Exploring the relationship between perceptions of social capital and enacted support online

Journal of Computer-Mediated Communication

Exploring the relationship between perceptions

of social capital and enacted support online

Michael A. Stefanone

Department of Communication, University at Buffalo,

Kyounghee Hazel Kwon

Culture & Communication, Drexel University,

Derek Lackaff

School of Communications, Elon University,

Online social networking sites enable users to connect with large, heterogeneous groups of people.

While extant research suggests individuals beneﬁt psychologically from the perception that they are

well connected, little is known about the nature of tangible resources embedded in these online

networks. In this study 49 participants sent 588 requests for instrumental help to their Facebook

friends to determine the accessibility of networked resources and online social capital. Almost 80% of

these modest requests went unanswered, and perceived bridging and bonding capital did not explain

enacted support. However, people who occupied socially prestigious positions were the most likely to

beneﬁt from their friend’s help. These results suggest that expansive mediated networks may yield

limited instrumental beneﬁts.

Key words: online social capital, social networks, enacted support, perceptions, behavior

doi:10.1111/j.1083-6101.2012.01585.x

Humans always ﬁnd themselves involved in social groups. Today, these groups are routinely mediated

by communication technology. Web 2.0 – the social web—is characterized best as the set of tools

that facilitate production and distribution of content produced by everyday people. In particular, there

is currently pervasive interest in the relationship between this content, (online) social networks, the

nature of people’s relationships mediated by websites like Facebook.com, and the changing role people

now play in the production and consumption of mass-mediated messages. Considering that sites like

Facebook.com facilitate the accumulation of expansive networks of acquaintances, there are pressing

questions about the relationship between the characteristics of online networks, access to social capital,

and outcomes like psychological wellbeing and access to resources embedded in these networks.

Social networks are the conduit for the entirety of human social behavior and are comprised by a

range of relationships with varying qualities. Granovetter (1973, 1982) was the ﬁrst to formalize the

nature of relationship strength in social networks by arguing that social networks consist of relationships

ranging from very weak in strength to very strong. Weak and strong tie relationships afford access to

different kinds of resources (Lin, 2001). For example, the strength of weak ties lies in their capacity

to connect people to novel information and resources that reside in and propagate across networks

(Hansen, 1999).

Considering how connected people have become online and the evidence suggesting that users

are consequently better off (Ellison, Steinfeld, & Lampe, 2007), we propose to explore the nature of

the relationships that comprise these networks and whether tangible, instrumental resources accrue

to users. The current research reports on a quasi-experimental study designed to test the relationship

between requests for instrumental help via Facebook, the relationship characteristics between request

senders and receivers, and actual, enacted support. Grounded in the theory of instrumental action (Lin,

1982) and limited to instrumental support, our goal is to establish a baseline for networked resources

by exploring the accessibility of social capital embedded in online networks.

Theories of Self-Interest

Social scientists interested in self-interest as a motivation for social action (Coleman, 1986) suggest

people make what they believe to be rational choices while pursuing objectives (Monge & Contractor,

2003). People are not always objective, systematic beings (see for example, Frijda, 1986). However,

rational choice theorists (e.g., Homans, 1950) outline a process in which people weigh outcomes

based on alternative actions and act based on the optimal solution to cost – beneﬁt analyses. This

approach suggests that people actively monitor and process environmental stimuli with the purpose of

maximizing their individual outcomes.

Much of the literature on self-interest guided behavior grew out of research on status attainment

(e.g., Blau & Duncan, 1967). The theory of instrumental action offers an alternative perspective

by explicating the nature of relationships and embedded resources. While status attainment was

operationalized as a function of ‘‘given’’ social network properties, status attainment can also be viewed

as a product of strategic relationship choices.

Instrumental Action

Lin’s (1982) theory of instrumental action suggests that people actively pursue opportunities and

resources for their personal beneﬁt. People have an intrinsic tendency to negotiate their social

environments in ways that maximize chances for personal gain. First, status in groups matters when it

comes to access to social resources. Consider the process of searching for employment opportunities.

The social resources proposition of the theory of instrumental action states that if someone uses a

contact higher in status than themselves to explore potential employment opportunities, they are likely

to ﬁnd a better job than someone who uses a contact of lesser status (Lin, 1999). Status in networks can

be inferred from formal hierarchies like organizational charts. However, social hierarchies also exist

outside the boundaries of organizations. For example, some people have ‘‘magnetic personalities’’ that

make them the center of attention, and physically attractive people are frequently the objects of others’

affection. Not surprisingly, it can be advantageous to be in positions like these.

Lin and Dumin (1986) focused explicitly on factors affecting access to social resources, concep-

tualized as the way a person’s network may connect them with a variety of different positions. They

operationalized strength of ties based on the nature of the relationship; relatives were coded as strong

ties, friends as moderate strength ties, and acquaintances as weak ties. As expected, social contacts with

high positions in networks and weaker tie afﬁliation (both friends and acquaintances) provided better

access to prestigious job opportunities, support for both the strength of positions and strength of ties

propositions reviewed above. Further, weak ties were more instrumental for people whose original

positions in the network were relatively low. It is unclear, however, whether or not weak ties mediated

by networking sites can be activated for instrumental gain.

Taken together, the evidence summarized above suggests that structural characteristics of ego

networks and positions in social hierarchies inﬂuence access to and use of resources embedded in

social networks, or networked resources (Wellman & Frank, 2001). Overall, weak ties have greater

instrumental functionality than strong ties, regardless of the structural location of those weak ties.

When using people for instrumental goals, the literature suggests that the likelihood of others enacting

support increases with social prestige, or status.

