Computers in Human Behavior 29 (2013) 11421149

Contents lists available at SciVerse ScienceDirect

Computers in Human Behavior

journal homepage: www.elsevier.com/locate/comphumbeh

Attachment style, social skills, and Facebook use amongst adults

Julian A. Oldmeadow , Sally Quinn, Rachel Kowert
Department of Psychology, University of York, Heslington, York YO10 5DD, UK

a r t i c l e i n f o a b s t r a c t

Article history: Research suggests that online communication may benet those high in social anxiety. The current study
Available online 13 November 2012 examined Facebook use from the perspective of adult attachment theory, exploring relationships
between attachment anxiety and avoidance and Facebook use. Social skills have been found to be related
Keywords: to attachment style and internet use thus we also examined the role of these skills. A diverse sample of
Attachment adult participants (N = 617; 50.1% female) completed a self report questionnaire measuring attachment
Social skills dimensions, social skills, and Facebook use and experience. In line with predictions grounded in attach-
Social network
ment theory, we found that individuals with high attachment anxiety used Facebook more frequently,
Online
Facebook
were more likely to use it when feeling negative emotions and were more concerned about how others
perceived them on Facebook. High attachment avoidance was related to less Facebook use, less openness
and less positive attitudes towards Facebook. These relationships remained when social skills were con-
trolled. These results suggest that Facebook may serve attachment functions and provide a basis for
understanding how online communication may be related to attachment styles.
2012 Elsevier Ltd. All rights reserved.

1. Introduction
With the growth in popularity of social networking sites (SNSs)
such as Facebook, Myspace and Twitter, new forms of social
interaction have emerged that differ in important ways from the
ofine interactions more typically studied by social psychologists.
Over a decade ago McKenna and Bargh (2000) highlighted four
characteristics that differentiate online from ofine social interac-
tions: increased anonymity, reduced importance of physical
appearance, reduced barriers due to geographical distance, and
greater control over the time and pace of interactions. These
characteristics, McKenna and Bargh speculated, could make online
social interactions particularly appealing to certain types of people,
such as those suffering from social anxieties that handicap them in
ofine interactions. In the past few years considerable research has
investigated the psychological characteristics of internet users,
particularly SNS users, with a focus on the personality correlates
and psychological outcomes of internet use (Nadkarni & Hofmann,
2012). Conrming McKenna and Barghs suspicions, there is
growing evidence that social anxiety may play an important role
in the use of SNSs (Buote, Wood, & Pratt, 2009; Caplan, 2007).
In the current study, we explored the use of Facebook from the
perspective of adult attachment theory, examining whether there
exist relationships between attachment style and Facebook use.
A number of characteristic features of Facebook (and other
SNSs), including maintaining geographically distant relationships,
Corresponding author. Tel.: +44 1904 433190; fax: +44 1904 433181.
E-mail address: julian.oldmeadow@psychology.york.ac.uk (J.A. Oldmeadow).
maximizing control over the pace and time of interactions, and
enabling perpetual contact with a social network, suggest that
Facebook could potentially serve needs for belonging, closeness
and security for those with high levels of attachment anxiety and/
or avoidance. To explore this possibility, an online survey was
administered to a large and demographically diverse sample of
the general adult population, eliciting data on respondents attach-
ment styles, social skills, and use and experience with Facebook.
These data revealed reliable and theoretically consistent relation-
ships between attachment style and Facebook use that lend support
to the hypothesis that the use of Facebook is inuenced to some ex-
tent by attachment style, particularly attachment anxiety. This, in
turn, suggests that Facebook may provide a sense of closeness or
belonging that satises attachment needs in individuals who are
otherwise anxious or avoidant in close personal relationships.
1.1. Psychological correlates of Facebook use
Dozens of papers have been published on the social and psycho-
logical correlates of Facebook use and attitudes (Nadkarni &
Hofmann, 2012). Some demographic correlates have been
reported, including gender, ethnicity, and parental education
(Hargittai, 2008; Moore & McElroy, 2012), as well as cross cultural
differences (Buote et al., 2009; Caplan, 2007). The majority of re-
search, however, has focused on personality correlates of Facebook
use (Moore & McElroy, 2012; Ross et al., 2009; Ryan & Xenos, 2011;
Vasalou, Joinson, & Courvoisier, 2010). All studies report signicant
associations between personality traits and aspects of Facebook
use, supporting Amichai-Hamburger (2002) claim that personality

