Method
Participants and procedure. We recruited Facebook users in
romantic relationships from Mechanical Turk (MTurk), an
online marketplace where users complete tasks for compen-
sation. Participants recruited through MTurk are more repre-
sentative of the U.S. population than are typical online
samples (Buhrmester, Kwang, & Gosling, 2011). We
obtained usable data1 from 217 participants2 (42.9% male,
56.7% female, 0.5% did not report their sex; relationship
duration M = 5.23 years, SD = 5.85; age M = 30.71 years,
SD = 9.23).
Participants reported on aspects of their Facebook profile
page, including their current profile picture and relationship
status, and completed measures assessing attachment,
motives for visibility, and control items. They were subse-
quently debriefed and compensated 50 cents, consistent with
MTurk payment standards. The order of the Facebook and
relationship items was counterbalanced.
Measures
Unless otherwise noted, all items were assessed with 7-point
scales (1 = strongly disagree, 7 = strongly agree).
Relationship visibility. Participants completed a measure of
desired relationship visibility (three items; α = .85; M = 4.29,
SD = 1.79; for example, “It is important to me that my Face-
book friends can tell that I am in a relationship”). They also
reported their profile picture content and current relationship
status on Facebook (coded 0 = non-dyadic, 1 = dyadic). The
two measures of actual visibility were correlated (r = .20)
and were averaged to create a composite measure of actual
relationship visibility (0 = no visibility [14.7%], 0.5 = either
a dyadic profile picture or a dyadic relationship status
[64.1%], 1 = both forms of visibility [21.2%]).
Attachment. Participants completed the short form of the
Experiences in Close Relationships Scale (Wei, Russell,
Mallinckrodt, & Vogel, 2007; 12 items; avoidance α = .82,
M = 2.47, SD = 1.17; anxiety α = .81, M = 3.50, SD = 1.36).
Motivations. We constructed items assessing motivations for
relationship visibility based on the impression management
literature and pilot testing. For the pilot test, we asked 52
participants from a small college in the Northeastern United
States to list reasons why people in relationships might or
might not post their relationship statuses on Facebook. Based
on these responses, we selected possible motivations that
aligned with self-presentation goals from the impression
management literature.
Social outcomes. One item assessed the extent to which
relationship visibility provides social status (“Other people
knowing that I am in a relationship gives me social status,”
M = 3.94, SD = 1.68).
Self-esteem. One item assessed the extent to which rela-
tionship visibility provides self-esteem (“Other people
knowing that I am in a relationship makes me feel better
about myself,” M = 4.47, SD = 1.79).
Development of identity. We measured two aspects of iden-
tity relevant to romantic relationships. Three items assessed
whether participants felt that others perceived that they had
high relationship quality (e.g., “Other people think that I
have a happy, stable relationship”; α = .88, M = 5.67, SD =
1.22). Two items assessed whether participants thought oth-
ers perceived them as being independent from their partners
(e.g., “Other people think that I am independent from my
partner”; r = .35, M = 4.45, SD = 1.27).
Control measures. We included a number of control mea-
sures, including the 10-Item Personality Inventory (Gosling,
Rentfrow, & Swann, 2003; 2 items for each of five personal-
ity dimensions: agreeableness r = .25, M = 5.27, SD = 1.20;
extraversion r = .58, M = 3.79, SD = 1.59; conscientiousness
r = .38, M = 5.45, SD = 1.22; neuroticism r = .57, M = 3.13,
SD = 1.50; and openness r = .36, M = 5.02, SD = 1.30). Par-
ticipants also completed the Rosenberg Self-Esteem Scale
(Rosenberg, 1965; 10 items; α = .92, M = 3.14, SD = .68; 1
= strongly disagree, 4 = strongly agree); a measure of self-
monitoring (Snyder & Gangestad, 1986; 18 items; α = .76,
M = 3.81, SD = .79); and a measure of self-concept clarity
(Campbell et al., 1996; 12 items; α = .94, M = 4.61,
SD = 1.43).
Time on Facebook. Participants reported how often they use
Facebook on an 8-point scale (1 = rarely, 8 = more than 2 hr
a day).
Results and Discussion
Attachment and relationship visibility. To assess the association
between attachment and relationship visibility, anxiety,
avoidance, and their interaction were entered into two mul-
tiple regressions predicting desired visibility and actual visi-
bility, controlling for time spent on Facebook (see Table 1
for correlations between all variables).3 Prior to analysis,
variables were standardized. As predicted, anxious individu-
als desired higher relationship visibility (β = .21, p = .002,
by guest on September 18, 2014psp.sagepub.comDownloaded from

Emery et al. 5
Table 1. Correlations Between Attachment Dimensions, Motives, and Relationship Visibility in Study 1.
Avoidance Anxiety
Social status
motive
Self-esteem
motive
Relationship
quality motive
Independent
motive
Desired relationship
visibility
Anxiety .30**
Social status motive −.03 .03
Self-esteem motive −.14* .12†.58**
Relationship quality motive −.59** −.40** .20* .23*
Independent motive −.03 −.09 −.15* −.16* .01
Desired relationship visibility −.24** .11†.20* .38** .30** −.15*
Actual relationship visibility −.15* .07 .009 .18* .24** −.16* .39**
†p < .15. *p < .05. **p < .001.
95% confidence interval [CI] = [0.08, 0.34]), and avoidant
individuals desired lower visibility (β = −.28, p < .001, 95%
CI = [−0.42, −0.15]), F(4, 212) = 8.97, p < .001, R2 = .15. The
interaction was not significant, (β = .06, p = .36, 95% CI =
[−0.08, 0.21]). Also, as predicted, avoidance was negatively
associated with actual visibility (β = −.19, p = .008, 95% CI =
[−0.34, −0.05]) and anxiety was marginally positively associ-
ated with actual visibility (β = .13, p = .06, 95% CI = [−0.008,
0.28]), again controlling for time spent on Facebook. The
interaction was not significant (β = .03, p = .62, 95% CI =
[−0.11, 0.19]), F(4, 212) = 2.09, p = .08, R2 = .04.