Social Status

Many deﬁnitions of social status have been proposed. For example, Moreno (1934) quantiﬁed concepts

of sociometric stars and isolates, where people situated in the center of star-shaped communication

networks beneﬁt from higher levels of status, while isolates i n networks exhibit lower levels. Lin,

Vaughn, and Ensel’s (1981) model incorporated the socioeconomic status of personal contacts used to

ﬁnd employment. This early iteration of the theory of instrumental action suggested that access to and

use of social resources play an important part in successful instrumental action. Essentially, the authors

argue that ‘‘if social ties have different instrumental consequences, then the status of the contact should

be a good indicator of the structural advantage of the tie’’ (p. 1166). We suggest that one form of

status – popularity – manifests itself via unreciprocated relationships. Celebrities are a good example

of popular people as they are the object of attention and affection from mass audiences. Similarly, some

Facebook users may beneﬁt from having higher social status than others, and this characteristic of their

social position may affect access to resources embedded in social networks.

In summary, Lin’s (1982) theory of instrumental action describes goal directed behavior which

beneﬁts the person taking action. Such behavior is deﬁned as instrumental in nature and is restricted to

actions that involve other people. Social resources are embedded in social networks (Lin et al., 1981; Lin,

2001), commonly known as social capital (Coleman, 1988) and these resources are used to maintain

or promote an individual’s welfare. Thus, the theory focuses on instrumental action initiated for the

purpose of gaining valued resources that reside in social systems.

Social Capital

According to Adler and Kwon (2002), social capital is roughly understood as ‘‘the good will that is

engendered by the fabric of social relations ... mobilized to facilitate action’’ (p.17). Other scholars

including Coleman (1988) and Kadushin (2004) emphasize that social capital is embedded in social

relations that develop during the pursuit of instrumental goals. In sum, social capital can be deﬁned

as networked resources that are created, maintained, and realized by social relations occurring via

mediated communication (Wellman & Frank, 2001). Lin (1999) proposes a clearly operationalizable

deﬁnition of social capital as ‘‘investment in social relations by individuals through which they gain

access to embedded resources to enhance expected returns of instrumental or expressive actions’’ (p.

39). Lin’s deﬁnition is particularly useful because it elucidates the social nature of capital.

Social support is one type of resource that is accessed through social networks, and refers to

availability of emotional and material support from others. Barerra (1986) suggests that social support

research should clearly differentiate among three major concepts: social embeddedness, perceived

social support, and enacted support. Social embeddedness refers to the structures of relationships

connecting people. This is typically measured with social network analytic techniques which facilitate

the quantiﬁcation of structural properties of communication networks (Walker, Wasserman, &

Wellman, 1994; Wellman, Carrington, & Hall, 1988; Wellman & Gulia, 1999; Wellman & Wortley,

1990).

Perceived social support is one of the broadest and most prevalent operationalizations in the

social support literature (Barrera, 1986) and reﬂects idiosyncratic perceptions of support, rather than

social structure. Perceived social support has been found to correlate with a range of psychosocial

and physiological responses and behaviors including coping (Tao et al., 2000), academic achievement

(Eggens, van der Werf, & Bosker, 2008), and even blood pressure (O’Donavan & Hughes, 2007). More

recently, research on social network sites has adopted measures differentiating between bridging and

bonding support (Williams, 2006).

Finally, enacted support refers to the actual provision or reception of support. Barerra (1986)

notes that enacted support is often measured using self-report data. As such, much extant research

has actually measured perceived-received support. Valid measures of enacted support should therefore

utilize behavioral observation or dyadic analysis. Thus, behavioral measures of enacted support are

used in the current study.

As mentioned above, different perspectives on the nature of social relationships and resources have

lead to the identiﬁcation and operationalization of two related forms of social capital: bonding and

bridging. Bonding capital is understood as embedded in internal, or closely connected social ties (Adler

& Kwon, 2002) and research shows that perceptions of bonding capital increases credibility assessments,

garners consensus from others, and enhances emotional support (Williams, 2006). Bonding capital can

be particularly advantageous for collective endeavors (Klandermans, 1984; McAdam & Paulsen, 1993;

Opp & Gern, 1989). For example, Coleman (1988) focused on a student revolution in Korea to discuss

the collective returns of bonding social capital within small clandestine groups. Gould (1991) also

illustrated the importance of neighborhood relations in exerting contagious motivation toward protest

participation. Thus, bonding capital is related to group solidarity, which in turn should be related to

enacted, mutual social support.

Bridging capital is associated with diverse social ties (Adler & Kwon, 2003) and is understood as

linkage capital because it facilitates connections to otherwise disparate social groups. The advantage of

bridging capital lies in its ability to connect people to novel, nonredundant social resources. For example,

information ﬂow between groups providing instrumental resources may be limited in homogeneous

networks exhibiting insulating properties opposed to heterogeneous networks where subgroups are

connected by liaisons (Granovetter, 1974). Accordingly, bridging social capital is understood as beneﬁts

stemming from network diversity.

Recently, Hampton, Lee, and Her (in press) explored the relationship between off- and online

behavior and network diversity, and framed their investigation in the context of advantages associated

broadly with ‘‘accessible social capital’’ (p. 14). They found that internet use, and in particular the use of

SNS, had positive relationships with network diversity (or, bridging capital). Their results also suggest a

negative relationship between SNS use and the number of ofﬂine neighborhood ties people maintained,

which suggests a replacement process whereby resources obtained by local, ofﬂine relationships are now

accessible by mediated interpersonal relationships. Their results are consistent with Wellman’s (2001)

argument that in a networked society, physical proximity is becoming less important in terms of access

to social capital. However, questions persist regarding the extent to which actual, enacted support

resources accrue to users today.

