Human Communication Research ISSN 0360-3989
ORIGINAL ARTICLE
Compliance Dynamics Within a Friendship
Network II: Structural Positions
Used to Garner Social Support
Edward L. Fink1 , Andrew C. High2 , & Rachel A. Smith3
1 Department of Communication, University of Maryland, College Park, MD 20742-7635, USA
2 Department of Communication Studies, The University of Iowa, Iowa City, IA 52242, USA
3 Communication Arts & Sciences, Pennsylvania State University, University Park, PA 16802, USA
Social networks play a critical role in people’s responses to influence attempts, determining
whether a person seeks the support of others as an alternative to compliance or as a way to
cope with being the target of an influence attempt. In 2 experiments (N = 458 and N = 105),
sociograms were used to represent social relationships and to investigate the social network
member who would be sought for social support after an influence attempt. Results showed
that targets were seen as likely to seek social support in more threatening situations and
from more useful (e.g., powerful, connected) network members. Differences found in the 2
experiments appear to represent differences between intergroup and intragroup networks.
doi:10.1111/hcre.12038
The social connections between agents and targets of influence attempts shape
observers’ expectations of how the influence event will ensue. For example, when
agents and targets of influence are members of the same social network, observers
may assume that the target is likely to comply with the agent to avoid rejection by
fellow members of the network (e.g., Schachter, 1951). On the other hand, if a target
of influence is part of a coalition with other network members, this alliance reduces
the expectation of group rejection (Allen, 1966), thereby decreasing the likelihood of
compliance.
In addition to influencing the likelihood of compliance, social networks provide a
means to respond to or help cope with an influence attempt. With the exception of the
literature on minority influence (see Mugny, 1984), social support obtained from network members has received little attention in studies of social influence. The present
study attempts to rectify this omission by considering different rationales for seeking support in a social network. We draw on the social influence, social network, and
social support literatures to examine the process by which the target of an influence
attempt responds by seeking support from network members.
Corresponding author: Edward L. Fink; e-mail: elf@umd.edu
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
21
Influence and Social Support
E. L. Fink et al.
According to Bochner and Insko (1966), targets of influence attempts may respond
to these attempts in a variety of ways, including by obtaining social support from
like-minded others to resist compliance. This resistance may be one mechanism by
which targets of influence attempts separate themselves from those who attempt to
alter their behavior (Latané, 1996). A second way targets may respond to influence
attempts is by turning to other members of their social network for assistance in coping with the distress or anxiety caused by the attempt (e.g., Krause, Goldenhar, Liang,
Jay, & Maeda, 1993; Lewis & Rook, 1999; Schachter, 1959). Because of the constraints
imposed on their behavior (Lewis & Rook, 1999), individuals rely on members of their
social network, especially friends (Toegel, Anand, & Kilduff, 2007), to provide assistance or comfort subsequent to an influence attempt. Here we employ a structural
explanation for social support: Seeking and receiving social support are influenced by
a person’s position in a social network.
Understanding the relationship between compliance dynamics and social support
helps us understand fundamental communication acts such as influencing, resisting
influence, creating coalitions, and engaging in voice (e.g., speaking up or complaining) as opposed to exit (e.g., leaving a group; see Hirschman, 1970). How do influence attempts affect seeking social support? Are there structural predictors of who
becomes a source of social support? Finally, what kinds of social support are sought
and provided? These are the questions that the two experiments reported here seek
to answer.
Influence and social support
People possess a naïve psychology of social influence, and people in every culture
use this naïve psychology to predict how influence attempts and responses to these
attempts proceed. Smith and Fink (2010) showed that participants use information
available in sociograms (i.e., visualizations of social networks), such as the perceived
power of network members, to predict responses to influence attempts. The current
study asks how social networks, as represented in sociograms, shape observers’ expectations of which targets of influence attempts are likely to seek support and which
network members are chosen to be support providers.
Which targets of influence are likely to seek social support?
Burleson and MacGeorge (2002) defined social support as “verbal and nonverbal
behavior produced with the intention of providing assistance to others perceived as
needing that aid” (p. 374). Social support may be sought after major life stressors or
minor hassles (Burleson & MacGeorge, 2002; Hale, Tighe, & Mongeau, 1997), and
support has been associated with a variety of positive outcomes, including emotional
improvement, enhanced self-esteem, resistance to alternative views, and relief from
distress (Burleson & MacGeorge, 2002; Xu & Burleson, 2001). However, little research
has examined influence attempts as stimuli that effect seeking social support or the
network characteristics associated with such activities.
22
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
The characteristics of social networks gain particular importance when the agent
of a stressor is a member of a target’s social network. Communication within a network may produce hardship and stress, and the adverse effects of such interactions
may be stronger than the ameliorative influences of supportive interactions (Rook
& Pietromonaco, 1987). Rook (1990) encouraged researchers to examine both social
support and social strain (i.e., actions by a member of a social network that cause
adverse reactions or relational problems), including attempts at social influence or
control (see also Rook & Pietromonaco, 1987).