Motivations for relationship visibility. We next examined associa-
tions between attachment and self-presentational motives. In
each analysis, avoidance, anxiety, and their interaction were
entered into a multiple regression predicting self-presenta-
tional motive, and all variables were standardized. Contrary
to predictions, neither avoidance (β = −.04, p = .55, 95% CI =
[−0.19, 0.10]), anxiety (β = .04, p = .60, 95% CI = [−0.11,
0.18]), nor their interaction (β = .000, p = 1.00, 95% CI =
[−0.15, 0.15]) was associated with perceived social status
from others knowing about the relationship, F(3, 213) = 0.16,
p = .92, R2 = .002. As expected, anxiety (β = .19, p = .007,
95% CI = [0.05, 0.33]) was positively associated with self-
esteem from others knowing about the relationship. However,
avoidance was also negatively associated with self-esteem
motive (β = −.20, p = .004, 95% CI = [−0.34, −0.06]); the
interaction was not significant (β = .08, p = .26, 95% CI =
[−0.06, 0.23]), F(3, 213) = 4.24, p = .006, R2 = .06. As pre-
dicted, both avoidance (β = −.52, p < .001, 95% CI = [−0.63,
−0.41]) and anxiety (β = −.24, p < .001, 95% CI = [−0.35,
−0.13]) were negatively associated with beliefs that others
perceived one has having high relationship quality. There was
no association with the interaction (β = .03, p = .52, 95% CI =
[−0.08, 0.15]), F(3, 213) = 48.22, p < .001, R2 = .40. Contrary
to predictions, avoidance was not associated with beliefs
about others’ perceptions of independence from one’s partner
(β = .003, p = .97, 95% CI = [−0.14, 0.14]), and neither was
anxiety (β = −.10, p = .15, 95% CI = [−0.26, 0.04]); the inter-
action was marginally significant (β = −.12, p = .08, 95% CI
= [−0.28, 0.02]), F(3, 213) = 1.54, p = .21, R2 = .02.
We then examined which motives mediated the associa-
tions between attachment and relationship visibility. We con-
ducted two multiple mediation analyses using bootstrapping;
we used 5,000 bootstrap re-samples and considered the medi-
ation significant if the CI did not include zero (Preacher &
Hayes, 2008). Self-esteem motive and perceived relationship
quality motive were entered as mediators simultaneously, as
the other two motives were not associated with anxiety or
avoidance. All analyses controlled for time spent on Facebook.
Self-esteem motive (95% CI = [−0.17, −0.02]) and per-
ceived relationship quality motive (95% CI = [−0.35, −0.05])
both mediated the association between avoidance and desired
visibility, controlling for anxiety, F(5, 211) = 15.35, p < .001,
R2 = .25. Avoidant individuals reported that other people
knowing about their relationships would make them feel
worse about themselves and thought others perceived them
to have poor relationship quality, and these accounted for
their low desired relationship visibility. Similarly, self-
esteem motive (95% CI = [0.01, 0.15]) and perceived rela-
tionship quality motive (95% CI = [−0.20, −0.02]) mediated
the association between anxiety and desired visibility, con-
trolling for avoidance, F(5, 211) = 15.35, p < .001, R2 = .27.
Anxious individuals perceived that others knowing about
their relationships would make them feel better about them-
selves and that other people perceived them to have poor
relationship quality, and these accounted for their high
desired relationship visibility (Table 2).
We then considered actual visibility as an outcome.
Perceived relationship quality motive (95% CI = [−0.06,
−0.01]) but not self-esteem motive (95% CI = [−0.02,
0.0009]) mediated the association between avoidance and
actual visibility, F(5, 211) = 4.58, p < .001, R2 = .10. Avoidant
individuals believed that others perceived them to have poor
relationship quality, which accounted for their low actual
relationship visibility. Perceived relationship quality motive
(95% CI = [−0.04, −0.002]), but not self-esteem motive
(95% CI = [−0.0006, 0.01]), mediated the association
between anxiety and actual visibility, F(5, 211) = 4.58, p <
.001, R2 = .08. Anxious individuals felt that others perceived
them to have low-quality relationships, which accounted for
their higher actual relationship visibility (Table 3).
by guest on September 18, 2014psp.sagepub.comDownloaded from

6 Personality and Social Psychology Bulletin
Table 3. Motives Mediating Association Between Attachment and Actual Relationship Visibility in Study 1.
Model Independent variable Dependent variable b SE t
Avoidance → Actual visibility Avoidance Self-esteem motive −.29* 0.11 −2.68
Avoidance Perceived relationship quality motive −.53** 0.06 −9.19
Self-esteem motive Actual relationship visibility .02 0.01 1.37
Perceived relationship quality motive Actual relationship visibility .07* 0.02 3.09
AvoidanceaActual relationship visibility −.05* 0.02 −2.65
AvoidancebActual relationship visibility −.008 0.02 −.39
Anxiety → Actual visibility Anxiety Self-esteem motive .24* 0.09 2.59
Anxiety Perceived relationship quality motive −.22** 0.05 −4.42
Self-esteem motive Actual relationship visibility .02 0.01 1.37
Perceived relationship quality motive Actual relationship visibility .07* 0.02 3.09
AnxietyaActual relationship visibility .03 0.02 1.81
AnxietybActual relationship visibility .04* 0.02 2.39
Note. Betas are unstandardized.
aThe total effect without mediators present
bThe direct effect with mediators present.
*p < .05. **p < .001.
Table 2. Motives Mediating the Association Between Attachment and Desired Relationship Visibility in Study 1.
Model
Independent
variable
Dependent
variable b SE t
Avoidance → Desired visibility Avoidance Self-esteem motive −.29* 0.11 −2.68
Avoidance Perceived relationship quality
motive
−.53** 0.06 −9.19
Self-esteem motive Desired relationship visibility .27** 0.06 4.27
Perceived relationship quality
motive
Desired relationship visibility .36* 0.12 3.07
AvoidanceaDesired relationship visibility −.42** 0.10 −4.14
AvoidancebDesired relationship visibility −.16 0.11 −1.40
Anxiety → Desired visibility Anxiety Self-esteem motive .24* 0.09 2.59
Anxiety Perceived relationship quality
motive
−.22** 0.05 −4.42
Self-esteem motive Desired relationship visibility .27** 0.06 4.27
Perceived relationship quality
motive
Desired relationship visibility .36* 0.12 3.07
AnxietyaDesired relationship visibility .26* 0.09 2.98
AnxietybDesired relationship visibility .28* 0.09 3.15
Note. Betas are unstandardized.
aThe total effect without mediators present.
bThe direct effect with mediators present.
*p < .05. **p < .001.
Testing alternative explanations. Next, we examined other vari-
ables that might also predict desired or actual relationship vis-
ibility. Prior to analysis, variables were standardized. We
entered anxiety, avoidance, their interaction, and time spent
on Facebook, along with agreeableness, extraversion, consci-
entiousness, neuroticism, openness, self-esteem, self-moni-
toring, and self-concept clarity, into a series of multiple
regressions in turn predicting desired visibility, actual visibil-
ity, and motives for visibility. The only control variables that
significantly predicted desired visibility were openness (β =
−.19, p = .008, 95% CI = [−0.33, −0.05]) and time spent on
Facebook (β = .21, p = .002, 95% CI = [0.08, 0.34]); extraver-
sion was marginally associated with desired visibility (β =
.15, p = .06, 95% CI = [−0.006, 0.31]). The association
between each attachment dimension and desired visibility
remained significant with control variables included. In the
multiple regression predicting actual visibility, the associa-
tion between avoidance and actual visibility remained signifi-
cant, and the association between anxiety and actual visibility
was still marginally significant. The only control variable
by guest on September 18, 2014psp.sagepub.comDownloaded from

Emery et al. 7
associated with actual visibility was self-concept clarity
(β = −.20, p = .04, 95% CI = [0.005, 0.39]).
