See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/265102817

Privacy in Interaction: Exploring Disclosure and Social Capital in Facebook

Article in Proceedings of the International AAAI Conference on Web and Social Media · May 2012
DOI: 10.1609/icwsm.v6i1.14268

CITATIONS READS
79 1,272

5 authors, including:

Jessica Vitak Nicole B Ellison

University of Maryland, College Park University of Michigan
93 PUBLICATIONS 8,116 CITATIONS 90 PUBLICATIONS 45,273 CITATIONS

SEE PROFILE SEE PROFILE

Rebecca Gray
Michigan State University
14 PUBLICATIONS 2,113 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

PERVADE: Pervasive Data Ethics for Computational Research View project

Inclusive Public Transit Toolkit to Assess Quality of Service Across Socioeconomic Status in Baltimore City View project

All content following this page was uploaded by Nicole B Ellison on 23 October 2014.

The user has requested enhancement of the downloaded file.

 Proceedings of the Sixth International AAAI Conference on Weblogs and Social Media

Privacy in Interaction: Exploring Disclosure

and Social Capital in Facebook
Fred Stutzman, Jessica Vitak, Nicole B. Ellison, Rebecca Gray, Cliff Lampe
H. John Heinz III College Dept. of Telecommunication, IS, & Media School of Information
Carnegie Mellon University Michigan State University University of Michigan
Pittsburgh, PA East Lansing, MI Ann Arbor, MI
[email protected] {vitakjes, nellison, grayreb2}@msu edu [email protected]

disclosures to one’s Facebook network, due to network

Abstract diversity or other reasons, may negatively impact the
extent to which users may access social capital resources.
In this paper, we explore the relationship between Facebook For instance, a user who chooses not to disclose a medical
users’ privacy concerns, relationship maintenance strategies,
diagnosis via the site will be less likely to receive
and social capital outcomes. Previous research has found a
positive relationship between various measures of Facebook supportive messages from her Facebook network.
use and perceptions of social capital, i.e., one’s access to This study extends previous work by analyzing the
social and information-based resources. Other research has relationship between Facebook users’ privacy attitudes and
found that social network site users with high privacy behaviors, their site engagement, and their perceptions of
concerns modify their disclosures on the site. However, no bonding and bridging social capital. This is done through
research to date has empirically tested how privacy two sets of analyses: first, we build and test a path model
concerns and disclosure strategies interact to influence to empirically test previous theoretical work (Ellison,
social capital outcomes. To address this gap in the literature, Vitak, Gray, Steinfield, and Lampe, 2011c) which argues
we explored these questions with survey data (N=230). that disclosures are a necessary requirement for accruing
Findings indicate that privacy concerns and behaviors
social capital on SNSs but that this relationship may be
predict disclosures on Facebook, but not perceptions of
social capital. In addition, when looking at predictors of mediated by users’ willingness to disclose on the site. We
social capital, we identify interaction effects between users’ then incorporate variables that assess the role that users’
network composition and their use of privacy features. engagement with their networks and network composition
play in predicting social capital, while accounting for
privacy considerations.
Introduction
Literature Review
Social network sites (SNSs) enable users to connect and
interact with proximate and distant ties, and the
Previous studies examining uses of and outcomes derived
communication features of these sites lower barriers for
from SNSs have examined social capital, privacy, and
requesting and providing support. Resources such as
information or social support are often framed as Facebook communication practices, but have done so in
separate analyses. This study contributes to the literature
instantiations of social capital, and researchers have
by integrating two important threads within human-
identified positive relationships between perceptions of
computer interaction (HCI) scholarship on SNS use:
social capital and various measures of Facebook use,
privacy and social capital outcomes.
including users’ network composition and specific forms
of engagement with that network (e.g., Burke, Kraut, and
Marlow 2011; Ellison, Steinfield, and Lampe 2011a). Facebook and Social Capital
However, privacy attitudes and behaviors play a critical Social capital is the total actual or potential resources
role in whether individuals choose to engage with and individuals have access to through their social network
share content within a network (Lampinen, Tamminen, and (Bourdieu 1985). Social capital includes physical (e.g.,
Oulasvirta 2009). Privacy concerns may have a direct driving a friend to the airport), emotional (e.g., giving a
impact on whether users exchange information and friend a hug), and informational (e.g., giving a friend
resources with their network. For example, Hogan (2010) advice about a big decision) resources, among others.
argues that, given the increasing diversity of users’ online Social capital can be understood as an investment in one’s
networks, some SNS users may only share content personal network (Lin 2001) with expected returns at some
appropriate for all their connections. However, limiting future point; in other words, reciprocity is a key component
of social capital. Resource requests and offers occur
Copyright © 2012, Association for the Advancement of Artificial through both offline and online channels and may vary
Intelligence (www.aaai.org). All rights reserved. based on the nature of the relationship. Social capital