Bonding and bridging social c apital develop through regular activity that (in)directly facilitates

interaction with other people and the social nature of contemporary Internet use lends itself to the

accumulation of bonding and bridging capital (Wellman et al., 2001). The widespread use of social

media increases perceived social capital (Ellison et al., 2007) which can create opportunities to expand

the size of recruitment pools for instrumental action. The level of interpersonal and collective capital

built online may contribute to instrumental action on- and ofﬂine. Our goal is to explore the nature

of networked resources embedded within mediated social networks and to explore the relationship

between perceptions of online bonding and bridging capital and actual, enacted support. The speciﬁc

hypotheses are presented next.

Hypotheses

The current study is composed of two sets of hypotheses. The ﬁrst explores the provision of instrumental

support via relationships mediated by SNSs. Research suggests that perceptions regarding online social

capital are positively associated with an individual’s psychological well-being. However, we are not

aware of any studies that examine the relationship between perceived social capital and the actual

capacity of generating enacted support.

Considering the positive relationship between perceptions and psychological well-being, it is likely

that people who are happier actually do have heightened access to resources. Thus, we hypothesize that

an individual’s perceived bonding social capital should have a positive relationship with the likelihood

of procuring instrumental beneﬁts from online networks. It is unlikely that bridging capital has a

relationship with enacted support. Thus,

H1: Higher levels of perceived bonding capital have a positive relationship with the provision of

enacted support.

Another set of hypotheses are posited by considering speciﬁc relationship characteristics between

SNS friends independent of perceptions of bonding social capital online. Many network scholars have

discussed the multidimensionality of relationship characteristics including Campbell and Marsden

(1984) who found that tie strength—operationalized as emotional closeness and communication

frequency—are distinct constructs. Wellman and Wortley (1990) also treated tie strength and contact

frequency a s distinct variables and they propose a range of explanations for interpersonal, enacted

support. Of particular interest to the current study is the distinction between relational explanations

which include ‘‘the strength of the relationship or ... access that two persons have to each other’’ (p.

560). Here, strength is characterized by voluntary, intimate relationships. On the other hand, access is

related directly to communication frequency and interaction (Galaskiewicz, 1985).

In light of this evidence, it is likely that tie strength and contact frequency operate as two separate

variables that explain the provision of support. Thus,

H2. Tie strength has a positive relationship with the provision of enacted support.

H3: Communication frequency has a positive relationship with the provision of enacted support.

While the hypotheses above consider whether Facebook friends provide instrumental resources or

not, they do not a ddress the quality of resources provided. We propose that the caliber of enacted

support is a function of how much time and effort is invested in fulﬁlling requests for help. Accordingly,

because people are more heavily invested in their strong tie networks, they should provide higher quality

support than weak ties. Thus,

H4: Tie strength has a positive relationship with the quality of enacted support.

Finally, relationships are often unbalanced in terms of liking and affection. Just because Frank

perceives a very close relationship with Judy, for example, Judy may not perceive a reciprocal level of

intimacy with Frank. In this elementary example, Judy beneﬁts from heightened social status and holds

an advantageous social position. We operationalize this perceptual gap as a form of social status. People

who occupy favorable positions in social hierarchies should be afforded enhanced access to resources

embedded in their social networks. This is analogous to the strength of positions proposition outlined

in the theory of instrumental action. Thus,

H5: Social status has a positive relationship with the quality of enacted support.

This study affords us the opportunity to start validating existing research designed to assess different

elements of social capital based on self-report data. Essentially all extant work on the relationship

between online social networks and social capital rely on people’s perceptions of access to these

resources (e.g., Williams, 2006). However, we are not aware of research that uses actual behavior as a

measure of the availability of and access to networked resources, and the quality of that support. Thus,

this study also proposes the following research question:

RQ1. What is the relationship between perceptions of online social capital and the quality of

enacted support on SNS?

Method

During Spring 2010, participants were drawn in two steps and consist of primary and secondary

participants, hereto referred to as ‘ego’ and ‘alters,’ respectively. First, ego’s (N = 50) were recruited

from communication classes at a large northeastern university and instructed not to discuss this study

with anyone else until completion of the study (a two week period). Participation was voluntary and

this project had the approval of the institutional review board for human subjects.

Each ego was instructed to examine their entire Facebook friend network and to think about their

six strongest and six weakest relationships on this site. The strong tie sample size was chosen based on

extant research suggesting that people generally have about six very close people in their lives (Bernard

et al., 1990), and the weak tie sample size was chosen to balance the design. Then they were required

to record the identities and contact information for each of these 12 online friends (alters). One ego

did not follow the procedure and was eliminated from the study. As a result, 49 egos sent a total of

588 requests to alters. One alter was chosen twice and was subsequently eliminated from the analysis

leaving 586 unique alters.

Next, egos completed a brief survey measuring demographic information and their perceptions

about a series of relationship characteristics for each of the 12 alters they identiﬁed. The speciﬁc

measures are described in the measures section, below. Finally, they were instructed to send a request

message to each alter. The standard request message explained that the sender needed help with a

class-related project, which read as follows:

Hey, [Alter’s First Name]- I need your help with a c lass project I’m working on. I need people to

provide labels for a series of online images. I’d really appreciate your help! Please g o to [study URL]

and take the quick survey and label as many images as you can. Your participation will be a huge

help. Thanks!

Figure 1 Screen shot of image-labeling task interface.

Each alter received only one request and all communication transpired via the Facebook message

tool which is analogous to e-mail.

Task

We limited the request to a low urgency, low stages task in an effort to establish a baseline response

to modest requests on Facebook. This conservative approach was chosen because of the dearth of

research in this area using actual behavior metrics. The request prompted alters to access a webpage and

complete a brief survey followed by an image-labeling task (see Figure 1, below). Each survey URL was

uniquely associated with each request, and the survey measured demographic variables and included

the same series of relationship-speciﬁc variables about the ego who sent the request for help (described

below). The image labeling task randomly generated a series of generic images and allowed visitors to

enter a text label for each image. The quality of enacted support was operationalized as the number of

images labeled, and the website recorded the number of images alters labeled. All egos were advised that

if their friends contacted them about the message, they were to maintain the ruse until the researchers

could debrief all a lters after a 2-week period.