We then tested whether these same variables might also
predict the motives underlying relationship visibility. The
only control variables significantly associated with self-
esteem motive were openness to experience (β = −.21,
p = .005, 95% CI = [−0.35, −0.06]) and self-concept clarity
(β = −.19, p = .04, 95% CI = [−0.37, −0.01]); the associations
between each attachment dimension and self-esteem motive
remained significant. In the multiple regression predicting
relationship quality motive, avoidance and anxiety remained
significantly associated with this motive, and the motive was
also associated with self-esteem (β = .22, p = .006, 95%
CI = [0.06, 0.36]).
Discussion. Taken together, the results of Study 1 showed that
anxiety and avoidance are uniquely associated with relation-
ship visibility. Avoidant individuals reported less desired and
actual relationship visibility; anxious individuals reported
more desired visibility and marginally more actual visibility.
These effects remained when including control variables.
Study 1 also tested possible motivations for desired and actual
relationship visibility. Avoidant and anxious individuals were
both motivated by their belief that others perceive them to
have poor relationship quality and their belief that other peo-
ple knowing about their relationships affects their self-esteem.
However, these motivations had divergent effects on relation-
ship visibility. That is, poor perceived relationship quality
among anxious individuals was associated with higher desired
and actual relationship visibility, whereas poor perceived rela-
tionship quality among avoidant individuals was associated
with lower desired and actual relationship visibility. Likewise,
anxious individuals believed that other people knowing about
their relationships would make them feel better about them-
selves, in turn predicting higher desired relationship visibility.
Avoidant individuals reported that other people knowing about
their relationships would make them feel worse about them-
selves, which was associated with lower desired visibility.
Study 2
Study 2 tested the causal association between attachment and
desired relationship visibility by experimentally priming
attachment. We expected that individuals primed with avoid-
ance would express lower desire for relationship visibility
than would those primed with anxiety. We anticipated that
those primed with anxiety would report higher desire for rela-
tionship visibility than would those primed with avoidance.
Method
Participants. We recruited Facebook users in relationships
from MTurk and obtained usable data4 from 586 partici-
pants5 (46.1% male; 53.9% female; relationship duration
M = 6.37 years, SD = 6.88; age M = 31.14 years, SD = 9.35).
Procedure. Participants were randomly assigned to receive
either an avoidance, anxiety, or control prime. In the avoid-
ance condition, participants visualized a time they felt
uncomfortable being too close to their partner, and in the
anxiety condition, they visualized a time their partner seemed
reluctant to get close to them (adapted from Bartz & Lydon,
2004). In the control condition, participants thought about
their plans for the weekend. They were then asked to write a
few sentences about what they had visualized. Participants
subsequently reported their desired relationship visibility,
were debriefed, and were compensated 50 cents.
Measures. Participants reported their desired relationship
visibility, anxiety, and avoidance on the same measures as in
Study 1 (desired visibility α = .84, M = 4.52, SD = 1.62;
avoidance α = .83, M = 2.38, SD = 1.08; anxiety α = .78, M =
3.51, SD = 1.24), and reported how often they used
Facebook.
Results and Discussion
As a manipulation check, we examined the effect of condi-
tion on the measures of avoidance and anxiety. Using a
between-subjects ANCOVA, we found a significant differ-
ence between conditions on the measure of avoidance, con-
trolling for anxiety, F(2, 582) = 4.40, p = .01, partial η2 = .02.
A Tukey’s least significant difference (LSD) post hoc test
showed that participants in the avoidance condition (M =
2.56, SE = 0.08, 95% CI = [2.41, 2.72]) scored significantly
higher on avoidance than did the anxiety condition (M =
2.33, SE = 0.07, 95% CI = [2.18, 2.47]), p = .03. In addition,
individuals in the avoidance condition reported higher avoid-
ance than did those in the control condition (M = 2.27, SE =
0.07, 95% CI = [2.14, 2.40]), p = .004. There was no differ-
ence between those in the anxiety and the control conditions
on avoidance, p = .59. A second between-subjects ANCOVA
revealed an overall effect of condition on anxiety that
approached significance, controlling for avoidance, F(2,
582) = 2.92, p = .06, partial η2 = .01. A Tukey’s LSD post hoc
test found that those in the anxiety condition (M = 3.65, SE =
0.08, 95% CI = [3.49, 3.82]) significantly differed from those
in the control condition (M = 3.38, SE = 0.08, 95% CI =
[3.23, 3.53]) on the measure of anxiety, p = .02. There were
no differences between those in the avoidance condition (M
= 3.55, SE = 0.09, 95% CI = [3.37, 3.72]) and those in the
anxiety condition, p = .40, on anxiety; however, the means
were in the predicted directions. There was also no differ-
ence between those in the avoidance condition and those in
the control condition, p = .16, on anxiety.
Testing our primary hypothesis, a between-subjects
ANCOVA revealed a significant main effect of condition on
desired visibility, controlling for time spent on Facebook,
F(2, 581) = 3.35, p = .04, partial η2 = .01 (Figure 1). A
Tukey’s LSD post hoc test revealed that individuals in the
avoidance condition expressed lower desire for relationship
by guest on September 18, 2014psp.sagepub.comDownloaded from

8 Personality and Social Psychology Bulletin
visibility (M = 4.27, SE = 0.12, 95% CI = [4.04, 4.50]) than
did those in the anxiety condition (M = 4.66, SE = 0.10, 95%
CI = [4.44, 4.88]), p = .02. Moreover, individuals in the con-
trol condition (M = 4.59, SE = 0.10, 95% CI = [4.40, 4.78])
expressed more desired visibility than did those in the avoid-
ance condition, p = .04. There was no significant difference
between those in the anxiety condition and those in the con-
trol condition, p = .64.
Discussion. In Study 2, we demonstrated a causal association
between attachment and relationship visibility. Participants
primed with avoidant attachment expressed lower desire for
relationship visibility than did those primed with anxious
attachment, suggesting that attachment alters desired visibil-
ity. Those in the avoidance condition also reported margin-
ally less desire for visibility than did those in the control
condition; the anxiety and control conditions did not differ.
These findings, however, are only from one member of the
couple. Given the importance of a dyadic approach in under-
standing relationship processes (e.g., Campbell & Simpson,
2013), it is necessary to consider whether partners might
influence each other’s relationship visibility. Therefore,
Study 3 focused on actor and partner effects of attachment on
relationship visibility.