Strained social relationships produce depression, dissatisfaction, distress, and
negative affect (DeLongis, Capreol, Holtzman, O’Brien, & Campbell, 2004; Walen
& Lachman, 2000). In contrast, social support, which occurs more frequently than
strain in most social networks, buffers the consequences of social strain (Rook,
1984; Walen & Lachman, 2000). As Walen and Lachman (2000) indicated, “Because
positive and negative aspects of relationships can occur within one network, studying
how they work in tandem is a logical next step, yet is not often done” (p. 25).
The literature on social support and social strain raise important questions. One
question asks about how the social network affects the experience of social support
or strain (Rook, 1990). More specifically, researchers have not investigated the structural characteristics of networks that influence the sources or types of social support
people desire when they are a target of interpersonal influence. We suggest that the
severity of influence attempts (i.e., social strain) and the perceptions of power within
a social network predict which targets of an influence attempt are likely to seek social
support; thus, those targets seeking social support are not a random subset of network
members. Social support may be sought by (1) all targets, (2) less powerful targets, (3)
targets with less supportive impact than the agents have persuasive impact (i.e., more
positive relative influence; see Nowak, Szamrej, & Latané, 1990; Smith & Fink, 2010,
p. 238), and (4) targets with more supportive impact than the agents have persuasive impact (i.e., more negative relative influence). Below we provide the theoretical
rationales for these alternative predictions.
Support sought by everyone
If influence attempts and social strain induce psychological stress by constraining a
target’s behavior (Krause et al., 1993; Lewis & Rook, 1999), then every target may seek
social support. Ensel and Lin (1991) argued that traumatic life events produce distress,
which causes people to mobilize their social networks to provide social support. Consistent with this view, Miller, Smerglia, Gaudet, and Kitson (1998) observed that stress
was positively associated with seeking support from social network members. Anxiety
or stress has also been found to increase affiliation motivation (see Schachter, 1959;
Taylor, 2006), and people faced with an influence attempt may translate this heightened motivation to seek support from members of their social network. Turning to
members of a social network may be an expected response for anyone under duress,
which leads to the following hypothesis:
Alternative H1a: Subsequent to an influence attempt, all targets are equally likely to seek
social support.
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
23
Influence and Social Support
E. L. Fink et al.
Support sought by less powerful targets
If influence attempts generate stress because of power dynamics, then targets with less
power may be especially stressed by the influence attempt and in relatively greater
need of social support. One aspect of a social situation that influences a target’s power
is his or her position within a social network, and network centrality has been found
to predict observers’ perceptions of a target’s power (Smith & Fink, 2010).
There are several different measures of network centrality, and this study focuses
on network members’ eigenvector centrality and betweenness centrality (Borgatti,
2005; Newman, 2010; Wasserman & Faust, 1994). Eigenvector centrality (Bonacich,
1972; Borgatti, 2005) reflects the idea that connections to well-connected others create more and better opportunities and resources than few connections or associations
with less-connected others; eigenvector centrality is an estimate of the number and
quality of a member’s connections. Betweenness centrality emphasizes the benefits
of nonredundant network paths (Burt, 1992). If equal amounts of information flow
between people in a network, and the information takes the shortest path within that
network, betweenness centrality represents the amount of information that flows past
a given person on its way through the network (Freeman, 1979). People with higher
betweenness centrality have greater access to and more control over information
because they are located at the intersection of different regions in a network (Borgatti,
2005).
Less power is associated with less eigenvector and betweenness centrality.
Compared to powerful people, individuals with less power are more constrained
by their environment (Anderson & Galinsky, 2006) and are more vulnerable to
negative actions from others (Allen, 1966). Relatively powerless people consider the
constraints on their behavior (Anderson & Galinsky, 2006), including how those
with power may respond to their actions (Berdahl & Martorana, 2006). Powerless
people have relatively few resources at their disposal and consequently lack sufficient
support to cope with stressors (Litwin & Landau, 2000). Because the constraints
imposed by an influence attempt may generate greater stress for those with low
power (as assessed by network centrality), low power targets may be more likely to
seek support. This view results in the following hypothesis:
Alternative H1b: Subsequent to an influence attempt, the less absolute power the target
has, the more likely the target is to seek social support.
Support sought by targets in positions of positive relative influence
A competing view regarding the targets that are likely to seek support is grounded
in a structural perspective to compliance dynamics. Based on dynamic social impact
theory (DSIT; Nowak et al., 1990), Smith and Fink (2010) defined the relative influence
of an agent over a target as the agent’s persuasive impact minus the target’s supportive
impact. Positive relative influence (i.e., an agent having more persuasive impact than
a target has supportive impact) should result in a target complying to an influence
attempt. In contrast, negative relative influence (i.e., a target having greater supportive
impact than an agent has persuasive impact) should result in target resistance.