0747-5632/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.chb.2012.10.006
 J.A. Oldmeadow et al. / Computers in Human Behavior 29 (2013) 11421149
is a major factor in internet use. Although the results across studies
are not always consistent, taken together they implicate at least
extroversion, shyness and self-esteem. Extroversion associates
with more time on Facebook and larger friendship networks,
whereas shyness and low self-esteem associate with smaller
friendship networks but more intensive Facebook use (Nadkarni
& Hofmann, 2012; Orr et al., 2009).
These patterns suggest Facebook use may serve different func-
tions for different people, on the one hand satisfying extroverts
needs for social stimulation and large networks, and on the other
hand facilitating social interaction and a sense of belonging
amongst shy individuals and those with low self-esteem. Based
on a review of the literature, Nadkarni and Hofmann (2012) sug-
gest Facebook use is motivated by two fundamental needs, the
need for self-presentation and the need to belong. In support, they
cite evidence that a sense of disconnection motivates Facebook
use, and that Facebook use may increase a sense of connection
and self-esteem (Sheldon, Abad, & Hirsch, 2011; Yu, Tian, Vogel,
& Kwok, 2010). Thus, there is converging evidence to suggest that
social networking sites like Facebook may offer a form of social
connecting that is particularly appealing to certain types of indi-
viduals whose needs for belonging and connection with others
are not fully realized in ofine social interactions.
While Facebook may not offer the anonymity and invisibility
that characterized the chat rooms and newsgroups to which
McKenna and Bargh (2000) referred, it does offer social interaction
without physical proximity and enables greater control over the
time and pace of interactions. In addition, particularly with the
use of mobile devices, individuals can connect to Facebook virtu-
ally anytime, anyplace and anywhere, a characteristic we have re-
ferred to elsewhere as the Martini Effect (Quinn & Oldmeadow,
submitted for publication). These characteristics, more so than
anonymity or physical appearance, may make online social inter-
actions particularly appealing for individuals who are anxious
about personal closeness or physical intimacy, who anticipate
rejection from others, and who therefore tend to avoid developing
close personal relationships. These kinds of response sets are
described and explained by adult attachment theory (Hazan &
Shaver, 1987; Mikulincer & Shaver, 2007).
1.2. Adult attachment theory
Adult attachment theory stems from Bowlbys (Bowlby, 1969/
1982, 1973, 1980) work on childrens attachment to their primary
caregiver, extending the principles of child-parent attachment to
close personal relationships in adulthood (Hazan & Shaver, 1987;
Mikulincer & Shaver, 2007). Infants and childrens experiences with
attachment gures shape their working models about themselves,
others and relationships, which inuence cognitive, emotional,
and behavioral response patterns in attachment-relevant contexts.
These working models are characterized along two dimensions,
referred to as attachment anxiety and avoidance (Brennan, Clark,
& Shaver, 1998; Collins & Allard, 2004). When needs for comfort
and security are consistently satised the infant develops a secure
bond towards the attachment gure characterized by a positive
view of the self as lovable and of others as dependable. However,
if needs are not met by the attachment gure a working model
develops in which the individual fears and expects rejection from
attachment gures. Individuals with this type of model are high in
attachment anxiety, dened as an anxious or fearful preoccupation
with relationships (Smith, Murphy, & Coats, 1999). If the experience
of neglect is repeated consistently enough the individual may devel-
op a working model of others as untrustworthy and undependable,
and of the self as autonomous and self-sufcient. They may then
avoid developing close relationships, which they fear will only lead
to disappointment. Individuals with this type of working model are
1143
high in attachment avoidance, dened as a tendency to dismiss or
avoid intimate relationships (Smith et al., 1999).
Attachment theory was developed to explain childrens attach-
ment to primary care givers, but it has been successfully adapted to
explain adult relationships more generally, and romantic partner-
ships in particular (Hazan & Shaver, 1987; Mikulincer & Shaver,
2007). The working models developed in the context of primary
caregivers during childhood generalize to other kinds of social rela-
tionships during adulthood. Individuals high in attachment anxiety
have a hyper activated attachment system. They expect and fear
rejection from others, so are particularly vigilant in detecting cues
to untrustworthiness or rejection, and particularly sensitive to crit-
icism and conict. They may be particularly concerned with man-
aging their self-presentation, engage in behaviors designed to
reinforce relationships (e.g. gift giving), and seek regular conrma-
tion of close others responsiveness. Individuals high in attachment
avoidance have a deactivated attachment system. They down-reg-
ulate their needs for closeness and reinforcement from others and
avoid situations that activate the attachment system. They tend to
avoid developing intimate relationships and are more likely to be
socially isolated. The patterns of thinking, feeling and behaving
that characterize attachment anxiety and avoidance suggest,
rstly, that online social interactions may be particularly attractive
to individuals with attachment insecurities and, secondly, that
individuals high in attachment anxiety and avoidance may use
SNSs such as Facebook in different ways.
1.3. Attachment style and Facebook use
In the domain of interpersonal relationships Facebook may offer
advantages over ofine interactions for individuals high in attach-
ment anxiety and/or avoidance. For example, having the ability to
carefully manage self-presentation could be a particularly attrac-
tive feature of Facebook for individuals high in attachment anxiety,
who desire closeness but anticipate and fear rejection. By manag-
ing how they present themselves to others, through text, pictures,
links, etc., such individuals may feel more condent in their ability
to maintain interpersonal relationships. Similarly, the ability to
control the time and pace of interactions and to maintain relation-
ships without physical proximity could appeal to individuals high
in attachment avoidance, who tend to avoid closeness and value
their autonomy. Thus, online social interactions, and Facebook in
particular, may contain a number of features that facilitate the
development and maintenance of interpersonal relationships for
individuals high in attachment anxiety and/or avoidance. Previous
research has alluded to this possibility, as individuals with an anx-
ious or avoidant attachment report approaching and responding to
online and ofine friendships differently (Buote et al., 2009).
Another way in which Facebook may appeal to individuals with
attachment issues is by providing a sense of belonging to a social
network. Research shows that social network site use can increase
a sense of belonging amongst both adults (Sheldon et al., 2011) and
children (Quinn & Oldmeadow, in press). Facebook provides a
means of staying in touch with what others are doing, observing
relationships between others, and maintaining ones own presence
within the network (through wall posts, messages, pokes, etc.).
This may appeal to those whose attachment issues extend to social
relationships more generally, such as social groups. Very little re-