Study 3
Study 3 extended the previous studies in two key ways. First,
we collected data from both members of romantic couples,
allowing us to test whether one partner’s attachment predicts
the other partner’s relationship visibility. Previous research
has demonstrated that one partner’s attachment influences
the other partner’s feelings about the relationship (Campbell
et al., 2001; Simpson, 1990). Second, we assessed romantic
partners’ daily Facebook use over 2 weeks, enabling us to
examine different forms of actual relationship visibility.
These varied in level of permanence, with relationship status
and profile picture as more enduring and daily posting as
more transient. Also, because people’s insecurity about their
partners’ feelings can vary from day to day (Lemay &
Dudley, 2011), we tested whether and how daily insecurity
influences relationship visibility.
In Study 3, we predicted that people higher in avoidance
and those with more avoidant partners would be less likely to
post a dyadic relationship status. Based on the marginal asso-
ciation between anxiety and actual visibility in Study 1, we
did not think that anxiety would be directly associated with
more permanent forms of relationship visibility such as hav-
ing a dyadic profile picture or relationship status. However,
we predicted that it would be associated with daily posting,
as one’s partner has less control over this form of visibility.
Finally, we expected that on days when people felt more
insecure about their partner’s feelings than they typically do,
they would share more relationship-relevant information on
Facebook. However, we expected this association to be mod-
erated by attachment, such that anxious individuals in par-
ticular would post more about their relationships on days
when they felt insecure.
Method
Participants. Participants were 108 heterosexual dating cou-
ples (216 individuals)6 recruited from a small university in
Canada; each partner was paid C$40 for participating. The
participants ranged in age from 19 to 31 (M = 21.05, SD =
0.94) and had been together from 2 to 73 months (M = 19.78,
SD = 15.49; 9% cohabiting). To be eligible for the study,
both partners had to participate and be current Facebook
users.
Procedure. On the first day of the study, participants com-
pleted background measures and “friended” the study’s
Facebook page. On joining the study, they consented to allow
us to download their Facebook profiles. Participants com-
pleted an online survey each night for 2 weeks, indepen-
dently from their partners. To maximize compliance with the
daily part of the protocol, reminder emails were sent to the
participants who had not completed their daily diaries by
10:00 p.m. each night. On average, participants completed
12.45 diaries across the 14 days (range = 1-14, SD = 3.72) for
a total of 2,689 days across participants.
Background Measures
Actual relationship visibility. From the downloaded Facebook
profiles, two trained coders rated each participant’s relation-
ship status (1 = in a relationship [with partner’s name], 2 = in
a relationship [without partner’s name], 3 = married or
Figure 1. Effect of attachment primes on desired relationship
visibility in Study 2.
by guest on September 18, 2014psp.sagepub.comDownloaded from

Emery et al. 9
engaged, 4 = single, 5 = in an open relationship, 6 = no status
listed, 7 = other). The same coders also rated whether or not
the couple was present in the photo (1 = dyadic photo, 0 =
non-dyadic photo). Because both couple members partici-
pated, after the ratings were complete, coders verified that the
other person in the photo was the person’s partner. Coders
were blind to participants’ scores on the survey measures.
Attachment. Participants completed the same measure of
attachment as in previous studies (avoidance α = .87,
M = 1.97, SD = .96; anxiety α = .79, M = 3.28, SD = 1.24).
Daily Measures
Facebook posts. Participants responded to one item (“I shared
information about my relationship or my partner on Face-
book today; that is, posted a status update, wall post, photo
comment, or photos about or with my partner”; 1 = not at all,
7 = very much).
Daily feelings of insecurity. Participants responded to one item
(“I felt insecure about my partner’s feelings for me today”;
1 = not at all, 7 = very much).
Time spent on Facebook. Participants reported the number of
minutes they spent on Facebook each day (range = 0-600,
M = 32.07, SD = 49.08).
Results and Discussion
Coding Facebook profiles. Coders had perfect agreement
regarding participants’ Facebook relationship statuses
(kappa = 1.00). Participants with an “in a relationship” status
(with or without their partner’s name) or who indicated they
were married/engaged were re-coded 1 (dyadic relationship
status), and those without a status or had an “other” status
were coded 0 (non-dyadic relationship status).
The coding of participants’ Facebook profile picture
resulted in high initial agreement (kappa = .97). The coders
only disagreed on two of the photos, and after discussion,
both of these photos were considered non-dyadic. As in
Study 1, having a dyadic relationship status and having a
dyadic profile picture were positively correlated (r = .25).
We averaged these two variables to create a composite mea-
sure of actual relationship visibility (0 = no visibility [39.3%],
0.5 = either a dyadic profile picture or a dyadic relationship
status [37.7%], 1 = both forms of visibility [23.0%]).
Attachment and relationship visibility. To test our first predic-
tion that avoidant individuals and the partners of people high
in avoidance would have less visible relationships on Face-
book, we used multilevel modeling with mixed models in
SPSS 20.0. The Actor Partner Interdependence Model
(APIM) guided the analyses (Kenny, Kashy, & Cook, 2006),
such that actor avoidance, actor anxiety, partner avoidance,
and partner anxiety were entered as predictors into the model
simultaneously. In these analyses, we report unstandardized
betas, which can be interpreted as the increase in the depen-
dent variable for every one-unit increase in the independent
variable. All analyses controlled for time spent on Facebook.
As expected, more avoidant participants had less visible rela-
tionships on Facebook (see Table 4). Moreover, controlling
for an individual’s own attachment, those who had a partner
who was higher in avoidance also had less visible relation-
ships on Facebook. Anxiety was not significantly associated
with a person’s own or his or her partner’s actual relationship
visibility on Facebook, and there were no significant interac-
tions between avoidance and anxiety predicting relationship
visibility.
Daily insecurity and relationship posting. Our next set of predic-
tions concerned fluctuations in a person’s feelings of insecu-
rity and daily posting about their relationship on Facebook.
We tested hypotheses with multilevel modeling using mixed
models in SPSS 20.0, testing a two-level cross model with
random intercepts with persons nested within dyads, and per-
son and days crossed to account for the fact that both partners
completed the daily surveys on the same days. To avoid con-
founding within- and between-person effects, we used tech-
niques appropriate for a multilevel framework, partitioning
the daily predictor (i.e., feelings of insecurity) into their
within- and between-variance components, which were per-
son-mean centered and aggregated, respectively (Rauden-
bush, Bryk, Cheong, & Congdon, 2004). As such, for the
analyses with daily-level predictors, the unstandardized beta
is interpreted as the change in the dependent variable for
every one-unit deviation from the person’s own mean on the
independent variable. We again controlled for time spent on
Facebook. On days when people felt more insecure about
their partner’s feelings than they typically do, they posted
more relationship-relevant information on Facebook
(b = .03, SE = 0.01, t = 2.17, p = .03, 95% CI = [0.003, 0.05]).