330
benefits may be mobilized through simple membership in a potentially increasing the visibility of each in one another’s
network, but as Burt (1992) notes, one’s network position News Feed (which users may or may not be aware of).
can result in greater or less access to resources. For Finally, thinking about the generalized reciprocity that
example, individuals who bridge two otherwise marks social capital exchange (Lin 2001), these “giving”
unconnected networks may control information diffusion behaviors should increase expectations of receiving
between those networks, a potentially powerful position. support provisions in the future. In other words, responding
Network composition is often related to the kind of to Facebook Friends’ requests should elevate the likelihood
resources embedded in one’s social network. In small, of one’s own requests being answered in the future.
densely connected networks, such as that of a tight-knit Therefore, we believe that responding to others, especially
family, there is often a high level of access to social and if is done in a public channel, may influence users’
emotional support. Putnam (2000) refers to these social perceptions of their access to resources such as bridging
support resources as bonding social capital. That said, the and bonding social capital.
homogeneity associated with close ties makes them less
likely to provide new ideas or information. Therefore, Privacy and Disclosure on SNS
larger, more loosely connected networks with many
In “nonymous” (e.g., Zhao et al. 2008) online spaces such
clusters and cross-cutting ties are more likely to provide
access to diverse perspectives and non-redundant as SNSs, where personal identity is highly prominent,
privacy is a critical component in determining how and
information. These types of networks are typically
with whom users interact. In highly contextual spaces such
associated with bridging social capital (Burt 1992).
as SNSs, privacy should be considered as a fluid process
Research has documented a positive link between
where individuals selectively control access to information
individuals’ use of Facebook and their perceptions of
about themselves by regulating their social interactions
social capital (Burke, Marlow, and Lento 2010; Burke et
al. 2011; Ellison et al. 2011a). The site structure enables (Altman 1975). On SNSs, this regulation can occur in a
number of ways. For example, users may designate their
the creation and maintenance of “social supernets” (Donath
profiles as “Friends only,” which limits access to their
2007)—large-scale networks enabled by technology—and
profile to only those users with whom they have formally
provides users with numerous public and private
connected. Most SNSs also enable tailoring of content
communication channels through which to request and
distribution so that individual updates or photos can be
offer resources to network members. As a whole, the
literature suggests that social capital is not only a function shared with a subset of one’s total network. Preliminary
research examining this strategy has found that users
of users’ network composition (i.e., Facebook “Friends”),
employing segmented privacy settings report larger
but also relates to their specific engagement practices on
Facebook networks and higher perceived bridging and
the site (i.e., how they interact with these “Friends”).
bonding social capital than those who do not use this
feature (Ellison et al., 2011c).
Relationship Investment on Facebook Some of the earliest work on privacy and SNSs
Whereas the literature suggests a relationship between identified a disconnect between users’ privacy concerns
several forms of Facebook-based engagement and and their disclosures on the site (Acquisti and Gross 2006;
perceptions of social capital, the mechanisms through Speikermann et al. 2001), in what some have labeled a
which Facebook use and social capital are associated have “privacy paradox” (Barnes 2006). More recent research
not been explicated. Burke et al. (2011) examine the role of suggests that privacy and disclosures are more closely
directed communication, a scale that includes six metrics related, with complicated tradeoffs between privacy,
such as the number of wall posts, comments, and “likes” intention, and disclosure (Stutzman and Kramer-Duffield
received from Facebook Friends. They found that directed 2010; Krasnova et al. 2010). When considering Facebook
communication (to a specific person) was associated with specifically, evolution in the relationship between privacy
greater social capital benefits than broadcasting updates to concerns and disclosures may be a result of changes in the
one’s entire network. site’s structure during the last six years, and the increasing
Recent work (Ellison, Vitak, Gray, Lampe, and Brooks, diversity of users’ networks as a result. Network
2011b) extends Burke et al.’s (2011) and Ellison et al.’s diversification can lead to context collapse, whereby
(2011a) findings with the development of a measure that various clusters of individuals with whom one has a
captures the extent to which individuals signal investment relationship (e.g., high school friends, coworkers, family)
in relationships with their friends by way of responding to are grouped together under a single label (e.g., Facebook
Facebook-mediated requests for social or informational Friends; Marwick and boyd 2011). One strategy for
support and acknowledging meaningful events in their managing context collapse is for users to actively distribute
lives (e.g., a friend’s birthday). These responses can be content to specific subsets of their network, a challenging
seen as a form of “social grooming”—signals of attention proposition given that most SNSs are oriented toward
that promote feelings of trust and closeness (Dunbar 1996). broadcasting to the entirety of one’s network.
Responding to a question from a Friend serves a social When considering the relationship between privacy and
grooming function as well as a technical one. Interactions disclosures in light of context collapse, users adopt a range
between two users help train the Facebook algorithm, of strategies for minimizing risk. Lampinen et al. (2009,