24
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
In a situation of positive relative influence, targets may comply with an influence
attempt rather than try to resist it because their relatively lower power (i.e., positive
relative influence) limits their ability to resist the agent. If such compliance dynamics
are stressful, then targets may seek support from network members to cope with
their inability to resist the influence attempt. This view considers social support
as a resource to cope with compliance-related distress, resulting in the following
hypothesis:
Alternative H1c: Subsequent to an influence attempt, the greater the positive relative
influence between agent and target (i.e., the less power the target has relative to the
agent), the more likely the target is to seek social support.
Support sought by targets in positions of negative relative influence
According to Bochner and Insko (1966), targets may respond to influence attempts
by engaging in some form of resistance, such as obtaining “social support from other
like-minded individuals” (p. 614). Social support would then be an asset that targets
can leverage in influence situations. Similarly, some scholars conceptualize social support as a form of social capital that is associated with robust resources and improved
wellbeing (Roustit et al., 2011). In a situation of negative relative influence, when
targets are likely to resist compliance, targets may attempt to garner support from
members of their social networks to resist compliance or even to counterpersuade
the influence agents. This perspective suggests that the target exercises the power of
network position, thereby “flexing one’s structural muscle.” Therefore:
Alternative H1d: Subsequent to an influence attempt, the greater the negative relative
influence between agent and target (i.e., the more power the target has relative to the
agent), the more the target seeks social support.
How does the type of influence attempt affect seeking social support?
In addition to characteristics of the agent and target, the tactic employed in the influence attempt may determine the extent to which participants seek social support.
For example, threats may be viewed as more serious relational transgressions than
attempts at persuasion. “Agents threaten a target when they indicate that they will
make the target’s situation worse if the target does not comply with their request”
(Fink et al., 2003, p. 297, emphasis in original). Threats are acts of dominance that
show indifference to a target’s autonomy, self-esteem, and social value (e.g., Fink et al.,
2003; Kaplowitz, Fink, & Lin, 1998). Thus, threats are messages that are both unconventional (Marwell & Schmitt, 1967) and powerful (Kaplowitz et al., 1998).
In contrast, persuasion focuses on altering someone’s beliefs, attitudes, or behavior
but not “hurt[ing] the target directly as a consequence of that change or lack thereof ” (Fink et al., 2003, p. 297). Persuasion may be perceived to be a fallback position
when agents hold less power than their targets (Smith & Fink, 2010). Fink et al. (2003)
reported that persuasion is considered to be less severe and involves minor relational
transgressions as compared to threats. The level of distress from these two kinds of
influence attempts differs, with the target more upset by a threat than by an attempt
at persuasion. Such considerations are relevant to social support because people are
more likely to seek support when they encounter more severe stressors (Bodie et al.,
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
25
Influence and Social Support
E. L. Fink et al.
2011). Other things being equal, threats should induce more distress, more resistance,
and therefore a stronger desire for support than do attempts at persuasion. The following hypothesis is proposed:
H2: Targets are more likely to seek social support after being subjected to a threat than to
an attempt at persuasion.
Who is likely to be sought for social support?
Research has documented the qualities of support providers and the features of
messages that are valued by support seekers (see Bodie et al., 2011; Burleson, 2003;
Burleson & MacGeorge, 2002; Jones, 2004), but comparatively little is known about
the structural characteristics of social networks that influence the selection of support
providers. The research on seeking support suggests that people prefer to receive
support from individuals who are similar to them in some way (Burleson & Denton,
1992; Burleson & Samter, 1996). The type of stressor also plays a role: Stressors
that involve intimate matters are likely to result in seeking support from confidants,
whereas situations that require a particular skill or specific knowledge are likely
to result in seeking support from trusted experts (Burleson & MacGeorge, 2002;
Cutrona & Suhr, 1994; Sullivan, 1996). Below we discuss three explanations for the
choice of a support provider: structural similarity, sociometric distance, and power.
Structural similarity
As suggested by Bochner and Insko (1966), targets of an influence attempt may turn to
“other like-minded individuals” (p. 614). In the context of a social network, members
who occupy similar structural positions often pay attention to one another because
their opinions, interests, and behaviors have implications for each other (Burt, 1992).
Structurally similar members often monitor each other, compete with one another,
and enact similar behaviors, even if they do not directly communicate with each other
(Abrahamson & Rosenkopf, 1997). These individuals are also likely to experience similar stressors, receive similar information, communicate with similar contacts, and
possess similar social resources. People in structurally similar positions in a social network may be attractive sources of social support because they possess the knowledge
and resources to offer support that matches the circumstances of the target. Accordingly, the following hypothesis is proposed:
Alternative H3a: The greater the similarity in network centrality between the target of an
influence attempt and another network member, the more likely the network member
is a source of social support for the target.