A person’s daily feelings of insecurity were not significantly
associated with his or her partner’s daily posting on
Facebook.
Next, we tested whether anxiety moderated the effect of
daily insecurity on relationship visibility. First, anxiety, b =
.29, SE = 0.04, t(126.80) = 6.71, p < .001, 95% CI = [0.21,
0.38], and avoidance, b = .26, SE = 0.05, t(138.85) = 4.95,
Table 4. Attachment Predicting Actual Relationship Visibility in
Study 3.
Model b SE t df p 95% CI
Avoidance −.13 0.03 −4.78 171.56 <.001 [−0.18, −0.08]
Anxiety −.01 0.02 −0.59 182.94 .56 [−0.06, 0.03]
Partner avoidance −.07 0.03 −2.61 173.37 .01 [−0.12, −0.02]
Partner anxiety −.001 0.02 −0.08 188.41 .93 [−0.04, 0.04]
Note. CI = confidence interval.
by guest on September 18, 2014psp.sagepub.comDownloaded from

10 Personality and Social Psychology Bulletin
p < .001, 95% CI = [0.016, 0.37], were both significantly
associated with daily insecurity. In addition, people who
were higher in anxiety posted more information about their
relationship on Facebook on a daily basis, b = .05, SE = 0.01,
t(1066.84) = 3.05, p = .005, 95% CI = [0.02, 0.07], and peo-
ple higher in avoidance posted less information about their
relationship, b = −.05, SE = 0.02, t(939.11) = −2.42, p = .02,
95% CI = [−0.09, −0.01]. We then entered attachment anxi-
ety, avoidance, daily insecurity, and the interaction between
daily feelings of insecurity and both anxiety and avoidance
as predictors of daily relationship posting. However, in con-
trast to our predictions, anxiety, b = −.001, SE = 0.01,
t(2119.33) = −0.05, p = .96, 95% CI = [−0.02, 0.02], did not
moderate the association between daily feelings of insecurity
and relationship posting. Avoidance also did not moderate
this effect, b = .003, SE = 0.01, t(2059.10) = 0.23, p = .83,
95% CI = [−0.02, 0.03].
Testing alternative explanations. In Study 3, we also tested for
combined effects of partner’s attachment styles; that is,
whether the effect of a person’s anxiety or avoidance was
strengthened or attenuated based on his or her partner’s anxi-
ety or avoidance (i.e., actor–partner interactions). We did not
find any significant interactions between an actor’s anxiety
and avoidance and their partner’s anxiety and avoidance, sug-
gesting that, in the current study, partners’ attachment styles
do not interact in a consistent way to predict relationship vis-
ibility. Moreover, we examined whether the daily effects held
when controlling for overall amount of information posted
about the self that day to determine whether there is a unique
effect of posting about the relationship. Although the effect of
daily insecurity on daily posting was slightly weakened (b =
.02, SE = 0.01, t = 1.65 p = .10, 95% CI = [−0.01, 0.04]), the
overall pattern of effects remained the same. The slight dimin-
ishing of the effect on relationship posting is likely because
daily posting about the relationship and daily posting about
the self were correlated (r = .39, p < .001).
Discussion. Study 3 investigated relationship visibility from a
dyadic perspective. Consistent with the previous studies,
avoidant individuals had less visible relationships, whereas
anxiety did not directly predict more permanent forms of vis-
ibility. Moreover, the partners of people high in avoidance
also had less visible relationships on Facebook, controlling
for their own avoidance and anxiety. On a daily basis, people
were more likely to make their relationships visible by post-
ing about their partners or their relationships on days when
they felt more insecure about their partner’s feelings for
them. However, neither anxiety nor avoidance moderated
these effects.
General Discussion
Some people ardently want others to see that they are in a
relationship; others prefer their relationships to be less
visible. Relationship visibility is conceptualized as a form of
impression management (Leary & Kowalski, 1990), which
reflects whether individuals want their relationships to
appear to be an important aspect of their self-concept. These
studies focused on relationship visibility in the context of
Facebook, a particularly rich environment for understanding
this construct, because users determine the visibility of their
relationships. Concealing a relationship corrodes relation-
ship quality (Foster & Campbell, 2005), so understanding
antecedents of public displays of a relationship provides
insights into the maintenance of romantic relationships.
Moreover, given the centrality of relationships in the self-
concept (Aron et al., 1991), considering the role of relation-
ships in the self-images that people convey to others is
crucial for understanding impression management.
Three studies tested the central predictions that avoidance
would predict less desired and actual visibility, whereas anx-
iety would predict greater desired visibility. Study 1 estab-
lished support for the predicted association between
attachment and relationship visibility and tested motivations
underlying relationship visibility. Anxious individuals
believed that others knowing about their relationships would
make them feel better about themselves, which was associ-
ated with desired visibility, and that other people perceived
that they had poor relationship quality, which was associated
with desired and actual visibility. Avoidant individuals
believed that others knowing about their relationships would
make them feel worse about themselves, which was associ-
ated with lower desired visibility, and that others thought
they had poor relationship quality, which was associated
with lower desired and actual visibility. Study 2 found exper-
imental evidence for the link between attachment and rela-
tionship visibility. Individuals primed with anxiety reported
greater desire for visibility than did those primed with avoid-
ance. Study 3 examined relationship visibility in a dyadic
daily diary paradigm. More avoidant individuals and the
partners of people higher in avoidance reported less overall
actual relationship visibility. Furthermore, on days when
people felt more insecure about their partner’s feelings for
them, they posted more about their relationships on Facebook
than usual.
Implications and Future Directions
The current research extends both impression management
theory and attachment theory by fusing these previously dis-
parate literatures. The majority of studies on attachment in
romantic relationships have only examined its effects on
dynamics within the relationship itself. This research, how-
ever, suggests that attachment also affects the extent to which
people want their romantic relationship to be a central part of
the image that outsiders form of them and is the first to our
knowledge to apply attachment theory to the realm of impres-
sion management. Notably, a person’s attachment avoidance
was associated with his or her partner’s impression
by guest on September 18, 2014psp.sagepub.comDownloaded from

Emery et al. 11
management, suggesting that in the context of relationships,
impression management may be a dyadic process. Our
examination of motives underlying impression management
was largely exploratory and found that the same motives
(self-esteem maintenance and beliefs about others’ percep-
tions of one’s relationship quality) mediated both anxious
and avoidant individuals’ relationship visibility.
Both anxious and avoidant individuals were motivated by
thinking that others perceived them to have poor relationship
quality, but this belief translated into high relationship visi-
bility for anxious individuals and low relationship visibility
for avoidant individuals. This research suggests that relation-
ship visibility may fulfill distinct impression management
goals for different individuals. The self-images that people
convey to others can include the current self-concept (so that
others hold correct perceptions of them) or the desired self
(the person they would like to be; Leary & Kowalski, 1990).