search has been done on attachment to social groups. However,
Smith et al. (1999) reasoned that individuals develop working
models of themselves in relation to groups that parallel those
underlying interpersonal attachment styles. They developed a
measure of group attachment by adapting an existing measure of
interpersonal attachment, and found moderate correlations be-
tween interpersonal and group attachment anxiety and avoidance.
These ndings suggest that attachment styles are not restricted to
1144 J.A. Oldmeadow et al. / Computers in Human Behavior 29 (2013) 11421149
interpersonal relationships, but extend to relationships to groups
as well.
A third way that attachment styles could inuence Facebook
use is indirectly through social skills. Attachment styles have been
shown to inuence a number of social skills, including the ability to
express emotions, empathy, the use of and understanding of body
language, and the ability to control ones social behavior (DiTomm-
aso, Brannen-McNulty, Ross, & Burgess, 2003). Although to our
knowledge there is no research linking social skills with social net-
work site use in particular, a number of studies show associations
between social skills and excessive internet use (Caplan, 2005;
Engelberg & Sjoberg, 2004; Iacovelli & Valenti, 2009; Liu & Peng,
2009). If attachment style inuences social skills, and individuals
with certain social skills are attracted to various forms of online
interaction, it is possible that attachment may inuence Facebook
use indirectly through social skills. If so, then Facebook use may
not specically cater to attachment needs and any relationships
between attachment style and Facebook use may generalize to
other forms of internet use.
1.4. The current study
In this study we examined relationships between attachment
styles, social skills and use and experience of Facebook. Based on
the view that Facebook provides a sense of closeness or belonging
that is attractive to individuals who are otherwise anxious in or
avoidant of close personal relationships, we hypothesized that
there would be direct relationships between attachment style,
measured as levels of attachment anxiety and avoidance, and the
use and experience of Facebook. Specically, we hypothesized that
(1) attachment anxiety would associate positively with intensity of
Facebook use, (2) that individuals high in attachment anxiety
would tend to use Facebook more when feeling negative emotions,
and also, (3) would express stronger concerns over how they might
be viewed by others on Facebook. Since individuals high in attach-
ment avoidance view themselves as autonomous and self-suf-
cient, it is possible that Facebook does not serve attachment
functions in the same way as for individuals high in attachment
anxiety. Therefore we hypothesized that, (4) attachment avoidance
would be uncorrelated with the intensity of Facebook use. Finally,
we expected that (5) social skills would associate with Facebook
use and experience, but that their impact would be largely inde-
pendent of the relationships between attachment style and Face-
book use and experience.
2. Method
2.1. Participants
Participants were recruited in several ways. Some were re-
cruited from within the psychology student population and re-
ceived course credit for their participation. Others were recruited
through posts on the authors social networking sites and received
no compensation. The remainder were recruited through Amazons
Mechanical Turk, a crowd sourcing website, where respondents
were paid $0.50 USD for taking the survey. A total of 742 complete
responses were received. One hundred twenty ve participants
who indicated they do not use Facebook were excluded from anal-
yses, leaving 617 participants (50.1% female). Age ranged from 18 to
64 years, with a mean of 27.63 (SD = 8.42). Sixty one percent of the
sample identied as Caucasian, 16.1 percent as Asian, 4.9 percent
Hispanic, 3.9 percent Indian/South Asian, 3.7 percent Black, and
5.9 percent as Other. Almost half (46.3%) were employed or self-
employed, and 39.7% were students (28.4% undergraduate; 10.9%
postgraduate). 6.5% of the total sample were psychology students.
2.2. Measures
2.2.1. Experiences in Close Relationships Scale (ECR; Brennan et al.,
1998)
The ECR is a 36-item scale that measures the two major dimen-
sions of attachment (anxiety and avoidance) in the context of close
adult relationships. An example item measuring attachment anxi-
ety is I need a lot of reassurance that close relationship partners
really care about me. An example of an item measuring attach-
ment avoidance is I get uncomfortable when someone wants to
be very close to me. Across hundreds of studies the scale has
shown excellent reliability and validity, and is the scale of choice
amongst leading adult attachment researchers (see Mikulincer &
Shaver, 2007). The main reason we chose this scale for the current
study is that it yields two separate dimension scores for each par-
ticipant, rather than assigning them to an attachment style cate-
gory (cf. Bartholomew & Horowitz, 1991). It is therefore an
appropriate measure for assessing linear relationships between
attachment dimensions, social skills, and Facebook use. In the cur-
rent study Cronbachs alphas for the anxiety and avoidance sub-
scales were .84 and .91 respectively, and the correlation between
the two was .246 (p < .001).
2.2.2. Social Skills Inventory (SSI; Riggio, 1986)
The standard SSI uses 90 items to measure six subcomponents
of social skills, three each in emotional and social domains: Emo-
tional Expressivity (EE), Emotional Sensitivity (ES), Emotional Con-
trol (EC), Social Expressivity (SE), Social Sensitivity (SS), and Social
Control (SC). The emotional and social components assess non-ver-
bal and verbal dimensions of social ability respectively. The
expressivity subscales focus on ones ability to communicate, both
verbally (SE) and non-verbally (EE); the sensitivity subscales assess
ones ability to interpret verbal (SS) and non-verbal (SE) communi-
cation; and the control subscales measure the ability to control and
regulate non-verbal (EC) and verbal (SC) displays (Riggio, 1989).
Because of the length of the SSI and the overall size of the online
questionnaire used in this study, we created an abridged version
of the SSI by choosing four items to measure each of the six sub-
scales (i.e. 24 items in total; see Appendix A). These items were se-
lected by identifying the four highest loading items from a
previous study in which we administered the full SSI to over 600
participants. Within the current sample a principle components
factor analysis with oblique rotation carried out on these 24 items
returned six factors with eigenvalues above 1, with the intended
four items loading separately on each factor. Cronbachs alphas
for the subscales were all above .8, except for emotional expressiv-
ity (alpha = .53). Correlations between subscales ranged from
.007 to .482. Thus, the four items used for each subcomponent
appeared to capture distinct and largely reliable aspects of social
skills, indicating they constituted an adequate measure of the in-
tended constructs. Factor scores were computed for each subscale
for use in subsequent analyses.
2.2.3. Facebook usage and experience
A range of questions were asked about participants usage and
experience with Facebook. These questions were devised by the
authors, drawing on previous research (Farahani, Aghamohamadi,
Kazemi, Bakhtiarvand, & Ansari, 2011; Hughes, Rowe, Batey, &
Lee, 2012; Moore & McElroy, 2012; Orr et al., 2009). Questions
focusing on Facebook usage included: How often do you use
Facebook? (1 = Less than once a month; 7 = daily); Think about