Sociometric distance
A path refers to the distance between two nodes (here, people) in a network (Newman,
2010; Wasserman & Faust, 1994). The shortest path (also called the geodesic path or
geodesic) is the path between two nodes with the fewest edges (i.e., undirected links).
Here we define the number of edges for the shortest path to be the shortest distance.
People with less sociometric distance from targets may be seen as more accessible, and targets may be expected to seek support from a sociometrically close other
26
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
rather than a distant other because people select the path of least resistance or effort
(Zipf, 1949). For example, seeking support from a friend is easier than a friend of a
friend because those directly related are more likely to respond positively (Cai & Fink,
2011; Cai, Fink, & Xie, 2012). DSIT includes a concept called immediacy, which refers
to closeness in space or time (Nowak et al., 1990). One’s sociometric neighborhood
surrounds and protects an individual: Greater immediacy protects minority groups
from being overwhelmed by majority groups. People cluster together to form coherent, bounded entities (Kincaid, 2004; Latané, 1996), and this clustering may reflect
targets turning to sociometrically close others for social support in the face of influence attempts.
People who are structurally closer to targets are viewed as both more accessible
and more similar to the target. In the current study, sociograms are used to describe a
simulated friendship network with ties between network members reflecting mutual
friendship. People who are close in their structural locations are likely to communicate more and become similar to their fellow network members (see, e.g., Berger
& Calabrese, 1975; Fink & Chen, 1995). Conversely, people who hold distinct values
and beliefs are often located further apart in a social network and interact relatively
infrequently.
Research on network evolution suggests that people develop friendships with
those who are like them (e.g., Bearman, Moody, & Stovel, 2004; Kandel, 1978), and
they seek to validate their opinions by communicating with similar others (Festinger, 1954). Similar people gravitate toward each other, thereby creating cohesive
subgroups of individuals located in close proximity (Kincaid, 2004; Latané, 1996).
Because of their accessibility and similarity, sociometrically close network
members may be considered valuable support providers. Their value may be understandable given Cutrona and Russell’s (1990) suggestion that the most effective
social support is provided by people who match their support to their experiences,
resources, and history of interaction with the support recipient (see also Thrasher,
Campbell, & Oates, 2004). Thus, the following hypothesis is proposed:
Alternative H3b: The smaller the network distance between the target of an influence
attempt and another network member, the more likely the network member is a source
of social support for the target.
Power
In general, people form connections with, communicate with, and seek comfort from
people whom they perceive to possess adequate and beneficial resources. Friendships
have costs (Burk, Kerr, & Stattin, 2008), so people avoid undesirable partners and
gravitate toward those whom they perceive to have positive attributes (Bearman et al.,
2004).
Network members with greater perceived power may be attractive support
providers. Those with more power have access to more resources, both material
and social; encounter less interference when pursuing rewards (Anderson & Galinsky, 2006); and have the ability to control others’ access to resources (Berdahl &
Martorana, 2006). People can use their associations with powerful others to secure
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
27
Influence and Social Support
E. L. Fink et al.
benefits (Dutta-Bergman, 2004). If targets use the resources of their social networks
to defend against influence attempts, then members in positions of relatively high
power may be valuable, and thus valued, support providers. Thus, we propose the
following hypothesis:
Alternative H3c: The greater the perceived power of a network member, the more likely
the network member is a source of social support for the target.
Experiment 1
Method
Participants
Participants were 458 undergraduate students from a variety of majors at a large
U.S. university (286 males, 171 females, and 1 with gender unreported). They were
recruited from basic communication courses and received extra credit for their
participation. Their mean age was 20.21 (Mdn = 20.00, SD = 1.85, age range 18–44).
Participants self-identified as Caucasian (87%), Asian (6%), African American (2%),
Hispanic or Latino (2%), Native American (1%), or they did not indicate a racial or
ethnic membership (2%).
Procedures
After receiving approval from the university’s IRB, the participants were asked to complete a Web-based survey, part of which presented simulated sociograms as stimuli.
Such simulations have been found to provide insight into social processes because
people have expectations for the antecedents and consequences of social behavior
(Fink et al., 2003; Kelley, 1992); they apply these expectations to novel and even hypothetical situations, a phenomenon Heider (1958) labeled “common sense psychology.”
To the extent that participants’ beliefs about social situations reflect observed behavior, those beliefs can be interpreted as valid representations of their social reality
(Kelley, 1992).
The simulations used in the two experiments reported here follow the logic
employed by Fink et al. (2003): These designs are “similar to what scientists in other
domains do: Simplify context to clarify process and structure” (p. 312). Social scientists have successfully examined communication using simulations (Stasser, 1988),
hypothetical scenarios (Burleson, 2003), and confederate-based interactions (Jones,
2004), without providing participants with detailed contextual knowledge. The
sociograms used in the current experiments follow this tradition and are used here
to provide insight into participants’ decisions regarding social support subsequent to
an attempt at interpersonal influence.