Measures

Data collection proceeded in two stages. First, when egos arrived at the lab to participate in the study,

an initial survey was administered. Stage 2 occurred when alters responded to requests for help. Each is

discussed next.

Stage 1.

The survey for egos included three Likert-type items used to measure tie strength for each of the 12

online friends they selected (Marsden & Campbell, 1984; Wellman & Wortley, 1990). Items included

‘‘This person is a ...’’ (1 = casual acquaintance, 7 = very good friend), ‘‘How close are you with

this person?’’ (very distant, very close), and ‘‘Do you interact with this person voluntarily rather than

because you are both members of the same social institutions?’’ (not voluntary, completely voluntary).

The interitem reliability was very high, Cronbach’s α for egos was .98 (M = 5.17, SD = 2.17).

Perceived Social Capital. We adopted Williams’ (2006) Internet Social Capital Scale (ISCS) to

measure ego’s perceptions of online bonding and bridging capital on Facebook (10 and 9 items,

respectively). An example of bonding capital items included ‘‘there is someone on Facebook I can turn

to for advice about making very important decisions’’ (Cronbach’s α for all responses = .88; M = 4.70,

SD = 1.31). An example of bridging capital items included ‘‘interacting with people on Facebook makes

me interested in things that happen outside of my town’’ (Cronbach’s α = .85; M = 4.42, SD = .10). A

conﬁrmatory factor analysis using varimax rotation yielded a two factor solution explaining about 57%

of the total variance, and all items were retained (see appendix A for descriptive statistics and factor

loadings for the ISCS scale).

Tie Strength. College students’ average Facebook social network size is greater than 250 people

(Stefanone, Lackaff & Rosen, 2010). Given that human cognitive capacity allows for the effective

management of only a limited number of strong ties (Hill & Dunbar, 2003; Roberts, Dunbar, Pollet,

& Kuppens, 2009), it is probable that the majority of Facebook friends are weak ties. Thus, if random

selection were utilized the sample of alters would have overrepresented weak ties. This would make

between-group comparisons difﬁcult. Accordingly, ego-reported tie strength was dummy coded as

either strong (= 1) or weak (= 0) based on median split.

Social Status. Status was operationalized as the discrepancy between the three items used to measure

ego and alters’ perceptions of relationship strength. Social status was calculated by subtracting ego

responses from alter responses for each item and averaging them (Cronbach’s α = .88). Higher scores

correspond to higher status positions in the interpersonal relationships. When egos reported liking

alters more, they suffered from low status. In these cases, the values were negative. Overall these scores

ranged from −7 (lowest possible status) to 7 (maximum). The mean status score for egos was .24 (SD

= 1.37) suggesting these relationships were fairly balanced overall.

We also included an item measuring the frequency of communication via Facebook. While

Facebook communication represents one aspect of relationship strength, it also reﬂects use of CMC to

maintain relationships. In this sense, Facebook contact frequency is understood as a Facebook-speciﬁc

characteristic of relationship strength. Facebook contact frequency was measured with a 7-point scale

(1 = has been more than a year since the last contact, 2 = only a few times a year, 3 = 1 or 2 times

amonth,4= 3–4 times a month, 5 = 1–2 times a week, 6 = more than 2 times a week, 7 = almost

daily). The average contact frequency was 2.55 (SD = 1.65).

Stage 2

When alters responded to requests, they were ﬁrst required to complete a brief survey. This survey

was similar to the survey egos completed and included the same demographic and relationship-related

questions. However, these items were framed such that alters evaluated the nature of their relationship

with the ego who made the request for help. All of these alters completed the survey (Cronbach’s α for

tie strength = .86) and then were automatically directed to a website for the image labeling task.

Outcomes. Two dependent variables were used to test the hypotheses and the research question

proposed above. In the ﬁrst analysis, the binary outcome of whether or not alters responded to requests

for help was used. Recall that 96 people responded to help requests by visiting the image-labeling

website. The dependent variable in the second analysis was measured as the quality of help provided.

This was measured as the number of images labeled by alters.

Results

26 of the 49 egos were male, and the mean age of participants was 20.8 years (SD = 1.7). Participants had

on average 426.9 Facebook friends (SD = 148.20), and spent about 35 minutes logged in per session (SD

= 19.5). The strong tie group of alters had an average relationship duration of 8.9 (SD = 6.0) years and

reported a contact frequency of 4.1 (SD = 1.7) times per week. The weak tie group reported an average

history of 5.8 (SD = 6.3) years and communicated 2.8 (SD = 1.8) times weekly. T-tests conﬁrmed that

Table 1 Correlations and descriptive statistics for secondary participant variables.

Model with Binary

DV (N = 586)

Model with Count

DV (N = 96)

M(SD) 1 2 3 4 5 6 7 8

1 Tie strength .50 ( .5) — 0.68

***

−.21

***

2 Comm Freq. 2.55 (1.65) — .26

***

3 DV1 .16 ( .37) –

4Sex(1= female) .67 ( .48) — −0.11 −0.06 −0.16 0.19

5 Age 20.66 (6.32) — 0.03 0.01 0.02

6 Tie strength .74 ( .46) — .42

***

−0.01

7 Social Status .21 (1.57) — .24

8 DV2 25.01 (42.09) —

Note:**p< .01, *** < .001; DV 1 = completed survey (Y, N); DV 2 = number of images labeled by

alters.

the strong tie sample of secondary participants (N = 294) was characterized by signiﬁcantly higher

communication frequency (p < .01) and greater emotional closeness (p < .001) opposed to the weak

tie sample. It is notable, however, that the weak tie sample was not as weak as anticipated; egos reported

communicating with these ties about three times per week. Table 1 below summarizes the relationship

between variables used in this study. As expected, reported tie strength had a positive relationship with

contact frequency and relationship duration.