the last day you logged onto Facebook. In total, how long did you
spend logged onto Facebook? (1 = about 15 min or less; 7 = 3 h
or more); How often do you post updates on Facebook?
(1 = Never; 7 = 45 times a week); How often do you change your
Facebook prole picture? (1 = Never; 7 = Daily); and Who do you
 J.A. Oldmeadow et al. / Computers in Human Behavior 29 (2013) 11421149
normally include as friends on your Facebook prole? (1 = Pretty
much anybody; 7 = Close personal friends or family only). An addi-
tional 18 items probed a variety of dimensions of participants
experience with Facebook, including their dependence on Face-
book (e.g. I get anxious if I cannot check Facebook for a while),
which emotions motivate them to use Facebook (e.g. I nd myself
wanting to check Facebook particularly when I feel (lonely,
stressed, bored, sad, anxious, happy)), concern over social evalua-
tion through Facebook (e.g. I worry about the impression people
might get of me through Facebook), pride in Facebook (I am
proud of my Facebook prole), and indifference (e.g. If I dont
check Facebook for a while I tend to forget about it). Responses
were given on 7-point scales (1 = Strongly disagree; 7 = Strongly
agree).
2.3. Procedure
Participants completed an online survey that included the
above measures as well as an additional section on video game
play. The latter is not relevant to the current study so will not be
discussed further. The order in which the various sections of the
survey (ECR, SSI, Facebook usage and experience) were presented
was randomized across participants.
3. Results
Several studies have reported gender differences in attachment
styles and social skills (Deniz, Hamarta, & Ari, 2005; DiTommaso
et al., 2003). In the current study independent t-tests revealed gen-
der differences in avoidant attachment (p < .001), and on four of
the six subcomponents of social skills (ES, EC, EE and SC,
ps < .05). To control for these effects, gender was held as a covariate
in our analyses of relationships between attachment, social skills
and Facebook usage and experience (see DiTommaso et al. for a
similar approach). Consequently, hypotheses 1 through 4 were
tested using partial correlations (controlling for gender), and
hypothesis 5 was tested using a hierarchical regression (also con-
trolling for gender).
3.1. Attachment style and social skills
Table 1 shows the associations (partial correlations controlling
for gender) between anxious and avoidant attachment scores and
each subcomponent of social skills. Attachment anxiety was nega-
tively associated with SE, SC and EE. However, anxiety was most
strongly associated with ES and EC. Those high in attachment anx-
iety were more likely to be sensitive to criticism (ES) and more able
to control or hide their emotions (EC). Attachment avoidance was
also negatively associated with SE, SC and EE. Those high in attach-
ment avoidance reported being less likely to enjoy socializing and
meeting new people (SE), less likely to take a leading role in groups
(SC), and less likely to show their emotions (EE).
Table 1
Partial correlations (controlling for gender) between attachment style and social
skills.
Social skill Anxiety Avoidance
Social expressivity 0.119*
Social sensitivity 0.443**
Social control 0.096*
Emotional expressivity 0.209**
Emotional sensitivity 0.002
Emotional control 0.338**
**
p < .001.
*
p < .05.
1145
3.2. Attachment, social skills, and Facebook usage and experience
The items relating to Facebook usage (e.g. How often do you use
Facebook?, How often do you post updates on Facebook?) were
summed to give a single measure of Facebook usage (alpha = .621).
The 18 items relating to Facebook experience were factor analyzed to
extract the principle components. Principle components factor anal-
ysis with oblique rotation was used. Five factors emerged with eigen-
values greater than 1, explaining 65.3 percent of the variance (see
Appendix B). Items loading highly on the rst factor included those
related to using Facebook particularly when feeling negative emo-
tions (lonely, stressed, sad, anxious). We called this factor FB com-
fort seeking. Items loading on the second factor included being
happy for others (employer, parents) to view ones prole. We called
this factor FB openness. The third factor included items relating to
concern about how one is perceived by others on Facebook (e.g. I am
concerned about how I appear to others on Facebook). We called this
factor FB evaluation concern. The fourth factor included items
relating to detachment from Facebook, such as I think about remov-
ing my Facebook prole, and If I dont check Facebook for a while I
tend to forget about it. We reverse-scored this factor and called it
FB attachment. The nal factor included items expressing a gener-
ally positive orientation towards Facebook, such as I like my Face-
book prole and I nd myself wanting to check Facebook
particularly when I feel happy. We called this factor FB positivity.
Factor scores were computed for each of these ve factors.
Relationships between Facebook usage/experience and attach-
ment and social skills are given in Table 2. In terms of usage, a
positive association was found with attachment anxiety, with
higher attachment anxiety associating with more intensive use of
Facebook. Usage also associated positively with three of the six
social skills (SS, SE and SC). Participants higher in social sensitivity,
social expressivity and social control tended to use Facebook more
intensively than those lower in these social skills.
Signicant associations were found between attachment style
and each aspect of Facebook experience. Notably, attachment anx-
iety was strongly associated with FB comfort seeking, that is, using
Facebook particularly when feeling negative emotions. Attachment
anxiety was also associated with FB evaluation concern, that is,
worrying about how one appears to others on Facebook. Both com-
fort seeking and evaluation concern also were associated with
attachment avoidance, but less so than attachment anxiety. Attach-
ment avoidance was negatively associated with FB attachment, FB
openness and FB positivity. That is, participants high in attachment
avoidance were more likely to think about removing their Facebook
prole, less happy for signicant others to view their prole, and
less likely to be happy with and like their Facebook prole.
There were numerous signicant associations between aspects
of Facebook experience and social skills (see Table 2). There was a
strong association between social sensitivity (e.g. I am often con-
cerned with what others are thinking of me) and FB evaluation
concern (e.g. I am concerned about how I appear to others on Face-
book). Social sensitivity also was associated with comfort seeking
(i.e. using Facebook when feeling negative emotions) and attach-
ment to Facebook. These relationships also held (though weaker)
with emotional control (e.g. I am very good at maintaining a calm
exterior even if I am upset). Thus, those high in social sensitivity
and/or emotional control were more concerned with how they
appear to others on Facebook, and tended to use Facebook
0.385**
0.061 particularly when feeling sad, lonely, stressed or anxious. In
0.241** contrast, emotional expressivity (e.g. I rarely show my feelings
0.484**
0.045
1
0.008 Cronbachs alpha is likely to underestimate the reliability of ordinal scales
(Zumbo, Gadermann, & Zeisser, 2007). Given this, and the relatively small number of
items (5), we considered an alpha of .62 to be adequate. Nonparametric correlations
between items were all positive, signicant (<.05), and ranged from .083 to .447.
1146 J.A. Oldmeadow et al. / Computers in Human Behavior 29 (2013) 11421149