People who make their relationships visible to others are per-
ceived to have higher relationship quality (Emery, Muise,
Alpert, & Le, in press). Given that both anxious and avoidant
individuals have poor relationship quality (Etcheverry et al.,
2013), anxious individuals may be c onstructing an image of
their desired selves, whereas avoidant individuals may be
creating images of their actual selves. However, future
research should test this prediction more directly.
Although our hypotheses were largely supported, it was
surprising that anxiety did not moderate the effects of daily
insecurity about a partner’s feelings on daily posting about
the relationship, given previous research on daily insecurity
among anxious individuals (Lemay & Dudley, 2011). Our
primary hypothesis for the daily analyses concerned anxious
individuals, but we were also surprised that avoidance did
not moderate these effects either. In other words, the effects
of daily insecurity on daily relationship visibility seem to
hold for all individuals, even those high in avoidance. At first
glance, this result appears to contradict our other findings
that avoidant individuals express low desire for relationship
visibility and tend not to make their relationships visible to
others. However, it suggests that there may be a difference
between more permanent forms of relationship visibility
(e.g., a profile picture or relationship status) and more tran-
sient ones (e.g., daily posting). Perhaps attachment, as a pro-
cess that describes how people chronically view the self and
others, is more directly relevant for more permanent relation-
ship visibility, whereas daily changes in how people view
their partners may be more pertinent to transient relationship
visibility.
These findings suggest the utility of Facebook as a context
for studying how individuals portray their relationships to
others. Facebook can be a valuable means of examining
behaviors that are difficult to investigate through other meth-
ods (Wilson, Gosling, & Graham, 2012). Facebook provides
a clear set of behaviors for operationalizing relationship visi-
bility (a dyadic profile picture, a relationship status, and post-
ing about one’s partner), but in the offline world, relationship
visibility may be subtler. Examining relationship visibility in
the offline world would require determining all possible ways
that people could refer to their partners in conversation or
subtly display their relationships.
However, this could be an important step for future
research. It would be fascinating to examine whether people
spontaneously mention their relationships in conversation to
make their relationships visible. We would expect specific
offline behaviors to parallel our findings with online behav-
ior: For example, our findings for posting a relationship sta-
tus or having a dyadic profile picture might map onto the act
of wearing a wedding ring, whereas posting about the rela-
tionship is likely analogous to mentioning one’s partner in
casual conversation. Creating a discussion paradigm to
observe when people mention their partners to new acquain-
tances would be a fruitful next step. Furthermore, such a
study would enable a more nuanced examination of what
constitutes relationship visibility. There may be a difference,
for instance, between mentioning one’s partner’s name in
conversation and talking about one’s relationship, and the
language used in the process of making one’s relationship
visible might reflect relationship quality or other motivations
underlying relationship visibility. Future research would
benefit from coding subtler forms of relationship visibility.
Strengths and Limitations
The present studies investigated the novel idea of relation-
ship visibility. Although previous research has examined
keeping one’s relationship secret (Foster & Campbell, 2005),
individual differences and motivations underlying the desire
and decision for others to know about one’s relationship have
not previously been investigated. This research suggests that
relationships form part of the self-images that people convey
to others, and it fuses distinct literatures on attachment orien-
tation and impression management.
These questions were examined across survey, experimen-
tal, daily experience, and dyadic methodologies; our samples
included diverse age groups and relatively equal numbers of
male and female participants. These varying methodologies
enable greater confidence in the internal and external validity
of our findings. Some findings were robust across different
samples and methodologies. For instance, the effects of both
anxious and avoidant attachments on desired relationship vis-
ibility were established correlationally in Study 1 and were
replicated experimentally in Study 2. Likewise, the associa-
tion between avoidance and actual visibility was consistent
across the different samples in Studies 1 and 3. At the same
time, the association between anxiety and actual visibility
was less clear. In Study 1, the direct association was margin-
ally significant, but in Study 3, we did not find any effects of
actor or partner anxiety on actual visibility. We suspect that
given the influence that anxious individuals’ partners exert on
them (e.g., Slotter & Gardner, 2012), it is their partners who
determine their actual relationship visibility, although we did
by guest on September 18, 2014psp.sagepub.comDownloaded from

12 Personality and Social Psychology Bulletin
not directly test this hypothesis. As a result, including partner
effects in Study 3 may have resulted in the previously mar-
ginal association becoming non-significant. This explanation
is likely considering that one partner’s anxiety is associated
with the other partner having less interest in closeness (Collins
& Read, 1990); that is, anxious individuals may tend to have
avoidant partners.
The sample sizes in all three studies exceeded 200 partici-
pants, increasing our confidence that our analyses possessed
sufficient statistical power. However, although Internet sam-
ples tend to be more representative than college samples, the
participants in this study tended to be primarily White, and
we only examined relationship visibility among heterosexual
individuals. Future research would benefit from more diverse
samples; for example, comparisons of relationship visibility
in heterosexual and non-heterosexual populations would be
valuable. For instance, in non-heterosexual populations,
relationship visibility might serve the double function of
showing that the relationship is central to one’s self-concept
and making a sexual minority identity visible to one’s social
network.
Although Study 3 examined the dyadic nature of relation-
ship visibility, the current research did not examine how part-
ners react to each others’ relationship visibility. That is, if
partners have different desires for relationship visibility or
different levels of actual relationship visibility, this discrep-
ancy could strain relationship quality. Likewise, it would be
informative to examine how partners negotiate the decision
and timing of posting a relationship status (as doing so
requires the approval of one’s partner), as well as which part-
ner posted a relationship status or dyadic profile picture first.
Examining these questions would shed further light on rela-
tionship visibility as a dyadic process.
Another limitation of the present research is that it solely
investigated relationship visibility on Facebook. Although
Facebook is a useful context for studying relationship visi-
bility, future studies should extend it to other offline environ-
ments. Facebook may differ from other contexts because
information reaches multiple audiences at once, and because
partners are often friends on Facebook, so they may be more
aware of one another’s level of relationship visibility than
they would be in the offline world.
This research focused on insecurely attached individuals,
but we tested the interaction between anxiety and avoidance
to examine the association between secure attachment and
relationship visibility. Unfortunately, the results of the cur-
rent research were inconclusive for secure attachment; none
of the interactions was significant. This unresolved question
is a limitation of the current research, and future research
might benefit from specifically investigating secure attach-
ment in the context of relationship visibility. If the control
condition in Study 2 includes securely attached individuals,
then perhaps secure individuals desire relationship visibility
just as much as do anxious individuals, but different motives
underlie this desire.