331
2011) describe both behavioral and mental strategies for social capital in context.
group context management. Users engaging in behavioral
strategies may actively employ site features to control The Study
access to disclosures by limiting their network size,
limiting access to specific parts of their profile, or creating In this research we draw on survey data to analyze the
friend lists to distribute content to subsets of their friend relationship between SNS privacy, disclosure practices,
network—while keeping these disclosures hidden from network composition and engagement, and social capital
others. Conversely, individuals that employ mental outcomes. In doing so, we provide new insight into the
strategies may choose to limit disclosures to only content complex relationships among attitudes and behaviors that
they deem appropriate for all network members, in a have the potential to either constrain or increase the social
process known as the lowest-common-denominator capital benefits of SNS use.
approach (Hogan 2010).
Method
Disclosure, Privacy, and Social Capital on FB
In April 2011, a random sample of 2,500 undergraduate
Research examining the relationship between Facebook students at a large, Midwestern university were invited, via
use and social capital has generally ignored the role that email, to participate in an online survey about their use of
privacy plays in users’ decision-making process regarding online communication tools. As incentive for participation,
content shared through the site. As has been argued all participants were invited to enter their email address for
previously (Ellison et al., 2011c), if one considers social a raffle of ten $15 Amazon gift cards. The total number of
capital to be the resources obtained through interactions usable responses amounted to 230 for a response rate of
with one’s social network, Facebook users must be willing 9.2% following AAPOR definition one (AAPOR, 2008).
to make these resource requests—i.e., disclose—in order This response rate is consistent with other recent studies
for their networks to respond appropriately. Clearly, employing online surveys of college students (e.g., Yoder
privacy concerns may serve as a barrier to some and Stutzman 2011). Respondents were 67% female and
disclosures, especially if the resource request is more 33% male, had an average age of 21.16 (SD = 4.37), and
personal in nature (such as a person requesting emotional 95% used Facebook. Compared to the population of
support following the death of a family member), and students, women are overrepresented in this sample.
therefore the effects of privacy attitudes and behaviors on
disclosures in a SNS are important to investigate.
Recent work has demonstrated evidence of the
Survey Content
relationship between privacy and social capital in the SNS Our survey was comprised of scales for bridging and
context, finding that use of segmented privacy settings on bonding social capital, a variable called Signals of
Facebook—such as limiting access to specific updates or to Relational Investment (SRI), privacy concerns, various
one’s profile more generally—is positively correlated with measures of Facebook engagement and use, privacy
perceptions of social capital (Ellison et al., 2011c). This settings, and demographics. Unless otherwise noted, scale
finding may reflect the fact that the ability to partition items were measured on a five-point Likert type scale (1 =
one’s online network and distribute content only to a Strongly Disagree, 5 = Strongly Agree).
specific audience makes users more comfortable disclosing Bridging social capital (10 items, α = .877, M = 3.87,
certain kinds of information to their network. In the design SD = .60), adapted from previous research (Williams
community, there are a number of active research streams 2006), indicates perceptions of bridging resources,
exploring the most effective ways to manage contexts measuring the extent to which participants feel they
within SNSs, with the goal of producing rule sets or interact with a diverse set of people, engage in diffuse
interfaces that actively foster the sharing of content to reciprocity, and have a view of themselves as a member of
intended, trusted audiences (e.g., Farnham and Churchill a broader group. For this study, the 10 items were prefaced
2011; Kelley et al. 2011; Ozenc and Farnham 2011) by the following instructions: “For the next series of
The design of features that manage the complexity of questions, think about your entire social network, including
heterogeneous networks may encourage users to maintain a relatives, close and distant friends, coworkers and
larger network of connections, which has important social acquaintances.” Sample items include: “Interacting with
capital implications. These diverse networks would be people in my social network makes me interested in things
more likely to represent ties and resources valuable to the that happen outside of my town” and “I am willing to
user, containing information such as employment spend time to support general community activities.”
opportunities. However, privacy controls such as the Bonding social capital (10 items, α = .865, M = 3.88, SD
“Circles” deployed in Google+ may have positive and = .64), adapted from previous research (Williams 2006),
negative social capital implications, as they limit access captures one’s ability to mobilize solidarity and one’s
and disclosures to certain subsets of an individual’s access to emotional support and limited resources. As in
network. As suggested by other researchers, the the case of the bridging social capital items, individuals
relationship between privacy and social capital may indeed were asked to think about their entire social network.
be “paradoxical” in that privacy can both cost and enhance Sample items include: “If I needed an emergency loan of