Table 2
Partial correlations (controlling for gender) between components of Facebook usage and experience, and attachment and social skills.

Usage Comfort seeking Evaluation concern Attachment Openness Positivity

Attachment
Anxiety 0.149** 0.401** 0.378** 0.087 0.047 0.066
Avoidance 0.008 0.151** 0.156** 0.083* 0.193** 0.141*
Social skills
EE 0.001 0.182** 0.096* 0.094* 0.076 0.062
ES 0.008 0.019 0.055 0.056 0.048 0.088*
EC 0.107 0.202** 0.157** 0.073 0.004 0.032
SE 0.157** 0.027 0.026 0.118* 0.123* 0.128*
SS 0.104* 0.223** 0.407** 0.136* 0.059 0.038
SC 0.084* 0.0001 0.052 0.028 0.169** 0.113*
**
p < .001.
*
p < .05.

Table 3
R2 and standardized beta weights for individual predictors at steps 2 and 3 in the model. Effects of gender were partialled out in step 1 and are not reported here.

Usage Comfort seeking Evaluation concern Attachment Openness Positivity

Step 2 Anxiety .151** .387** .369** .107* 0.009 0.085
Avoidance 0.044 0.06 0.066 .111* .186** .165**
Step 3 Anxiety .142* .343** .225** 0.065 0.030 .108*
Avoidance 0.046 0.068 .175** 0.013 0.154* .122*
EE 0.024 .097* 0.039 0.09 0.011 0.024
ES 0.048 -0.07 .114* 0.073 0.008 0.069
EC -0.04 -0.063 0.016 0.014 0.012 0.000
SE .179** .094* 0.034 0.119* 0.010 0.060
SS 0.009 0.037 .334 0.103* 0.062 0.035
SC 0.032 0.031 .111* 0.022 .131* 0.050
R2 change .034* .025* .103** .029* 0.019 0.014
Total R2 .061** .194** .262** .050** .054** .045*
**
p < .001.
*
p < .05.