The sociograms. Sociograms were employed to represent friendships between 11
hypothetical individuals who were described as being in the same university class.
Classmates were identified by letters, and participants were not given any other
information about them. In the sociogram, the classmates were the nodes and their
28
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
Figure 1 Sociogram used in Experiment 1 to represent 11 students in a class, with the circles representing the students and lines representing friendships. Numbers are shown on four
circles to indicate the four positions that could be held by agents and targets in the different
experimental conditions. The two letters identify the two least preferred sources of support.
friendships were the links. The links appeared as undirected lines, so the friendships
were implicitly defined as symmetric.
The letters used to identify classmates were selected because they are the most
commonly used letters in the English language. Following the pilot study reported in
Smith and Fink (2010), the two hypothetical actors (i.e., the agent and target of social
influence) were designated by the letters T and R, which were observed to have equivalent and neutral power in a pilot study (see Smith & Fink, 2010). In this experiment,
six sociograms were used, and T and R were situated in different locations in these
sociograms. Figure 1 shows the four locations used for the two nodes identified as the
agent and target. The sociograms depicted T and R in one of the six combinations of
positions: 1–2, 1–3, 1–4, 2–3, 2–4, and 3–4. Thus, in four sociograms (1–3, 1–4,
2–3, and 2–4) the agent and target nodes were not pictured as directly linked, which
means that they were not considered to be friends. Two variations of each combination pair were generated so that each agent-target order was represented (e.g., T in
position 1 as the agent and R in position 2 as the target versus T in position 1 as the
target and R in position 2 as the agent).
The positions of the agents and targets within this sociogram vary in their centrality. According to Smith and Fink (2010), position 1 in Figure 1 has high eigenvector
centrality but low betweenness centrality; position 2 has high eigenvector centrality and high betweenness centrality; position 3 has low eigenvector centrality but
high betweenness centrality; and position 4 has low eigenvector centrality and low
betweenness centrality.
Experimental design. Each participant answered questions about one sociogram.
Participants were asked to consider an influence attempt between two classmates, T
and R. Participants were presented with one of four statements, which varied in both
order and tactic: Classmate T threatened Classmate R; Classmate T attempted to persuade Classmate R; Classmate R threatened Classmate T; or Classmate R attempted to
persuade Classmate T. The experiment was a 6 (number of sociograms) × 2 (threaten
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
29
Influence and Social Support
E. L. Fink et al.
vs. attempt to persuade) × 2 (agent = R vs. agent = T) between-subjects factorial
design, which generated 24 experimental conditions.
Measures of covariates and dependent variables
Many of the measures described below were calculated from structural positions
within the sociogram and were, therefore, not solicited as perceptions from participants. In particular, measures of network centrality, persuasive impact, supportive
impact, relative influence, network similarity, and shortest path were all calculated
directly from positions within the sociogram. Participants provided measures of
perceived power, responses by the target of the influence attempt, and the probability
that the target would seek social support from network members.
Target’s network centrality. Using UCINET (Borgatti, Everett, & Freeman, 2002),
the normalized betweenness centrality (× 100; M = 29.59, SD = 25.29) and normalized eigenvector centrality (× 100; M = 35.86, SD = 30.67) were computed for the target in each condition. The two centrality measures for the four nodes included in this
experiment were relatively independent (r = .11, ns); thus, both betweenness centrality and eigenvector centrality were used in the analyses presented below.
Agent’s persuasive impact. Drawing on concepts and modified formulae from
DSIT (Nowak et al., 1990), the agent’s persuasive impact (ip ) is pi /di 2 , where pi is
the agent’s perceived power before the influence attempt, and di is the shortest path
distance between the agent and the target.
Target’s supportive impact. Drawing on concepts and modified formulae from
∑
DSIT (Nowak et al., 1990), the target’s supportive impact (is ) is N s 1/2 [ (si /di 2 )/N s ],
where Ns is the number of nodes in the sociogram with exclusive connections to the
target (i.e., paths from nodes that go through the target to reach the agent), si is the
perceived power of the source of support, and di is the shortest path distance between
the target and the source of support.
Relative influence. This variable equals ip – is .
Network similarity. Using UCINET, the normalized betweenness centrality
(× 100; M = 16.85, SD = 20.92) and normalized eigenvector centrality (× 100;
M = 32.50, SD = 26.03) were estimated for each network member in each condition.
In this hypothetical network, the correlation between these two centrality measures
is r = .16 (ns); thus, both scores were used in the analyses. The absolute value of the
difference between the network member’s and the target’s betweenness centrality
(M = 28.89, SD = 22.29) and eigenvector centrality (M = 33.91, SD = 24.91) was calculated, with lower scores indicating greater similarity between their positions in the
network.