Overall, 98 of the 588 requests for help were answered secondary participants. Interestingly, 10

participants did not receive any responses to their requests for help. Nine participants had only one

friend respond. The m ajority—16 participants—had 2 responses from their friends. Although two

participants received responses from six of their friends, none of the primary participants received more

than six responses to their requests for help.

Multilevel Logistic Analysis. Hypotheses H1 through H3 were tested with the entire pool of 588

secondary participants from stage 2 of data collection as units of analysis. Because these data were

nested, we conducted multilevel logistic analysis using the Bernoulli distribution and Penalized Quasi-

Likelihood estimation (PQL). To increase the robustness of the models, robust standard error was

used. Multilevel models have two levels of variables. In our analysis, tie strength and Facebook contact

frequency were used as level-1 variables similar to regular logistic regression models. In addition, we

considered the random component in the intercept and primary participants’ perceptions of online

bonding and bridging capital as level-2 factors. All independent variables were grand-mean centered.

Because tie strength and contact frequency were highly correlated (r = .682, p < .001), we present

three separate models. Tie strength and contact frequency were included separately in the ﬁrst two

models. The third model included a combined measure of tie strength and contact frequency as a single

independent variable. Table 2 below summarizes the results from these three models.

First, the results show that the random component of the intercept was signiﬁcant in all three

models, indicating that enacted support differed depending on which primary participant made the

request (Model 1, χ

(46) = 70.54, p < 0.05; Model 2, χ

(46) = 69.11, p < 0.05; Model 3, χ

(46)

= 69.93, p < 0.05). However, hypotheses 1 was not supported. Primary participants’ perceived level of

bonding capital did not explain enacted support.

When tie strength was included in the model alone, the model was signiﬁcant (Model 1, b = 1.23,

p < 0.001); strong ties were 3.42 times more likely to enact support than weak ties. This is support for

Table 2 Multilevel model for secondary participant resource provision (Binary outcome, N = 586).

b SE t Odds Ratio C.I.

Model 1 Level 2

Bridging −0.19 0.17 −1.11 0.83 (0.59,1.17)

Bonding 0.16 0.14 1.14 1.17 (0.89,1.54)

Level 1

Tie strength

1.23 0.26 4.75 3.42 (2.06,5.68)

Model 2 Level 2

Bridging −0.14 0.18 −0.81 0.87 (0.61,1.24)

Bonding 0.15 0.12 1.21 1.16 (0.91,1.48)

Level 1

FB Contact Freq

0.39 0.06 6.92 1.48 (1.33,1.66)

Model 3 Level 2

Bridging −0.16 0.18 −0.89 0.86 (0.60,1.22)

Bonding 0.15 0.13 1.18 1.16 (0.90,1.49)

Level 1

Tie strength 0.53 0.36 1.50 1.71 (0.85,3.43)

FB Contact Freq

0.29 0.09 3.31 1.34 (1.13,1.60)

Note:

p < .05,

p < .001.

hypothesis 2. However, the effect of tie strength was mitigated when Facebook contact frequency was

considered (Model 3), b = .53, p = .116. When entered separately (Model 2) and when considered

together with tie strength (Model 3), Facebook contact frequency was the most signiﬁcant predictor of

resource provision, supporting hypothesis 3 (b = .39, p < 0.001 in Model 2; b = .29, p < .01 in Model

3). Contact frequency and tie strength together increased the odds of resource provision 1.34 times.

Using contact frequency alone in the model increased the odds of resource provision 1.48 times.

Negative Binomial Regression Analysis. Hypotheses 4 and 5 were tested using a negative binomial

regression. When the dependent variable is a count variable, like the data collected for the current study,

poisson regression is a common analytic approach. However, poisson regression models require strict

adherence to the assumption of dispersion by which the expected mean value should be approximately

equal to the observed variance. When quantifying the equality of variance to the expected mean, it

should be approximately equal to 1 when the residual deviance is divided by the degrees of freedom. If

the result is greater than 1, the data fails to ﬁt the poisson distribution assumption due to the excess

variation. When excess variation is produced from stochastic components, or random errors, the excess

variation is understood as overdispersion. In cases of overdispersion, negative-bionimal regression

analysis produces more robust results compared to poisson regression (for a discussion of issues related

to overdispersion, see Berk & MacDonald, 2008).

Our data demonstrated excess variation. The ratio of residual deviance (347.46) to degrees of

freedom (84) was 4.136, obviously higher than 1. This excess variation occurs from random errors

that are inherent in our experimental design: The 84 cases are part of a multilevel dataset in which

secondary participants are nested in the primary participant’s personal networks. Accordingly, the

excess variation in our data is understood as overdispersion. Thus, a negative binomial regression

analysis was performed. Secondary participants’ sex and age variables (M = 20.66, SD = 6.32) were

entered as controls and tie strength and social status were entered as predictors. The continuous

Table 3 Negative binomial regression model (N = 96).

b SE Wald Wald C.I.