Conclusion
Romantic relationships do not exist in isolation; people
experience them through the lens of their broader social
environments, and in turn, they must decide whether to con-
vey information about their relationships to others. The
desire and decision to make a relationship visible to others
reflect people’s fears or aspirations for closeness with their
romantic partners, with avoidant individuals eschewing rela-
tionship visibility and anxious individuals yearning for it.
These studies suggest that people’s views of themselves and
others shape the centrality of their relationships in the self-
concepts that they stage for the outside world.
Acknowledgment
The authors thank Brent Mattingly for his comments on a previous
version of this article.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support
for the research, authorship, and/or publication of this article: This
work was partly supported by a Social Science and Humanities
Research Council (SSHRC) postdoctoral fellowship and an SSHRC
Banting postdoctoral fellowship awarded to Amy Muise and was
partially funded by the University of Guelph–Humber Research
Grant.
Notes
1. Initially, we received 355 responses; however, 97 did not meet
our eligibility criteria (were not in a relationship, did not have
a Facebook account, or did not have a partner with a Facebook
account), submitted multiple surveys, or failed an attention
check. An attention check is a question designed to determine
whether participants are reading questions (Oppenheimer,
Meyvis, & Davidenko, 2009), which may especially be a prob-
lem on MTurk. We also excluded 41 non-heterosexual partici-
pants from analyses.
2. Recent research showing associations between relationship
quality and posting a dyadic profile picture used 115 participants
from MTurk (Saslow, Muise, Impett, & Dubin, 2013, Study 1);
given that we were also testing mediation and ruling out alterna-
tive explanations in this study, we expected that we would need
a slightly larger sample than Saslow and colleagues (2013) col-
lected to detect similar effects.
3. In this and subsequent studies, we re-ran all key analyses includ-
ing gender, relationship duration, and the interaction between
these and both attachment orientations to determine whether
gender or relationship duration moderated any of our effects.
In Study 1, there was a significant interaction between anxiety
and relationship duration predicting actual relationship visibility
(β = −.20, p = .02, 95% confidence interval [CI] = [−0.32, −0.03]),
such that highly anxious individuals were more likely to make
their relationships visible when they had been in relationships
by guest on September 18, 2014psp.sagepub.comDownloaded from

Emery et al. 13
for a shorter time, compared with a longer time. Conversely, less
anxious individuals were more likely to make their relationships
visible when they had been in those relationships for a longer
versus a shorter time. In all other key analyses in Studies 1 to 3,
gender and relationship duration were not significant modera-
tors. The pattern of all effects studies remained the same when
gender and relationship duration were controlled.
4. We initially received 927 responses, but prior to data analysis,
data were cleaned as in Study 1. We also excluded responses
from 4 participants who guessed the purpose of the study and
22 participants who did not complete the essay prime. In addi-
tion, participants indicated whether they could think of a time
such as the one in the prime; responses from 10 participants
who indicated that they were “completely unable to” (a “1” on
a 7-point scale) were removed. We excluded responses from 96
non-heterosexual participants.
5. The attachment prime in this study was adapted from Bartz and
Lydon (2004, Study 1), so we also used this study as a basis
for determining sample size. Bartz and Lydon (2004, Study 1)
reported a sample size of 245 participants. However, MTurk
samples can reduce statistical power, necessitating larger
samples to detect effects comparable with those in traditional
samples. This difference arises from MTurk participants pay-
ing significantly less attention to study materials (Goodman,
Cryder, & Cheema, 2013), which may particularly affect experi-
mental manipulations. We estimated that we would need a larger
sample than reported by Bartz and Lydon (2004) to be confident
that we possessed sufficient power to avoid Type II errors.
6. The measures in this study were collected as part of a larger
investigation of Facebook use and jealousy in romantic rela-
tionships (Muise, Christofides, & Desmarais, 2014). As such,
the sample size was based on Marshall, Bejanyan, Di Castro,
and Lee’s (2013, Study 2) daily diary study on jealousy due to
Facebook use, which recruited 108 couples.
Supplemental Material
The online supplemental material is available at http://pspb
.sagepub.com/supplemental.
References
Aron, A., Aron, E. N., Tudor, M., & Nelson, G. (1991). Close rela-
tionships as including other in the self. Journal of Personality
and Social Psychology, 60, 241-253.
Back, M. D., Stopfer, J. M., Vazire, S., Gaddis, S., Schmukle, S. C.,
Egloff, B., & Gosling, S. D. (2010). Facebook profiles reflect
actual personality, not self-idealization. Psychological Science,
21, 372-374.
Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles
among young adults: A test of a four-category model. Journal
of Personality and Social Psychology, 61, 226-244.
Bartz, J. A., & Lydon, J. E. (2004). Close relationships and the
working self-concept: Implicit and explicit effects of priming
attachment on agency and communion. Personality and Social
Psychology Bulletin, 30, 1389-1401.
Baumeister, R. F., & Jones, E. E. (1978). When self-presentation is
constrained by the target’s knowledge: Consistency and com-
pensation. Journal of Personality and Social Psychology, 36,
608-618.
Baxter, L. A., & Widenmann, S. (1993). Revealing and not reveal-
ing the status of romantic relationships to social networks.
Journal of Social and Personal Relationships, 10, 321-337.
Bogaert, A. F., & Hafer, C. L. (2009). Predicting the timing of com-
ing out in gay and bisexual men from world beliefs, physical
attractiveness, and childhood gender identity/role. Journal of
Applied Social Psychology, 39, 1991-2019.
Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s
Mechanical Turk: A new source of inexpensive, yet high-qual-
ity, data? Perspectives on Psychological Science, 6, 3-5.
Butzer, B., & Campbell, L. (2008). Adult attachment, sexual sat-
isfaction, and relationship satisfaction: A study of married
couples. Personal Relationships, 15, 141-154.
Campbell, J. D., Trapnell, P. D., Heine, S. J., Katz, I. M.,
Lavallee, L. F., & Lehman, D. R. (1996). Self-concept clar-
ity: Measurement, personality correlates, and cultural bound-
aries. Journal of Personality and Social Psychology, 70(1),
141-156.
Campbell, L., & Simpson, J. A. (2013). The blossoming of rela-
tionship science. In J. A. Simpson & L. Campbell (Eds.), The
Oxford handbook of close relationships (pp. 3-12). Oxford,
UK: Oxford University Press.
Campbell, L., Simpson, J. A., Kashy, D. A., & Rholes, W. S.
(2001). Attachment orientations, dependence, and behavior
in a stressful situation: An application of the Actor–Partner
Interdependence Model. Journal of Social and Personal
Relationships, 18, 821-843.
Carpenter, C. J., & Spottswood, E. L. (2013). Exploring romantic
relationships on social networking sites using the self-expan-
sion model. Computers in Human Behavior, 29, 1531-1537.
Collins, N. L., & Allard, L. M. (2001). Cognitive representations
of attachment: The content and function of working models.