Shortest path. The minimal number of edges between the target and each network
member (or shortest path) was assessed. The distance varied from one to six, with
higher scores indicating greater sociometric distance.
30
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
Perceived power. Participants reported their perceptions of each hypothetical
network member’s power using five different statements with response alternatives
ranging from 0 (no power) to 10 (highest power). The five statements corresponded to
French and Raven’s (1959) conceptions of power in small groups (e.g., reward power,
coercive power, referent power, legitimate power, and expert power). The responses
for each person were averaged. Confirmatory factor analyses (CFAs) using maximum
likelihood estimation were conducted for each hypothetical classmate’s power. All
the solutions produced acceptable model fit. For example, the power assessments for
the two main characters, Classmate T: χ2 (5, N = 456) = 8.82, χ2 /df = 1.76, p = .12;
NFI = .99, CFI = 1.00, RMSEA = 0.04, 90% CI [.00, .08]; and Classmate R: χ2 (4,
N = 458) = 5.56, χ2 /df = 1.39, p = .35; NFI = 1.00, CFI = 1.00, RMSEA = 0.02, 90%
CI [.00, .07], fit well; these summaries include all versions of the sociogram. An
overall perceived power score was created by averaging the five responses for the
agent (α = .84) and the target (α = .81), with higher scores indicating more power.
Responses by the target of the influence attempt. Based on Bochner and Insko
(1966), participants were told that the target had at least five different ways (strategies)
to react to the influence attempt: (a) comply with the agent’s request; (b) disparage
(ridicule, discredit, or criticize) the agent; (c) persuade the agent to not want what he
or she wants; (d) obtain social support from similar others; and (e) do nothing; this last
option was not mentioned by Bochner and Insko. Participants were asked to provide
examples, in their own words, of each of these options. These open-ended answers
were evaluated to determine whether the participants understood the different
strategies; in all cases their answers were appropriate to the categories.
Next, participants were asked to indicate the likelihood, from 0% to 100%, that the
target would engage in each of the five different options and then to indicate which
option the target would try first. This experiment focuses on Option 4: obtain social
support. Based on the five options, the likelihood of selecting social support as the
first choice and the overall likelihood of seeking social support (on the 0–100% scale)
were significantly correlated, r(456) = .47, p < .05; the overall likelihood of selecting
Option 4 was used for the analyses below.
Source selection. Participants were asked to indicate the probability, from 0% to
100%, that the target would seek social support from each of the nine remaining members of the network, which excludes the influence agent and target.
Analytic issues
To test AH3a–AH3d, the data first needed to be transposed and restructured to allow
for mixed models (e.g., repeated-measures analyses of covariance, i.e., ANCOVAs).
Participants indicated the likelihood that the target would turn to the remaining network members for social support, excluding the target and agent. The nine remaining
probabilities were restructured as the rows of the data matrix: Each participant had
nine rows of data. Noninteger degrees of freedom are reported for the repeated measures ANCOVAs.
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
31
Influence and Social Support
E. L. Fink et al.
Results
Likelihood of seeking social support
Targets were perceived to have less than a 50% probability (M = 43.40%, SD = 28.61)
of seeking social support after an influence attempt. In contrast to AH1a, not all targets were perceived to have an equal likelihood of seeking support. An overall test
was conducted, with perceived probability of seeking social support as the dependent
variable and message condition as the independent variable; there were 24 conditions.
The overall model was statistically significant, F(23, 434) = 4.14, p < .001, R2 = .18.
The highest means for seeking social support appeared in two conditions: the agent
in position 4 threatening either a target in position 1 (M = 68.50, SD = 22.60) or position 3 (M = 65.47, SD = 19.29). The lowest means for seeking social support appeared
in three conditions: target in position 4 attempting to be persuaded by an agent in
position 1 (M = 25.50, SD = 21.45), position 2 (M = 24.83, SD = 19.82) or position 3
(M = 22.36, SD = 22.99).
AH1b, AH1c, AH1d, and Hypothesis 2 test the factors associated with seeking
social support. To test these hypotheses, a one-way ANCOVA was conducted with
perceived probability of seeking social support as the dependent variable, tactic
(threaten vs. attempt to persuade, H2) as the independent variable, and target’s
perceived power (AH1b) and relative influence (AH1c and AH1d) as covariates. The
model was statistically significant, F(3, 454) = 23.77, p < .001, R2 = .14.
Power
Targets perceived to be more powerful in general were thought to be more likely to
seek social support, unstandardized b = 2.57, SE = 0.50, p < .001. AH1b, which predicted the opposite, was not supported.