Age −0.01 0.02 0.39 ( −. 06, .03)

Sex * 0.64 0.23 7.18 (.17, 1.10)

Tie Strength 0.43 0.27 2.47 (− .11, .97)

Social Status** 0.45 0.10 20.64 ( .26,.65)

Note: Maximum Likelihood Ratio χ

(4) = 13.22, p = .01, * p < 0.01; **p < .001.

variables age and social status were grand-mean centered. The results indicate that sex (female) had a

positive relationship with the intensity of support given (b = 0.64, p < .01). In other words, female

friends were more supportive in providing their time and effort than male friends. Social status was

also a signiﬁcant determinant of the quality of received support indicating that for a one unit change in

status, the difference in the logs of expected counts of labeled images changed by .45 while holding the

other independent variables constant (p < .001). Thus, only H5 was supported (Table 3).

Discussion

Computer-mediated interaction and online networking sites have enhanced the ability to maintain

a broader spectrum of relationships ranging from the most intimate to extremely superﬁcial. This

exploratory study focused on social capital operationalized as enacted online support and begins to

explore the utility of vast networks articulated via SNSs like Facebook. We investigated perceptions of

online social capital and a series of relationship characteristics to begin explaining the likelihood of

enacted instrumental support by Facebook friends. To our knowledge this is the ﬁrst research to explore

the instrumental utility of online networks using behavior metrics as dependent variables.

We began by exploring the relationship between perceptions of bonding social capital and enacted

support. While extant research suggests that people who believe they have more social capital online

beneﬁt psychologically, the results of the current study suggest that limited instrumental resources may

accrue to such people. While this study limited requests for enacted support to the provision of a service,

the results indicate that relatively few people respond to such low-stakes requests for help. Further, we

failed to show any signiﬁcant relationship between perceived social capital and enacted support. In this

study, the intuition ‘‘I have many good friends from whom I can get help’’ did not explain the ability to

mobilize resources when actually needed. It seems that perceptions about the quality of relationships

may be too general to be linked to actual resource acquisition. Another possible explanation is that

bonding capital and embedded resources may be independent constructs that engage different aspects

of one’s social life, even if they currently share the term ‘‘social capital’’ within the literature. In a sense,

this failure to associate perceived social capital and enacted support mirrors the cleavage within the

existing social capital literature.

As Granovetter (1973) suggests, strong ties are limited in their ability to provide novel resources

because they are likely to be interconnected among themselves, for example family members, thus

circulate redundant information. However, they also represent relationships with greater levels of

investment. In the current study, the request was for a service, not information. As such, it was likely

that strong ties would be more likely to respond to requests for help. It is important to note here that the

weak ties solicited in this study were not as superﬁcial as expected. Participants were given instructions

to think about their six weakest ties, and told that these ties may be with people that they have not

even met to encourage them to think about very weak ties. However, primary participants reported

communicating with their weak ties about three times a week. Regardless, these ties were still far less

likely to provide support when asked.

One interesting ﬁnding is that the amount of interaction that occurs via Facebook explained enacted

support. Facebook contact frequency accounted for the most variance in enacted support, even more so

than emotional closeness. This Facebook-speciﬁc context of social exchange inﬂuenced the likelihood of

receiving enacted support. This ﬁnding is particularly relevant considering the recent work by Hampton

et al. (in press) which suggests that beneﬁts accrue to SNS users even though these people report having

relationships with fewer of their local neighborhood members. The development of mediated networks

characterized by frequent online communication may be replacing people’s traditionally local social

support networks, consistent with Wellman’s (2001) notion of networked individualism.

Although online social spaces are governed by traditional interpersonal communication norms,

technological factors may uniquely affect the dynamics of social exchange and action. Facebook provides

multiple channels for interpersonal and group communication, for example. The outcome of requests

might have been different if the request was solicited through an alternate Facebook communication

channel such as a public wall post or a group message. These channel differences resonate with

earlier theories about social presence in mediated communication, and future research would beneﬁt

by revisiting these technology-oriented theories and examine how technology factors interact with

relationship characteristics.

Finally, we identiﬁed a social status effect on the quality of enacted support. Although tie strength

explained support, the quality of the enacted resources was explained largely by the level of social status

maintained by those requesting help. This can be understood as a kind of interpersonal or social power

within informal groups. In student friendship networks where status (socioeconomic and otherwise) is

rather homogeneous across members, power is determined by a person’s popularity o r attractiveness,

among o ther attributes. Our results suggest those who acquire higher social status are likely to have

greater access to social capital, analogous to other social contexts where authority or high levels of

prestige are associated with enhanced access to resources (Lin, 2001).

Instrumental action is commonly observed online. People routinely receive recommendations,

suggestions, and requests from their friends, community organizers, the commercial sector, and a

range of other third parties. Along with social media practices with which sharing, authoring, and

recommending are major activities, the regime of instrumental action is increasingly expanded online.

Our study explored how speciﬁc relationship characteristics explain such behavior. Because our study

was based on a simple manipulation of the strongest and weakest ties, the investigation could not

capture the inﬂuence of multifarious relational characteristics. Likewise, the measure of social status

was restricted to the relationship between pairs of participants. Status can also be conceptualized more

broadly by measuring a person’s standing at the whole network level.

Further, sending personal requests via e-mail or Facebook messages may not be the most valid

or effective way to determine the accessibility of social capital. For example, response rates would

probably increase if requests for help were made face-to-face, in real time or if the communication was

characterized by more than a single attempt to request help. Although we were not able to measure

whether secondary participants actually received the request for help, the weak ties used in this study

reported communicating with primary participants about three times per week so it is likely they that

did actually receive requests. In addition, although statistical tests conﬁrmed that the strong tie sample

of secondary participants differed from the weak tie group, primary participants were instructed to

think about their six strongest and weakest ties. As noted in the results section, primary participants

indicated that they frequently communicated with their weak ties. Thus, these relationships did not

represent the extremes in terms of relationship strength. Still, the difference between these groups was

statistically meaningful, and strong ties were signiﬁcantly more likely to enact support.