332
$100, I know someone in my social network I could turn included three additional items to tap into Facebook users’
to” and “The people I interact with in my social network concerns about private information being revealed publicly
would be good job references for me.” on their profile as well as concerns about potential or
Signals of Relational Investment (Ellison et al., 2011b) current employers viewing incriminating content about
(SRI; 5 items, α = .80, M = 3.71, SD = .71) reflects users’ them online. Items were measures on a three-point scale (1
intent to respond to Facebook Friends’ resource requests. = Not Concerned, 3 = Very Concerned).
While not explicitly stated in the items, it is assumed that Control variables. We controlled for sex (women = 1)
these responses typically take the form of directed and age. We included a self-esteem scale (Rosenberg 1989;
communication, such as comments on status updates, 7 items, α = .91, M = 4.16, SD = .66) as a control variable
although communication outside Facebook or via other because research has established self-esteem as a strong
Facebook channels (such as private messaging) is also predictor of perceptions of social capital (Burke et al.
possible. Sample items include: “When I see someone 2010; Ellison et al. 2011a).
asking a question on Facebook that I know the answer to, I
try to respond” and “When I see someone asking for advice Hypotheses
on Facebook, I try to respond.”
We explore the relationship between privacy attitudes and
Network diversity (24 items, α = .90, M = 4.06, SD =
.51) attempts to capture network diversity by assessing behaviors, disclosures, and accrued social capital in the
social network site Facebook. While there is a large body
users’ access to various types of people and resources
of literature exploring the relationship between use of
within their social network. The 24 items were preceded by
social network sites and positive outcomes (e.g., social
the following instructions: “Please think about all of your
capital), the effects of privacy and disclosure behavior on
social connections. How easy would it be for you to find
positive outcomes are under-studied.
someone who…” Sample options include: “knows a
language you are interested in learning,” “has a political Privacy, disclosure, and positive outcomes
belief system that differs from your own,” and “can help The accrual of social capital in a SNS is a function of one’s
you fix your computer.” Items were measured on a five- activity and network composition on the site (Burke et al.
point scale (1 = Very Difficult, 5 = Very Easy). 2010; Ellison et al. 2011a). Being able to access the
Facebook use variables include time spent on the site embedded supportive possibilities of one’s network via
(M = 97.25 minutes, median = 60, SD = 113.29; “In the Facebook requires disclosure. However, one’s willingness
past week, on average, approximately how many minutes to disclose in a social network site may be affected by
PER DAY have you spent actively using Facebook?”); one’s attitudes towards privacy and one’s use of privacy
network size (M = 476, median = 450, SD = 290.59; controls. We hypothesize that:
“Approximately how many TOTAL Facebook Friends do
you have?”); and the number of “actual” Friends in one’s H1A: Bridging social capital is a function of disclosure on
network (see Ellison et al., 2011a; M = 150, median = 100, social network sites, and disclosure is a function of privacy
SD = 147.20; “Approximately how many of your TOTAL attitudes and behaviors.
Facebook Friends do you consider actual friends?”). H1B: Bonding social capital is a function of disclosure on
We created a variable measuring the ratio of “actual social network sites, and disclosure is a function of privacy
Friends” to total Friends in a Facebook network (M = .362, attitudes and behaviors.
SD = .277, median = 27.3%). Finally, we created an Social capital and relational investment
original scale, Facebook disclosures (4 items, α = .80, M = By engaging in forms of directed communication such as
2.39, SD = 1.14) to capture the extent to which Facebook replying to Facebook Friends’ requests for information or
users share information with their Facebook network. writing “Happy Birthday” on their wall, individuals engage
Sample items include: “When I’m having a bad day, I post in a form of social grooming that serves multiple purposes.
about it on Facebook” and “When I have an In addition to the disclosures that individuals broadcast to
accomplishment I’m proud of, I share it on Facebook.” their network through public channels such as status
Privacy behaviors included dichotomous measures of updates, interactions between network members are also
engagement with two privacy settings on Facebook: (1) critical to individuals’ perceptions of social capital (Lin
Friends only privacy settings, with 72% of the sample 2001). Because these responses are likely to come in forms
reporting using this setting; and (2) Segmented privacy that are seen by the recipient’s network, these behaviors
settings, with 68% of the sample responding “yes” to the may lead to expanded networks via Friends of Friends, and
item, “Have you ever changed the privacy settings so that may increase the likelihood of access to resources in the
only some of your Facebook Friends can view specific future due to norms of generalized reciprocity. Therefore,
types of content (e.g., wall, photos, notes)?” we expect that engagement in SRI will positively predict
Privacy concerns (7 items, α = .84, M = 1.81, SD = .53) users’ perceptions of bridging and bonding social capital.
is a measure adapted from Stutzman et al.’s (2011) privacy
attitudes scale that probes SNS users’ concerns about H2A: Bridging social capital is significantly and positively
potential privacy risks associated with participation in related to Signals of Relational Investment.
these sites, such as “cyberstalking” and “hacking.” We H2B: Bonding social capital is significantly and positively