Relative influence
AH1c and AH1d made opposing predictions about the role of relative influence in
predicting social support. As predicted by AH1d, relative influence was negatively
related to the likelihood that targets were thought to seek social support (unstandardized b = −0.37, SE = 0.14, p < .01). Targets who had more supportive impact than their
agents had persuasive impact (i.e., negative relative influence) were predicted to be
more likely to seek social support. AH1d was supported; AH1c was not.
Tactic
As predicted by H2, targets were more likely to seek support after an influence attempt
using threats (M = 47.72, SD = 27.83) than persuasion (M = 38.88, SD = 28.78), F(1,
454) = 9.56, p < .01, partial η2 = .02. This hypothesis is supported.
Sources of social support. Alternative Hypotheses 3a–3c (AH3a–AH3c) predicted
the valence of the attributes of network members—similar network centrality, smaller
sociometric distance, and greater perceived power—that may explain who is sought
for social support. A mixed-model ANCOVA for the likelihood of selecting a network
member was conducted, with network member in the simulated friendship network
32
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
being a repeated measure within participants. Member’s similarity for eigenvector
centrality and betweenness centrality (AH3a), sociometric distance (AH3b), and perceived power (AH3c) were covariates; the identifying letter of the network member
(i.e., A, D, E, H, I, L, N, S, and U) was the independent variable.
As predicted by AH3a, network members whose eigenvector centrality
and betweenness centrality were more similar to the target had a greater likelihood of being identified as a source of social support, coefficient = −0.67,
SE = 0.02, t(3608.20) = −37.33, p < .001, R2 = .28, and coefficient = −0.14, SE = 0.02,
t(3705.69) = −6.59, p < .001, R2 = .01, respectively. Contrary to AH3b, greater sociometric distance from the target led to being selected as a source of social support,
coefficient = 2.29, SE = 0.44, t(2722.06) = 5.15, p < .001, R2 = .01. As predicted by
AH3c, network members with greater perceived power had a greater likelihood
of being a source of social support, coefficient = 2.09, SE = 0.23, t(3284.57) = 9.10,
p < .001, R2 = .02.
In addition to these hypotheses, the letter used to identify the network member
was also statistically significant, F(8,1101.06) = 45.00, p < .001, R2 = .25. The pattern
of means was explored to investigate the effects of the network positions associated
with the letters. After adjusting for other model variables, the perceived power
of positions H (M = 14.70, SE = 1.36, 95% CI [12.03, 17.38]) and L (M = 22.11,
SE = 1.38, 95% CI [19.39, 24.83]) had a lower likelihood of being selected to be support providers than other network members. The next lowest member, the person in
position S (M = 30.55, SE = 1.37, 95% CI [28.04, 33.05]), had overlapping confidence
intervals with almost all of the other network members. These results suggest that
some quality about the power associated with the network members in positions H
and L warrants investigation (see Figure 1 for their sociometric location).
A post hoc analysis examined this issue. First, paired-sample t tests were used
to assess systematic differences in the perceived power of the members in positions H
and L in comparison to the target. Although across scenarios position H was perceived
to be less powerful than the target, paired-sample t(456) = −3.54, p < .001, the results
for position L were not statistically significant, paired-sample t(456) = 0.21, ns. These
two positions were not the lowest in eigenvector centrality or betweenness centrality;
the two positions with the least network centrality appear above and below position
3, at the right-hand side of the sociogram. The perceived power of position H was
within sampling error of these two right-hand positions.
Discussion
Experiment 1 investigated the conditions under which participants expected targets
of influence attempts to seek social support and the attributes of the network members
they selected to provide support. We embraced a structural explanation for seeking
support. This experiment found that the likelihood of seeking support was associated
with the relative influence between the agent and target (i.e., negative relative influence), threats more than attempts at persuasion, and more powerful targets. These
findings suggest that social support may bolster resistance to compliance. In other
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
33
Influence and Social Support
E. L. Fink et al.
words, powerful targets may use their structural resources to seek support from others
to maintain their autonomy in the wake of an influence attempt. The network members selected for support are consistent with this argument. Targets were found to seek
support from network members with network centrality similar to that of the target,
with greater sociometric distance from the target, and who were perceived to be more
powerful.
An unanswered question from Experiment 1 concerns the avoidance of two network members, those members identified as H and L. A previously discussed post hoc
analysis showed that these two members did not hold positions of particularly low network centrality nor were they perceived to be the least powerful network members.
These results suggest that other structural qualities make network members attractive
support providers.
Experiment 2
Types of social support
To this point, we have discussed social support as a single resource; however, support
providers can communicate comfort or assistance in different ways, and different situations may call for distinct forms of comfort. These different forms of support, including reappraising stressors, suggesting healthy habits, and offering forms of distraction,
may be beneficial for support seekers (Burleson & Goldsmith, 1998; Callaghan &
Morrissey, 1993; Dewe & Guest, 1990; Xu & Burleson, 2001). This issue was not investigated in Experiment 1.