or emotions [reversed]) was negatively associated with both com- 4. Discussion

fort seeking and evaluation concern. Those high in emotional
expressivity were less likely to use Facebook when feeling negative
emotions and less concerned with how they appear to others on
Facebook.
There were further associations between social skills and open-
ness, positivity, and attachment to Facebook. In particular, those
high in social expressivity (e.g. I love to socialize) were more at-
tached to their Facebook prole and more open and positive about
it than those low in social expressivity. Those high in social control
(e.g. I am often chosen to be the leader of a group) were also
more open and positive about their Facebook proles than those
lower in social control.
3.3. Regression analyses
Hierarchical regression analyses assessed the contributions of
attachment dimensions and social skills to explaining each aspect
of Facebook usage and experience. In each model gender (dummy
coded) was entered rst to partial out gender effects. In the second
step, attachment anxiety and avoidance were entered. The six so-
cial skills subscales were entered in step 3. The total amount of
variance explained in the nal model and beta weights of each
individual predictor are shown in Table 3. In each case the attach-
ment dimensions explained a signicant proportion of variance
over and above gender (step 2). With the exception of FB openness
and FB positivity, adding the social skills subscales in step 3 signif-
icantly improved prediction of each outcome. Nevertheless, attach-
ment anxiety remained a signicant predictor of FB usage, comfort
seeking, evaluation concern and positivity independent of social
skills (step 3). Attachment avoidance remained a signicant pre-
dictor of evaluation concern, openness and positivity independent
of social skills.
The aim of the present study was to explore any links between
attachment style and Facebook use and experience. In line with our
rst three hypotheses, we found that individuals with a high
attachment anxiety were more likely to spend more time on the
site, to use it when they were feeling negative emotions and to
show concern over how other Facebook users perceived them.
We also found no relationship between the avoidant attachment
style and Facebook usage, supporting our fourth hypothesis.
Although social skills were found to have some relationship with
Facebook use and experience, they were found to have little effect
on the relationships between the two attachment styles and Face-
book use and experience, supporting our nal hypothesis. This sug-
gests that the way people experience and use Facebook may have a
direct relationship with their style of attachment.
Securely attached individuals have been found to be high in so-
cial and emotional expressivity, high in social control, and low in
social sensitivity (Deniz et al., 2005; Dereli & Karakus, 2011). In
the current study social expressivity and social control both were
associated with more openness and positivity towards Facebook.
In addition, while those high in social expressivity and social con-
trol tended to use Facebook more than those low in social expres-
sivity and social control, there were no associations between
these traits and comfort seeking or evaluation concern. This pattern
gives rise to an image of a securely attached individual, outgoing,
sociable and expressive, who uses Facebook in a positive way. They
are happy for important others to view their Facebook prole, they
are proud of their Facebook prole and enjoy using Facebook when
feeling happy. Since Nadkarni and Hofmann (2012) suggest that
Facebook may help to increase a sense of connection to others,
our data suggest that securely attached individuals may use Face-
book as a tool to augment existing relationships in the ofine world.
 J.A. Oldmeadow et al. / Computers in Human Behavior 29 (2013) 11421149
Individuals high in attachment anxiety are typically character-
ized by anxiousness about being alone, being sensitive to criticism
received from others and having a need for reassurance (Brennan
et al., 1998). In the present study, we found that attachment anx-
iety was positively related to Facebook usage, comfort seeking
and to evaluation concern. This would suggest that Facebook is
being used more frequently by those with an anxious attachment
style as a way to alleviate their feeling of being alone. In addition,
their motivation for logging on and using Facebook may be moti-
vated by their mood that is, they may seek to improve their neg-
ative mood by interacting with others. Social networking sites can
offer users instant access to social interaction with a myriad of
people (Urista, Dong, & Day, 2009) and it is perhaps this feature
that is particularly attractive to users high in attachment anxiety.
This instant access offers them interaction with others that may
be needed, particularly when their mood is negative. In a qualita-
tive study, Urista et al. (2009) found that receiving comments on
a photograph, for example, can result in the user feeling as though
they are popular. An anxious Facebook user may log onto their pro-
le when they are in a negative mood to check if any of their
friends have commented on their content, which may improve
their mood by making them feel popular to some degree. What re-
mains to be seen however is whether or not this Facebook usage
results in a changed mood and if so, to what degree. It has been
suggested elsewhere that the link between user experience and
benets gained from Facebook usage may relate to the characteris-
tics of the user and how they use the site (DiTommaso et al., 2003).
Future studies should consider examining the relationships be-
tween Facebook use and mood change taking into account each
attachment style. Since those high in attachment anxiety are sen-
sitive to criticism and also require frequent reassurance, it could
be hypothesized that for those high in attachment anxiety, Face-
book interactions would only improve mood if the interactions
are constantly positive and continually boost self esteem.
We found that Facebook users who were high in attachment
anxiety were also more likely to be concerned about how other
Facebook users perceive them. Since those high in attachment anx-
iety are sensitive to criticism, social interactions via Facebook may
be particularly attractive to these individuals. The online environ-
ment can provide users with greater control over interactions
(McKenna & Bargh, 2000) and Facebook may therefore allow these
individuals to have greater control over their self-presentation. For
example, they can take their time over what to post on their status
update, what photos to put on their own prole and how to com-
ment on others photos and posts. This may also go some way in
explaining why we found that high attachment anxiety was related
to higher usage of Facebook. For example, upon posting a photo-
graph, the anxious attachment Facebook user may log on regularly
to check if anyone has commented on their photo - being con-
cerned about how others may have perceived them in this photo-
graph. In addition, they may spend more time on the site
perfecting their Facebook self-portrayal, in an attempt to ensure
that others would view them in a positive light.
The present study also provided evidence that regardless of so-
cial skills, individuals high in attachment avoidance used Facebook
less often, were less likely to be open about their Facebook prole
and less likely to hold a positive attitude about Facebook. These
ndings would seem to t with the characteristics associated with
this type of attachment that is, someone who is autonomous,
reluctant to share emotions and sees others as untrustworthy.
However, we also found a signicant positive relationship between
high attachment avoidance and evaluation concern. This would
suggest that although those high in attachment avoidance are typ-
ically autonomous, they do have some level of concern over how
others view them, at least on Facebook. Since we found no relation-
ship between this style of attachment and social sensitivity, it sug-
1147
gests that for the present sample those high in attachment
avoidance generally do not have concerns over how others perceive
them. There may therefore be some unique features of Facebook
that, for those high in attachment avoidance, result in being con-
cerned about what others think of them. It must, however, be con-
sidered that our study focused only on Facebook users, who may on
the whole be relatively low in attachment avoidance. The charac-
teristics typically associated with high attachment avoidance may
therefore be slightly weakened within the present sample. This
might go some way in explaining why there seems to be a relation-
ship between attachment avoidance and evaluation concern.
We included measures of social skills because previous research
shows relationships between social skills and attachment style
(Deniz et al., 2005; Dereli & Karakus, 2011; DiTommaso et al.,
2003) and between social skills and internet use (Caplan, 2005;
Engelberg & Sjoberg, 2004; Harman, Hansen, Cochran, & Lindsey,
2005). We therefore wanted to examine whether attachment style
would predict Facebook use over and above social skills. Indeed,
there were signicant relationships between social skills and
attachment style, and between social skills and aspects of Facebook
use and experience. Social expressivity was positively related to
Facebook usage, indicating those who are more outgoing tend to
use Facebook more intensively, and there was a strong association
between emotional sensitivity and FB evaluation concern. Those
who tend to be emotionally sensitive generally tend also to be con-
cerned with how others view them on Facebook. Importantly,
though, these associations between social skills and Facebook use
did not account for the relationships between attachment style
and Facebook use. The relationships of attachment anxiety and
avoidance to Facebook use and experience were largely unaffected
by the inclusion of social skills in the regression models. This sug-
gests that attachment style inuences Facebook use and experi-
ence largely independently of social skills, and supports the view
that Facebook may serve attachment functions.
Social networking sites such as Facebook can be many things to
many people. The current study provides initial evidence that Face-
book may serve attachment functions for people high in attach-
ment anxiety, and to some extent avoidance. Although the
relationships observed between attachment and Facebook use
and experience were largely independent of social skills, this does
not rule out a role for other individual difference variables. Further
research should aim to partial out the effects of attachment style
from other variables such as loneliness, extroversion, and more
general social anxiety to further clarify the unique role that attach-
ment plays in SNS use.
The advantage of an attachment theory approach to SNS use is
that it offers a more theoretically grounded framework for under-
standing how and why people use SNSs than simply identifying
associations between personality or other traits and SNS use. An
attachment theory approach invites one to ask which components
of the attachment behavioral system are implicated in SNS use, and
what kinds of SNS use serve attachment functions. As already sug-
gested, further research could examine whether using SNSs actu-
ally leads to positive outcomes like improved mood for
individuals with different attachment styles. Another avenue
would be to examine more closely the effects of activating the
attachment system on SNS use. Our data show that individuals
high in attachment anxiety reported being motivated to use Face-
book particularly when feeling negative emotions such as loneli-
ness. However, a more experimental approach that seeks to
activate the attachment system and observe subsequent SNS activ-
ity would help to clarify whether SNS use is motivated by attach-
ment needs and not just negative mood per se.
A word of caution is warranted when generalizing from Face-
book use to other SNSs. Facebook is currently the most popular
SNS (Kazeniac, 2009) and 83 percent of our survey respondents
1148 J.A. Oldmeadow et al. / Computers in Human Behavior 29 (2013) 11421149