333
related to Signals of Relational Investment. that our analysis is associational in nature.
Network composition and privacy management The construction of the path models follows a two-step
As Ellison et al. (2011a) demonstrate, the accrual of social approach. In the first step, disclosure, measured by the
capital in social network sites is a function of network Facebook disclosure scale, is regressed on privacy attitudes
composition, particularly the number of “actual” Friends in (privacy concerns scale) and privacy behavior (use of
the site. However, a simple measure of actual Friends is segmented privacy settings). In the second step, social
challenging in heterogeneous networks; for example, an capital is regressed on privacy attitudes, behavior, and
individual with 20 close Friends out of 25 total Friends disclosure. In both cases, we find support for a direct effect
may feel more comfortable about disclosing than an of privacy attitudes and behaviors on disclosure and
individual with 20 close Friends out of 800 total Friends. disclosure on social capital, upholding hypotheses 1A and
Therefore, we explore network composition using a ratio 1B. The regression estimates are reported in Table 1, and
measure that captures the proportion of actual Friends to the path models are presented in Figures 1 and 2.
total Friends in a social network site. We propose there is Notably, we do not find direct effects of privacy
an interaction between the composition ratio and use of attitudes or behaviors on social capital, which indicates
privacy settings in the accrual of social capital. In that the relationship between privacy and social capital is
particular, we explore the interaction between friend ratio mediated by one’s ability to disclose successfully on the
and use of privacy settings (measured as having a Friends SNS. In other words, to reap the benefits of SNS use, one
only profile), and Friend ratio and use of privacy settings must disclose on the site, and one’s ability to disclose is a
for network segmentation (measured as employing function of privacy attitudes and behaviors. This analysis
Facebook friend lists to segment content sharing). As these must be approached carefully as it is not causal, has a
analyses are exploratory in nature, we do not specify limited specification (it tests only the basic relationship and
directionality in the effect, but rather simply test for the is not carefully controlled), and has low explanatory power
presence of a significant relationship. (R2s are below .1). To improve this analysis, we conducted
a more robust specification of the model, focusing on the
H3A: There is a significant interaction between friend ratio importance of engagement, network composition, and
and the use of Friends only privacy settings in the accrual privacy in the accrual of social capital.
of bridging social capital.
H3B: There is a significant interaction between friend ratio Hierarchical Regression Analysis
and the use of Friends only privacy settings in the accural We employed hierarchical OLS regression analysis to
of bonding social capital.
H4A: There is a significant interaction between friend ratio Step 1 2 1 2
and the use of segmented privacy settings in the accrual of DV Bridging SC Bonding SC
bridging social capital. Privacy 0.13* 0.03 0.15** -0.11
H4B: There is a significant interaction between friend ratio Attitudes (0.11) (0.08) (0.11) (0.08)
and the use of segmented privacy settings in the accrual of Privacy 0.18** 0.13 0.19* 0.10
bonding social capital. Behaviors (0.13) (0.09) (0.13) (0.10)
Facebook 0.22*** 0.26***
Analysis Disclosure (0.49) (0.05)
Constant 2.69*** 3.13*** 2.61*** 3.33***
Path Models
(0.22) (0.21) (0.22) (0.22)
To test hypotheses 1A and 1B, we employed a path Observation 211 211 203 203
analysis to explore the relationship between privacy R-squared 0.059 0.079 0.069 0.089
attitudes and behavior, disclosure activities, and social E 0.97 0.96 0.96 0.95
capital. Path analysis is an extension of multiple regression Standard errors in parentheses, *** p<0.001, ** p<0.01, * p<0.05
analysis that allows exploration of hypothesized directional
Table 1: Summary of Path Regression
relationships between variables. While path analysis can be
used to supplement causal analysis, it is important to note

Figure 1. Path Model for Bridging Social Capital Figure 2. Path Model for Bonding Social Capital

334
explore hypotheses 2-4. In both the bridging and bonding Hypotheses 2A and 2B: SRI
models, we utilized the same four-step approach, adding We report the results of all models in Table 2. We find SRI
(1) controls, (2) Facebook-specific variables; (3) privacy is a significant positive predictor of both bridging and
concerns, use of a Friends only profile, and an interaction bonding social capital, with the effect being stronger for
between the friend ratio and Friends only privacy status; bridging social capital. Therefore, we find support for
and (4) use of segmented privacy settings (replacing hypotheses 2A and 2B.
Friends only) and an interaction between the friend ratio Hypotheses 3A and 3B: Friends only
and segmented privacy settings.
In step three, we add the privacy concerns scale, the binary
The logic for these regressions follows the hypothetical
variable measuring use of a Friends only profile, and the
specification. The first step of the regression establishes interaction between use of a Friends only profile and the
the baseline for the model. In the second step, we explore
Facebook friend ratio. The interaction effect is significant
the impact of SRI (Signals of Relational Investment),
and positive for bonding social capital, but not bridging,
which provides insight into the relationship between social
providing support for H3B only. We find that individuals
capital and these signals of attention in Facebook. In the
with higher proportions of actual Friends in their networks
third and fourth steps, we explore the impact of two
that employ a Friends only profile report greater bonding
separate types of privacy settings, with an interaction social capital than those who have set their profile to be
effect, on the outcome. As this analysis is exploratory, we
publicly visible to individuals beyond their “Friends” on
do not specify a directional effect for the predictions.
the site. Practically speaking, individuals that limit their
However, we note that the use of Friends only profiles and
audience and have a high proportion of actual Friends in
segmented privacy settings operate quite differently in the
their network will be the prime beneficiaries of the
context of social capital. Whereas changing one’s
bonding outcomes of SNS use, perhaps because they feel
activities to be visible to Friends only may lessen more free to disclose to this group.
constraints on disclosure within a defined set of ties, the
use of segmented privacy settings on some kinds of content Hypotheses 4A and 4B: Segmented privacy settings
may limit the audience of this content. In step four, we include the privacy concerns scale, the