Finally, it is possible that the decision to have primary participants deceive their friends was

problematic. It is possible that these individuals did not maintain the deception for the duration of the

study. However, there were no systematic biases in the distributions of responses received for primary

participants. That is, the response rates were normally distributed, although skewed toward lower

response rates.

Future research would beneﬁt from employing a more systematic sampling approach and test other

relationship criteria. This study was also limited in terms of the support requested. This study mandated

that participants ask their friends to provide a service which required a modest time commitment.

There are differences, however, in the kinds of resources embedded in strong and weak tie networks.

Recall that this study was designed to provide baseline data on the likelihood of resource mobilization

because of the dearth of extant literature on this topic. As such, we used a conservative approach to

operationalizing enacted support and used a low-stakes, low-urgency request.

Future research should explore the differential returns as a function of the type of support requested.

Clearly the results in the current study are limited in that they address provision of support for one

speciﬁc kind of request. Developing continuous dependent variables would also add strength to these

results. To better understand the dynamics of instrumental support and mobilization, continuous

dependent variables like time spent helping might be more useful.

Although some of the measurements used herein could beneﬁt from reﬁnement and the small

sample size for the multilevel analysis is a limitation, this study make a novel contribution to our

understanding of online social networks and provides a foundation for future research in this area. Our

operationalization of instrumental action as a speciﬁc, quantiﬁable social behavior is novel within the

social capital research space, but has parallels with social research in other disciplines such as behavioral

economics. Behavioral research approaches are uniquely suited to the study of social capital, a s online

communication forums are generally amenable to experimental investigation. The forms and practices

of social goal seeking and reciprocation in these spaces like making recommendations, sharing links, and

sending e-mails can be examined and manipulated in a manner that is both controlled and naturalistic.

Note

1 Analyses for multilevel models were performed using the software HLM. SPSS was used for the

negative binomial model.

References

Adler, P. S., & Kwon, S. (2002). Social capital: Prospects for a new concept. The Academy of

Management Review, 27(1), 17–40.

Barrera, M. Jr. (1986). Distinctions between social support concepts, measures, and models, American

Journal of Community Psychology, 14(4), 413–445.

Berk, R., & MacDonald, J. (2008). Overdispersion and poisson regression. Journal of Quantitative

Criminology, 24, 269–284.

Bernard, H. R., Johnson, E. C., Killworth, P. D., McCarty, C., Shelley, G. A., & Robinson, S. (1990).

Comparing four different methods for measuring personal social networks. Social Networks, 12,

179–215.

Blau, P. M., and Duncan, O. D. (1967). The American occupational structure. New York:Wiley.

Coleman, J. S. (1986). Social theory, social research: A theory of action. American Journal of Sociology,

91, 1309–1335.

Coleman, J. C. (1988). Social capital in the creation of human capital. American Journal of Sociology,

94, S95-S120.

Ellison, N. B., Steinﬁeld, C., & Lampe, C. (2007). The beneﬁts of Facebook ”friends:” Social capital and

college students’ use of online social network sites. Journal of Computer-Mediated Communication

12(4), 1143–1168.

Eggens, L., van der Werf, M., & Bosker, R. (2008). The inﬂuence of personal networks and social

support on study attainment of students in university education. Higher Education, 55, 553–573.

Frijda, N. H. (1986). The emotions. London: Cambridge University Press.

Galaskiewicz, J. (1985). Social organization of an urban grants economy: A study of business

philanthropy and nonproﬁt organizations. Orlando, FL: Academic Press.

Gould, R. V. (1991). Multiple networks and mobilization in the Paris Commune, 1871. American

Sociological Review, 56, 716–729.

Granovetter, M. (1973). The strength of weak ties. American Journal of Sociology, 78, 1360–1379.

Granovetter, M. (1974). Getting a job: A study of contacts and careers. Chicago, IL: University of

Chicago Press.

Granovetter, M. (1982). The strength of weak ties: A network theory revisited. In Social structure and

network analysis (pp. 105–130), Sage: Beverly Hills.

Hampton, K. N., Lee, C., & Her, E. J. (In press). How network media affords network diversity: Direct

and mediated access to social capital through participation in local social settings. New Media and

Society.

Hansen, M. T. (1999). The search-transfer problem: The role of weak ties in sharing knowledge across

organization subunits. Administrative Science Quarterly, 44, 82–111.

Hill, R. A., & Dunbar, R. I. M. (2003). Social network size in humans. Human Nature, 14, 53–72.

Homans, G. C. (1950). The human group. New York: Harcourt, Brace & World.

Kadushin, C. (2004). Too much investment in social capital? Social Networks, 26, 75–90.

Klandermans, B. (1984). Mobilization and participation: Social-psychological expansions of resource

mobilization. American Sociological Review, 49, 583–660.

Lin, N., Vaughn, J. C., & Ensel, W. M. (1981). Social resources and occupational status attainment.

Social Forces, 59, 1163–1181.

Lin, N., & Dumin, M. (1986). Access to occupations through social ties. Social Networks, 8, 365–385.

Lin, N., (1982). Social resources and instrumental action. In Marsden, P.V., Lin, N. (Eds.), Social

structure and network analysis. Beverly Hills: Sage, pp. 131–145.

Lin, N. (1999). Social networks and status attainment. Annual Review of Sociology, 25, 467–487.

Lin, N. (2001) Social capital: A theory of social structure and action. Cambridge: Cambridge University

Press.

Marsden, P. V., & Campbell, K. E. (1984). Measuring tie strength. Social Forces, 63, 482–501.

McAdam, D., & Paulsen, R. (1993). Specifying the relationship between social ties and activism.

American Journal of Sociology, 99, 640–667.

Monge, P. R., & Contractor, N. S. (2003). Theories of communication networks. New York: Oxford

University Press.