indicated they use Facebook. However, there are other SNSs and
we cannot be sure that the relationships we observed generalize
to other sites. In theory these relationships should generalize to
other sites that also offer the features of enhanced control, self-pre-
sentation and perpetual contact that we theorize make Facebook
attractive to individuals with high attachment anxiety. However,
they may not generalize to other sites that serve more professional
than purely social functions, such as Linkedin or Academia. Further
research would be needed to precisely dene which types of inter-
net sites and/or usage are related to attachment styles.
5. Conclusion
To conclude, the current study provides evidence of theoreti-
cally consistent relationships between dimensions of adult attach-
ment and Facebook use and experience that are largely
independent of social skills. These ndings support the view that
Facebook (and possibly other SNSs) can serve attachment functions
and may be particularly attractive for individuals high in attach-
ment anxiety. Further research should investigate which attach-
ment functions SNS use serves, or which parts of the attachment
behavioral system SNS use responds to. The current study provides
a promising rst step towards understanding SNS use from the
theoretically rich perspective of adult attachment theory.
Appendix A
 I always seem to know what peoples true feelings are no matter
how hard they try to conceal them.
 I can accurately tell what a persons character is upon rst meet-
ing him or her.
 I can instantly spot a phony the minute I meet him or her.
Emotional Control (EC):
 I am not very skilled in controlling my emotions ().
 It is very hard for me to control my emotions ().
 I am very good at maintaining a calm exterior even if I am upset.
 I am rarely able to hide a strong emotion ().
Social Expressivity (SE):
 I love to socialize.
 I always mingle at parties.
 At parties I enjoy talking to a lot of different people.
 I enjoy going to large parties and meeting new people.
Social Sensitivity (SS):
 I am very sensitive of criticism.
 It is very important that other people like me.
 I am generally concerned about the impression I am making on
others.
 I am often concerned what others are thinking of me.

Abridged Social Skills Inventory listing the items used to assess
each component of social skills. Responses were given on 5-point
scales ranging from 1 (not at all like me) to 5 (exactly like me).
Emotional Expressivity (EE):
 I usually feel uncomfortable touching other people ().
 Sometimes I have trouble making my friends and family realize
how angry or upset I am with them ().
 I often touch my friends when talking to them.
 I rarely show my feelings or emotions ().
Emotional Sensitivity (ES):
 I can easily tell what a persons character is by watching his or
her interactions with others.
Social Control (SC):
 When I am with a group of friends I am often the spokesperson
for the group.
 I nd it very difcult to speak in front of a large group of people ().
 I am usually very good at leading group discussions.
 I am often chosen to be the leader of a group.
() Represents a reverse scored item.
Appendix B
Structure matrix for the principle components factor analysis
on Facebook experience items. The highest loading items in each
factor are in bold.

Items Factors
I get anxious if I cannot check Facebook for a while 0.564 0.001 0.410 0.647 0.149
If I dont check Facebook for a while I tend to forget about it 0.268 0.030 -0.143 0.788 0.242
The longer I am unable to check Facebook the more I want to check it 0.528 0.008 0.402 0.722 0.188
I nd myself wanting to check Facebook particularly when I feel lonely 0.757 0.128 0.282 0.165 0.308
I nd myself wanting to check Facebook particularly when I feel stressed 0.823 0.023 0.342 0.266 0.214
I nd myself wanting to check Facebook particularly when I feel bored 0.328 0.191 0.085 0.073 0.628
I nd myself wanting to check Facebook particularly when I feel sad 0.828 0.022 0.343 0.198 0.171
I nd myself wanting to check Facebook particularly when I feel anxious 0.838 0.049 0.376 0.223 0.203
I nd myself wanting to check Facebook particularly when I feel happy 0.414 0.103 0.213 0.293 0.482
I think about removing my Facebook prole 0.147 0.266 0.058 0.600 0.185
I worry about what people might say about me on Facebook 0.400 0.034 0.774 0.081 0.044
I worry about the impression people might get of me through Facebook 0.299 0.099 0.895 0.076 0.049
I am concerned about how I appear to others on Facebook 0.310 0.005 0.914 0.162 0.126
How I appear to others on Facebook is important to me 0.292 0.093 0.824 0.261 0.270
I am proud of my Facebook prole 0.116 0.417 0.233 0.277 0.738
I like my Facebook prole 0.076 0.428 0.099 0.385 0.770
I would be happy for a potential employer to view my Facebook prole 0.045 0.814 0.018 0.144 0.142
I would be happy for my parents to view my Facebook prole 0.019 0.785 0.017 0.068 0.187
 J.A. Oldmeadow et al. / Computers in Human Behavior 29 (2013) 11421149 1149