Step 1 2 3 4 1 2 3 4
DV Bridging Social Capital Bonding Social Capital
Sex (1=Female) 0 13* 0 07 0 05 0 06 0 05 0 02 0 00 0 03
(0 08) (0 08) (0 08) (0 08) (0 07) (0 07) (0 08) (0 08)
Age -0 02* -0 01 -0 01 -0 01 -0 00 -0 01 -0 01 -0 01
(0 01) (0 01) (0 01) (0 01) (0 01) (0 01) (0 01) (0 01)
Self-Esteem 0 29*** 0 30*** 0 31*** 0 32*** 0 51*** 0 51*** 0 50*** 0 51***
(0 06) (0 06) (0 06) (0 06) (0 06) (0 05) (0 06) (0 06)
Network Diversity 0 46*** 0 42*** 0 42*** 0 39*** 0 35*** 0 26*** 0 25*** 0 25***
(0 07) (0 08) (0 08) (0 08) (0 07) (0 07) (0 07) (0 08)
SRI 0 22*** 0 24*** 0 22*** 0 13** 0 15** 0 13**
(0 07) (0 07) (0 07) (0 07) (0 07) (0 07)
Minutes on FB/Day 0 00 0 00 -0 00 0 00 0 00 0 00
(0 00) (0 00) (0 00) (0 00) (0 00) (0 00)
FB Disclosure -0 01 -0 03 -0 02 0 04 0 04 0 04
Scale (0 06) (0 06) (0 06) (0 05) (0 05) (0 06)
FB Friend Ratio -0 02 -0 39 0 49* 0 39*** -0 19 0 39
(0 13) (0 27) (0 27) (0 13) (0 28) (0 26)
Privacy Concerns 0 12* 0 09 -0 05 -0 06
Scale (0 07) (0 07) (0 07) (0 07)
Friends-Only -0 23 -0 20
Status (1=Y) (0 14) (0 14)
RatioXFriendsOnly 0 49 0 73**
Interaction (0 31) (0 31)
Segmented Privacy 0 32** 0 02
Settings (1=Y) (0 13) (0 13)
RatioXSegmented -0 68** -0 01
Interaction (0 31) (0 30)
Constant 1 08*** 0 25 0 24 -0 02 0 37 0 18 0 51 0 29
(0 36) (0 40) (0 43) (0 41) (0 35) (0 38) (0 41) (0 41)

Observations 208 193 191 191 202 186 184 184

R-squared 0 34 0 43 0 45 0 46 0 43 0 52 0 53 0 52
Standard errors in parentheses, *** p<0 001, ** p<0 01, * p<0 05
Table 2. Hierarchical Regression Estimates for Bridging and Bonding Social Capital.