In G. O. Fletcher & M. S. Clark (Eds.), Blackwell handbook
of social psychology: Interpersonal processes (pp. 60-85).
Malden, MA: Blackwell.
Collins, N. L., & Read, S. J. (1990). Adult attachment, working
models, and relationship quality in dating couples. Journal of
Personality and Social Psychology, 58, 644-663.
Emery, L. F., Muise, A., Alpert, E., & Le, B. (in press). Do we
look happy? Perceptions of romantic relationship quality on
Facebook. Personal Relationships.
Etcheverry, P. E., Le, B., Wu, T.-S., & Wei, M. (2013). Attachment
and the investment model: Predictors of relationship commit-
ment, maintenance, and persistence. Personal Relationships,
20, 546-567.
Facebook Newsroom. (2014). Company info. Retrieved from http://
newsroom.fb.com/company-info
Foster, C. A., & Campbell, W. K. (2005). The adversity of secret
relationships. Personal Relationships, 12, 125-143.
Fraley, R. C., & Shaver, P. R. (1998). Airport separations: A
naturalistic study of adult attachment dynamics in separating
couples. Journal of Personality and Social Psychology, 75,
1198-1212.
Goodman, J. K., Cryder, C. E., & Cheema, A. (2013). Data col-
lection in a flat world: The strengths and weaknesses of
Mechanical Turk samples. Journal of Behavioral Decision
Making, 26, 213-224.
Gosling, S. D., Rentfrow, P. J., & Swann, W. B. (2003). A very
brief measure of the Big-Five personality dimensions. Journal
of Research in Personality, 37, 504-528.
by guest on September 18, 2014psp.sagepub.comDownloaded from

14 Personality and Social Psychology Bulletin
Hazan, C., & Shaver, P. (1987). Romantic love conceptualized
as an attachment process. Journal of Personality and Social
Psychology, 52, 511-524.
House, W. C. (1980). Effects of knowledge that attributions will
be observed by others. Journal of Research in Personality, 14,
528-545.
Karl, K. A., Hall, L. M., & Peluchette, J. V. (2013). City employee
perceptions of the impact of dress and appearance: You are
what you wear. Public Personnel Management, 42, 452-470.
Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data
analysis. New York, NY: Guilford Press.
Kirkpatrick, L. A., & Davis, K. E. (1994). Attachment style, gender,
and relationship stability: A longitudinal analysis. Journal of
Personality and Social Psychology, 66, 502-512.
Leary, M. R., & Kowalski, R. M. (1990). Impression management:
A literature review and two-component model. Psychological
Bulletin, 107, 34-47.
Lehmiller, J. J. (2012). Perceived marginalization and its associa-
tion with physical and psychological health. Journal of Social
and Personal Relationships, 29, 451-469.
Lemay, E. P., & Dudley, K. L. (2011). Caution: Fragile! Regulating
the interpersonal security of chronically insecure partners.
Journal of Personality and Social Psychology, 100, 681-702.
Leslie, L. A., Huston, T. L., & Johnson, M. P. (1986). Parental
reactions to dating relationships: Do they make a difference?
Journal of Marriage and Family, 48, 57-66.
Loving, T. J., & Agnew, C. R. (2001). Socially desirable responding
in close relationships: A dual-component approach and measure.
Journal of Social and Personal Relationships, 18, 551-574.
Marshall, T. C., Bejanyan, K., Di Castro, G., & Lee, R. A. (2013).
Attachment styles as predictors of Facebook-related jeal-
ousy and surveillance in romantic relationships. Personal
Relationships, 20, 1-22.
Mashek, D. J., & Sherman, M. (2004). Desiring less closeness with
intimate others. In D. J. Mashek & A. Aron (Eds.), Handbook of
closeness and intimacy (pp. 343-356). Mahwah, NJ: Lawrence
Erlbaum.
Mikulincer, M. (1995). Attachment style and the mental representa-
tion of the self. Journal of Personality and Social Psychology,
69, 1203-1215.
Muise, A., Christofides, E., & Desmarais, S. (2014). “Creeping”
or just information seeking? Gender differences in partner
monitoring in response to jealousy on Facebook. Personal
Relationships, 21, 35-50.
Nadkarni, A., & Hofmann, S. G. (2012). Why do people use
Facebook? Personality and Individual Differences, 52, 243-249.
Oppenheimer, D. M., Meyvis, T., & Davidenko, N. (2009).
Instructional manipulation checks: Detecting satisficing to
increase statistical power. Journal of Experimental Social
Psychology, 45, 867-872.
Papp, L. M., Danielewicz, J., & Cayemberg, C. (2012). “Are
we Facebook official?” Implications of dating partners’
Facebook use and profiles for intimate relationship satisfac-
tion. Cyberpsychology, Behavior, and Social Networking, 15,
85-90.
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resam-
pling strategies for assessing and comparing indirect effects
in multiple mediator models. Behavior Research Methods, 40,
879-891.
Raudenbush, S. W., Bryk, A. S., Cheong, Y. F., & Congdon, R. T.
(2004). HLM 6: Hierarchical linear and nonlinear modeling.
Chicago, IL: Scientific Software International.
Rosenberg, M. (1965). Society and the adolescent self-image.
Princeton, NJ: Princeton University Press.
Saslow, L. R., Muise, A., Impett, E. A., & Dubin, M. (2013). Can
you see how happy we are? Facebook images and relationship
satisfaction. Social Psychological & Personality Science, 4,
411-418.
Simpson, J. A. (1990). Influence of attachment styles on romantic
relationships. Journal of Personality and Social Psychology,
59, 971-980.
Simpson, J. A., Rholes, W. S., & Nelligan, J. S. (1992). Support
seeking and support giving within couples in an anxiety-pro-
voking situation: The role of attachment styles. Journal of
Personality and Social Psychology, 62, 434-446.
Slotter, E. B., & Gardner, W. L. (2012). How needing you changes
me: The influence of attachment anxiety on self-concept
malleability in romantic relationships. Self and Identity, 11,
386-408.
Snyder, M., & Gangestad, S. (1986). On the nature of self-moni-
toring: Matters of assessment, matters of validity. Journal of
Personality and Social Psychology, 51(1), 125-139.
Wegner, D. M., Lane, J. D., & Dimitri, S. (1994). The allure of secret
relationships. Journal of Personality and Social Psychology,
66, 287-300.
Wei, M., Russell, D. W., Mallinckrodt, B., & Vogel, D. L. (2007).
The Experiences in Close Relationship Scale (ECR)–Short
Form: Reliability, validity, and factor structure. Journal of
Personality Assessment, 88, 187-204.
Wilson, R. E., Gosling, S. D., & Graham, L. T. (2012). A review
of Facebook research in the social sciences. Perspectives on
Psychological Science, 7, 203-220.
by guest on September 18, 2014psp.sagepub.comDownloaded from