References Iacovelli, A., & Valenti, S. (2009). Internet addictions effect on likeability and
rapport. Computers in Human Behavior, 25(2), 439443.
Kazeniac, A. (2009). Social networks: Facebook takes over top spot, Twitter climbs.
Amichai-Hamburger, Y. (2002). Internet and personality. Computers in Human
Compete Pulse, http://blog.compete.com/2009/02/09/facebook-myspace-
Behavior, 18(1), 110.
twitter-social-network/, (Retrieved 23.06.2012).
Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among yound adults:
Liu, M., & Peng, W. (2009). Cognitive and psychological predictors of the negative
A test of a four category model. Journal of Personality and Social Psychology, 61,
outcomes associated with playing MMOGs (massively multiplayer online
226244.
games). Computers in Human Behavior, 25(6), 13061311.
Bowlby, J. (1969/1982). Attachment and loss: Vol. 1. Attachment. NY: Basic Books.
McKenna, K. Y. A., & Bargh, J. A. (2000). Plan 9 from cyberspace. The implications of
Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation: Anxiety and anger. NY:
the internet for personality and social psychology. Personality and Social
Basic Books.
Psychology Review, 5(1), 5775.
Bowlby, J. (1980). Attachment and loss: Vol. 3. Sadness and depression. NY: Basic
Mikulincer, M., & Shaver, P. R. (2007). Attachment in adulthood: Structure, dynamics
Books.
and change. NY: The Guildford Press.
Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult
Moore, K., & McElroy, J. C. (2012). The inuence of personality on Facebook usage,
romantic attachment: An integrative overview. In J. A. Jimpson & W. S. Rholes
wall postings, and regret. Computers in Human Behavior, 28(267274).
(Eds.), Attachment theory and close relationships (pp. 4676). NY: The Guildford
Nadkarni, A., & Hofmann, S. G. (2012). Why do people use Facebook? Personality and
Press.
Individual Differences, 52(3), 243249.
Buote, V., Wood, E., & Pratt, M. (2009). Exploring similarities and differences
Orr, E. S., Sisic, M., Ross, C., Simmering, M. G., Arseneault, J. M., & Orr, R. R. (2009).
between online and ofine friendships: The role of attachment style. Computers
The inuence of shyness on the use of Facebook in an undergraduate sample.
in Human Behavior, 25, 560567.
CyberPsychology & Behavior, 12(3), 337340.
Caplan, S. E. (2005). A social skill account of problematic internet use. Journal of
Quinn, S., & Oldmeadow, J. A. (submitted for publication). The Martini Effect and
Communication, 55(4), 721736.
social networking sites.
Caplan, S. E. (2007). Relations among loneliness, social anxiety, and problematic
Quinn, S., & Oldmeadow, J. A. (in press). Is the i-generation a we generation?
Internet use. CyberPsychology & Behavior, 234242.
Children, social networking sites and belonging. British Journal of Developmental
Collins, N. L., & Allard, L. M. (2004). Cognitive representations of attachment: The
Psychology.
content and function of working models. In G. J. O. Fletcher & M. S. Clark (Eds.),
Riggio, R. E. (1986). Assessment of basic social skills. Journal of Personality and Social
Blackwell handbook of social psychology: Interpersonal processes (pp. 6085). MA:
Psychology, 51, 649660.
Blackwell.
Riggio, R. E. (1989). Manual for the social skills inventory. Palo Alto: Consulting
Deniz, M. E., Hamarta, E., & Ari, R. (2005). An investigation of social skills and
Psychologists Press.
loneliness levels of university students with respect to their attachment styles
Ross, C., Orr, E. S., Sisic, M., Arseneault, J. M., Simmering, M. G., & Orr, R. R. (2009).
in a sample of turkish students. Social Behavior and Personality, 33(1), 1932.
Personality and motivations associated with Facebook use. Computers in Human
Dereli, E., & Karakus, O. (2011). An examination of attachment styles and social
Behavior, 25, 578586.
skills of sniversity students. Electronic Journal of Research in Educational
Ryan, T., & Xenos, S. (2011). Who uses Facebook? An investigation into the
Psychology, 9(2), 16922095.
relationship between the big ve, shyness, narcissism, loneliness, and Facebook
DiTommaso, E., Brannen-McNulty, C., Ross, L., & Burgess, M. (2003). Attachment
usage. Computers in Human Behavior, 27, 16581664.
styles, social skills and loneliness in young adults. Personality and Individual
Sheldon, K. M., Abad, N., & Hirsch, C. (2011). A two-process view of Facebook use
Differences, 35, 303312.
and relatedness need-satisfaction: Disconnection drives use, and connection
Engelberg, E., & Sjoberg, L. (2004). Internet use, social skills, and adjustment.
rewards it. Journal of Personality and Social Psychology, 100, 766775.
CyberPsychology & Behavior, 7(1), 4147.
Smith, E. R., Murphy, J., & Coats, S. (1999). Attachment to groups: Theory and
Farahani, H. A., Aghamohamadi, S., Kazemi, Z., Bakhtiarvand, F., & Ansari, M.
measurement. Journal of Personality and Social Psychology, 77(1), 94110.
(2011). Examining the relationship between sensitivity to rejection and using
Urista, M. A., Dong, Q., & Day, K. D. (2009). Explaining why young adults use
Facebook in university students. Procedia Social and Behavioral Sciences, 28,
MySpace and Facebook through uses and gratications theory. Human
807810.
Communication, 12(2), 215230.
Hargittai, E. (2008). Whose space? Differences among users and non-users of social
Vasalou, A., Joinson, A. N., & Courvoisier, D. (2010). Cultural differences, experience
network sites. Journal of Computer Mediated Communication, 13, 267297.
with social networks and the nature of true commitment in Facebook.
Harman, J. P., Hansen, C. E., Cochran, M. E., & Lindsey, C. R. (2005). Liar, liar: Internet
International Journal of Human-Computer Studies, 68(10), 719728.
faking but not frequency of use affects social skills, self = esteem, social anxiety,
Yu, A. Y., Tian, S. W., Vogel, D., & Kwok, R. C.-W. (2010). Can learning be virtually
and aggression. CyberPsychology & Behavior, 8(1), 16.
boosted? An investigation of online social networking impacts. Computers and
Hazan, C., & Shaver, P. R. (1987). Romantic love conceptualized as an attachment
Education, 55, 14941503.
process. Journal of Personality and Social Psychology, 52, 511524.
Zumbo, B. D., Gadermann, A. M., & Zeisser, C. (2007). Ordinal versions of coefcients
Hughes, D. J., Rowe, M., Batey, M., & Lee, A. (2012). A tale of two sites: Twitter vs.
alpha and theta for likert rating scales. Journal of Modern Applied Statistical
Facebook and the personality predictors of social media usage. Computers in
Methods, 6, 2129.
Human Behavior, 28, 561569.