335
binary variable measuring the use of segmented privacy Next, we ran a series of hierarchical regression analyses
settings, and the interaction between segmented privacy incorporating privacy, disclosures, interaction with one’s
settings and friend ratio. We find support for hypothesis network (SRI), and network characteristics. What emerged
4A, noting that both the main effect and interaction effect from these analyses provides some additional support to
for segmented privacy settings are significant. Notably, the privacy paradox argument that has previously been
while the use of segmented privacy settings is positively made for both the Internet generally and SNSs specifically.
and significantly associated with bridging social capital, The significant interaction effects between one’s friend
the interaction effect is significant and negative. That is, ratio and specific privacy behaviors suggests that privacy
we find that individuals with higher proportions of actual and social capital are indeed related in ways that have not
Friends in their networks that employ segmented privacy previously been examined. If users want to extract the
settings tend to have lower perceptions of bridging social greatest benefits from their network, it is important that
capital than those who are not employing these settings. they are interacting with that network, as seen in the strong
This finding is not surprising, in that individuals with a positive beta for SRI in both models. SRI assesses
high proportion of actual Friends and who segment behaviors such as responding to Friends’ request and
disclosures may be less likely to engage with the wide engaging in social grooming activities.
variety of social media content associated with bridging When looking at the model predicting bonding social
social capital. That is, users who have a high proportion of capital, we find that perceptions of bonding are higher for
actual Friends and segment their networks may be less those with a high actual-to-total friend ratio when they
likely to be exposed to the diverse content and novel restrict their profile to Friends only. As bonding social
interactions associated with bridging social capital. We do capital is associated with provisions of emotional support
not find support for hypothesis 4B. and, generally, more intimate relationships, it makes sense
that users with tight-knit Facebook networks may feel
Discussion more comfortable disclosing, and enacting this privacy
strategy may lead them to be more open in their requests
This study bridges two separate streams of research related for various resources. These intimate networks should
to how people engage with SNSs and the potential benefits contain higher levels of trust, which is associated with
that can be accrued through use of sites such as Facebook. more open sharing and thus offers of support.
We show that people’s attitudes and behaviors regarding When looking at the model predicting bridging social
privacy, their disclosures and interactions on SNSs, and the capital, a different picture emerges. First, we find a
resources that can be derived through these sites have a positive relationship between the actual-to-total friend
complex, interdependent relationship. Furthermore, we ratio, suggesting that when users’ networks include a
show that one’s network composition—operationalized in greater percentage of closer ties, they have greater
this study as the ratio of “actual” Friends to total Friends perceptions of bridging social capital. This may seem
on Facebook—interacts with users’ engagement in various counter-intuitive at first, but it may be useful to think of
privacy-enhancing behaviors when predicting users’ relationships on a spectrum from strong to weak. While
perceptions of both bonding and bridging social capital. weak ties provide benefits in terms of diverse people and
In the first analysis, we found support for a directional information, ties that are too weak are unlikely to respond
relationship between privacy attitudes and behaviors, to a resource request. Looking at the interaction, we find
disclosure in Facebook, and social capital. In conducting that these more tight-knit networks who employ segmented
this analysis, we demonstrate that the relationship between privacy settings to limit content have lower bridging than
privacy and social capital is mediated by disclosure, and those who do not partition access to content. One potential
that privacy does not act exogenously to constrain social explanation for this finding is that by restricting access,
capital outcomes. While this finding is inherently limited resource requests may not be reaching the proper audiences
due to its simple specification and limited explanatory who would be able to respond accordingly.
power, it is insightful for designers as it demonstrates that SRI was a significant predictor of both bridging and
individuals with different privacy needs can equally reap bonding social capital. While the direction of the
social capital benefits of participation in social network relationship is indeterminate in this analysis, SRI has a
sites. Although we do not claim causality, we observe that stronger coefficient in relation to bridging social capital
privacy’s relationship to social capital is indirect, having than to bonding. Conscious signals of attention on the site
more of an impact on the behaviors that lead to social are more powerful for weak ties, as opposed to strong ties,
capital exchanges. Concerns about sharing information and who no doubt share multiple channels of interaction.
engaging in strategies to protect one’s content are more Many SNSs are now engaging in design processes
likely to be the deciding factor in whether a disclosure is intended to help people segment their audience and allow
made through a public channel on Facebook, which will users to control who sees their content. Because bonding
subsequently impact whether network members will social capital is more strongly associated with individuals
respond. These interactions constitute social capital in who have a high actual-to-total Friends ratio and have
action, and our findings shed insight into the process of employed Friends only privacy settings, tools could be
social capital accrual, not just the outcomes. created to enhance content sharing within known, limited