Moreno, J. L. (1934). Who shall survive? Foundations of sociometry, group psychotherapy, and

sociodrama. Washington, D. C.: Nervous and Mental Disease Publishing.

O’Donovan, A., & Hughes, B. (2007). Social support and loneliness in college students: Effects on pulse

pressure reactivity to acute stress. International Journal of Adolescent Medicine and Health, 19,

523–528.

Opp, K. D., & Gern, C. (1989). Dissident groups, personal networks, and spontaneous cooperation:

The East German revolution of 1989. American Sociological Review, 58, 659–680.

Roberts, S., Dunbar, R., Pollet, T., & Kuppens, T. (2009). Exploring variation in active network size:

Constraints and ego characteristics. Social Networks, 31, 105–164.

Stefanone, M. A., Lackaff, D., & Rosen, D. (2010). The relationship between traditional mass media

and ‘social media:’ Reality television as a model for social network site behavior. Journal of

Broadcasting and Electronic Media, 54(3), 508–525.

Tao, S., Dong, Q., Pratt, M. W., Hunsberger, B., & Pancer, M. S. (2000). Social support: Relations to

coping and adjustment during the transition to university in the people’s republic of china. Journal

of Adolescent Research, 15, 123–144.

Walker, J., Wasserman, S., & Wellman, B. (1994). Statistical models for social support networks. In S.

Wasserman & J. Galaskiewicz, (Eds.), Advances in social network analysis (pp. 53–78). Thousand

Oaks, CA: Sage.

Wellman, B., Carrington, P., & Hall, A. (1988). Networks as personal communities. In B. Wellman &

S.D. Berkowitz, (Eds.), Social structures: A network approach (pp. 130–84). Cambridge: Cambridge

University Press.

Wellman, B. & Wortley, S. (1990). Different strokes from different folks: Community ties and social

support. American Journal of Sociology, 96, 558–588.

Wellman, B., & Gulia, M. (1999). The network basis of social support: A network is more than the sum

of its ties. In B. Wellman (Ed.). Networks in the global village (pp. 83–118). Boulder, CO: Westview

Press.

Wellman, B. (2001a). Physical place and cyber-place: Changing portals and the rise of networked

individualism. International Journal for Urban and Regional Research, 25(2), 227–252.

Wellman, B., & Frank, (2001b). Network capital in a multi-level world: Getting support from personal

communities. In N. Lin, K. Cook, & R. Burt (Eds.), Social capital: Theory and research (pp.

233–273). Hawthorne, NY: Aldine de Gruyter.

Williams, D. (2006). On and off the net: Scales for social capital in an online era. Journal of Computer

Mediated Communication, 11(2), article 11.

About the Authors

Michael A. Stefanone (e-mail: [email protected]), Ph.D., is an assistant professor at the Department of

Communication, University at Buffalo. Dr. Stefanone’s research explores the social psychology of new

media use. His main interest is in group-level computer-mediated communication and Internet-based

communication tools like social network sites.

Address: University at Buffalo, 359 Baldy Hall, University at Buffalo, Buffalo, NY 14260

Kyounghee Hazel Kwon (e-mail: [email protected]), Ph.D., is an assistant teaching professor at the

Culture and Communication Department, Drexel University. Dr. Kwons research areas are commu-

nication technology and social informatics with an emphasis on social networking and collaboration.

Address: Drexel University, Culture & Communication, 3228–50 Powelton Ave, Philadelphia, PA

19104

Derek Lackaff (email: [email protected]), Ph. D., is an assistant professor at Elon University, where

he teaches in the School of Communication’s graduate program in Interactive Media. His research

examines social media technologies and the development of sustainable social, economic, and media

institutions. Address: School of Communications, Elon University, 2850 Campus Box, Elon, NC 27244

Appendix

Table A Descriptive statics and factor loadings for ISCS.

M SD Factor 1 Factor 2

Bridging

Interacting with people on Facebook makes me interested in things

that happen outside of my town.

4.39 1.47 0.728

Interacting with people on Facebook makes me interested in what

people unlike me are thinking.

4.27 1.44 0.696

Interacting with people on Facebook makes me want to try new

things.

3.97 1.45 0.801

Talking with people on Facebook makes me curious about other

places in the world.

4.71 1.61 0.872

Interacting with people on Facebook makes me feel connected to the

bigger picture.

4.5 1.49 0.826

Interacting with people on Facebook reminds me that everyone in the

world is connected.

4.99 1.56 0.725

I am willing to spend time to support community activities occurring

on Facebook.

3.89 1.41 0.691

Interacting with people on Facebook gives me new people to talk to. 3.87 1.71 0.68

On Facebook, I come in contact with new people all the time. 3.53 1.69 0.764

Bonding

There are several people on Facebook I trust to help solve my

problems.

4.62 1.97 0.629

There is someone on Facebook I can turn to for advice about making

very important decisions.

5.14 1.91 0.413

There is no one on Facebook that I feel comfortable talking to about

intimate personal problems.

2.41 1.84 0.609

When I feel lonely, there are several people on Facebook I can talk to. 4.31 1.69 0.341

If I needed an emergency loan of $500, I know someone on Facebook

I can turn to.

3.62 2.23 0.579

The people I interact with on Facebook would put their reputation on

thelineforme.

4.57 1.68 0.502

The people I interact with on Facebook would be good job references

for me.

4.09 1.7 0.583

The people I interact with on Facebook would share their last dollar

with me.

4.4 1.87 0.365

I do not know people on Facebook well enough to get them to do

anything important.

2.78 1.73 0.44

The people I interact with on Facebook would help me ﬁght an

injustice.

4.9 1.67 0.323

Note: Conﬁrmatory factor analysis with varimax rotation yielded 2 factor solution explaining about

57% of the variance.