336
 audiences. Recently, Facebook and Google+ introduced Dunbar, R. Grooming, Gossip, and the Evolution of Language. Harvard
user-friendly limited sharing functionality within their University Press, Cambridge, 1996.
platforms. These tools may provide opportunities for Ellison, N.B., Steinfield, C., and Lampe, C. The benefits of Facebook
“Friends:” Social capital and college students’ use of online social
accruing bonding social capital via the disclosure process network sites. JCMC 12, 4 (2007), 1143-1163.
outlined in this paper. Ellison, N., Steinfield, C., and Lampe, C. Connection strategies: Social
Limitations capital implications of Facebook-enabled communication practices. New
Our research explores perceived privacy behaviors in our Media & Society, 13, 6 (2011a), 873-892
sample, and many of our measures of interest are not Ellison, N., Vitak, J., Lampe, C., Gray, R., and Brooks, B. Cultivating
discernable from behavioral data such as server-level logs. social resources: The relationship between bridging social capital and
Facebook use among adults. A Decade in Internet Time Symposium on
Our research questions thus necessitate the use of self-
the Dynamics of the Internet and Society. Oxford Internet Institute, 2011b.
report data. Our findings are associational in nature and
Ellison, N., Vitak, J., Steinfield, C., Gray, R., and Lampe, C.. Negotiating
findings should not be interpreted as causal. Finally, as is
privacy concerns and social capital needs in a social media environment.
common with survey research, response to our solicitation Privacy online Perspectives on privacy and self-disclosure in the social
was low, but in line with comparable studies. In this work, web (eds. S. Trepte & L. Reinecke). Springer, New York (2011c), 19-32.
we have attempted to follow best-practice social science Farnham, S. D. and Churchill, E. F. (2011). Faceted identity, faceted
research methods (e.g., random sampling), which lives: Social and technical issues with being yourself online. Proc. CSCW
strengthens the contribution in light of these limitations. 2011, ACM Press (2011), 359-368.
Conclusion Hogan, B. The presentation of self in the age of social media:
Distinguishing performances and exhibitions online. Bulletin of Science,
When people use social network sites like Facebook, they Technology & Society, 30, 6 (2010), 377-386.
have the opportunity to access resources such as emotional Kelley, P., Brewer, R., Mayer, Y., Cranor, L., and Sadeh, N. An
support and novel information. However, people also have investigation into Facebook friend grouping. Proc. INTERACT 2011,
concerns about privacy that may limit how much they are Springer (2011), 216-233.
willing to disclose to their Facebook networks. By Krasnova, H., Spiekermann, S., Koroleva, K., and Hildebrand, T. Online
surveying Facebook members about their attitudes towards social networks: Why we disclose. Journal of Information Technology,
and use of the site, we have shown that those who engage 25, 1 (2010), 109-125.
in activity with their Friends feel they have more access to Lampinen, A., Tamminen, S., and Oulasvirta, A. All my people right
here, right now: Management of group co-presence on a social
resources. We have shown that privacy concerns on their
networking site. Proc. GROUP 09, ACM Press (2009), 281-290.
own may not affect people’s perceptions of access to
Lampinen, A., Lehtinen, V., Lehmuskallio, A., and Tamminen, S. We’re
bridging and bonding social capital; however, privacy does in it together: Interpersonal management of disclosure in social network
have a relationship with one’s willingness to disclose services. Proc. CHI 2011, ACM Press (2011), 3217-3226.
information, which both positively and negatively affects Lin, N. Social Capital A Theory of Social Structure and Action.
social capital perceptions. Our approach highlights the Cambridge University Press, NY, 2001.
complex and interactive nature of these relationships in Marwick, A. and boyd d. I tweet honestly, I tweet passionately: Twitter
ways earlier research does not. As SNS designers work to users, context collapse, and imagined audience. New Media Society 13, 1
make their sites more valuable to users, they must (2011), 114-133.
address—and engage—this new privacy paradox. Ozenc, F. K. and Farnham, S. D. Life “modes” in social media. Proc.
CHI 2011, ACM Press (2011), 561-570.
Putnam, R. Bowling Alone The Collapse and Revival of American
Community. Simon & Schuster, NY, 2000.
References Rosenberg, M. Society and the Adolescent Self-Image. Wesleyan
AAPOR. Standard Dentitions Final Dispositions of Case Codes and University Press, Middletown, CT, 1989.
Outcome Rates for Surveys (5th Ed.). Lenexa, KS: AAPOR, 2008. Spiekermann, S., Grossklags, J., and Berendt, B. E-privacy in 2nd
Acquisti, A. and Gross, R. Imagined communities: awareness, generation E-commerce: privacy preferences versus actual behavior.
information sharing, and privacy on the Facebook. Proc. PET 2006, Proc. ACM CCE, ACM Press (2001), 38-47.
Springer (2006), 36-56. Stutzman, F., Capra, R., and Thompson, J. Factors mediating disclosure in
Altman, I. The Environment and Social Behavior. Wadsworth, Belmont, social network sites. Computers in Human Behavior, 27, 1 (2011), 590-
CA, USA, 1975. 598.
Barnes, S. A privacy paradox: Social networking in the United States. Stutzman, F. and Kramer-Duffield, J. Friends Only: Examining a privacy-
First Monday, 11, 9 (2006). enhancing behavior in Facebook. Proc. CHI 2010, ACM Press (2010),
1553-1562.
Bourdieu, P. The forms of capital. Handbook of Theory and Research for
the Sociology of Education (ed. J.C. Richardson). Greenwood, New York Tufekci, Z. Can you see me now? Audience and disclosure regulation in
(1985), 241-258. online social network sites. Bulletin of Science and Technology Studies,
11, 1 (2008), 544-564.
Burke, M., Kraut, R., and Marlow, C. Social capital on Facebook:
Differentiating uses and users. Proc. CHI 2011, ACM Press (2011), 571- Williams, D. On and off the 'net: Scales for social capital in an online era.
580. JCMC 11, 2, 2006.
Burke, M., Marlow, C., and Lento, T. Social network activity and social Yoder, C. and Stutzman, F.. (2011). Identifying Social Capital in the
well-being. Proc. CHI 2010, ACM Press (2010), 1902-1912. Facebook Interface. Proc. CHI 2011, ACM Press (2011), 585-588.
Burt, R. S. Structural Holes The Social Structure of Competition. Zhao, S., Grasmuck, S., and Martin, J. Identity construction on Facebook:
Harvard University Press, Cambridge, MA, USA, 1992. Digital empowerment in anchored relationships. Computers in Human
Behavior, 24, 5 (2008), 1816-1836.
Donath, J. Signals in Social Supernets. JCMC 13, 1 (2007).

337

View publication stats