Structural differences in sociograms
The results of Experiment 1 were based on one sociogram, which has a particular
network structure. In Experiment 2, we used a sociogram that was created to reflect
different patterns of centrality compared to the sociogram used in Experiment
1. Whereas the sociogram in Experiment 1 included nodes that were situated in
positions of high and low betweenness or eigenvector centrality, the sociogram
used in Experiment 2 featured nodes in positions of high degree centrality, high
closeness centrality, and high betweenness centrality. Furthermore, the sociogram
employed in the second experiment has higher correlations between the measures
of centrality than the sociogram used in the first experiment. These differences
represent structural properties of the networks, not subjective perceptions of the
participants.
Intergroup versus intragroup structure
Structural differences between these sociograms may create differences in the way
people respond to hypothetical influence attempts. More specifically, the sociogram
employed in Experiment 1 (see Figure 1) may represent two separate groups, thereby
arousing intergroup tensions, whereas the sociogram used in Experiment 2 (see
Figure 2) represents a single, cohesive group, which may elicit intragroup pressures.
34
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
Figure 2 The sociogram used in Experiment 2 to represent friendship connections between
classmates (credited to Krackhardt, as found in Krebs, 2011).
Turner (1985) noted that “simply imposing a shared group membership on people
can be sufficient to generate attraction between them” (p. 84) and that sharing a
common fate, a shared threat (e.g., a common enemy, intergroup competition), proximity, or similarity is enough to evoke feelings of intragroup belonging or intergroup
tension.
Prior research on group communication affirms that group membership influences its members, causing them to abide by the norms, pressures, and values of the
group (Turner, 1985). Imposing group membership on people, even on an arbitrary
basis, creates intragroup cohesion. Although group members accentuate their differences from and discriminate against members of outgroups, they report more positive attitudes, greater liking, increased similarity, and heightened altruistic motivation
toward members of their ingroup (Howard & Rothbart, 1980; Tajfel, 1981; Turner,
1985).
The goals and interests of group members are often shared, which further promotes intragroup cooperation and harmony (see, e.g., Turner, 1985). Conversely,
intergroup conflict or competition magnifies differences and creates tensions between
groups (Turner, 1982, 1985). Ingroup members are perceived to be more attractive
simply by virtue of their membership in a common group; intergroup tensions create
feelings of depersonalization, reduced attraction, lessened success, and less prestige
for outgroup members (Lott & Lott, 1965; Turner, 1985). Therefore, social networks
that create differences in group entitativity (Campbell, 1958) should create differences
in behavior following an influence attempt. Whereas networks that foster notions
of intergroup tension may generate conflict or resistance to influence attempts,
networks that elicit perceptions of intragroup cohesion may encourage compliance
and social support to manage the stresses of compliance. These ideas can be tested by
examining any differences in results between the two experiments presented here.
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
35
Influence and Social Support
E. L. Fink et al.
Identifying and labeling the network nodes
Another difference between the sociograms in the two experiments is that the
sociogram in Experiment 2 uses names, rather than letters (as in Experiment 1), to
label the nodes in the sociogram; this change may have the advantage of making it
easier for participants to envision an actual friendship network. On the other hand,
the use of names may predispose participants to employ stereotypes associated with
particular names. This issue is discussed below.
Strategies of social support
Barbee and colleagues (e.g., Barbee & Cunningham, 1995; Barbee et al., 1993) used
the term interactive coping to encapsulate the range of supportive behaviors network
members can communicate. These behaviors vary on the dimensions of approach and
avoidance. Approach behaviors, which involve actively and deliberately working to
improve people’s situation, include finding answers to problems (solve) and addressing
a person’s emotional state (solace). Avoidant behaviors, which involve disengagement
to allow distressed people to cope as they see fit, include downplaying the details of
a problem (dismiss) and providing distractions from negative or harmful emotions
(escape).
People generally prefer and react more positively to approach compared to
avoidant support behaviors (e.g., Derlega, Winstead, Oldfield, & Barbee, 2003);
however, this is not a universal preference. Some individuals indicate a preference
for avoidant support, and certain feelings, such as grief or anxiety, which further
complicate stressful situations, correspond to a greater appreciation of avoidant
support (Barbee, Gulley, & Cunningham, 1990; Winstead & Derlega, 1991).
Other research indicates that avoidance is used in strategic ways during conflict
(Wang, Fink, & Cai, 2012), and avoidance can be used in similar ways in episodes that
require supportive communication. People who prefer avoidant support do not want
others to be directly involved with their stressors; rather, they prefer to be distracted
from their problems or desire to be left alone to cope as they see fit. Furthermore,
research on false consensus and social projection shows that participants often assume
that their preferences are shared by others (see Krueger, 2000, for a review). The weight
of the evidence and our naive psychology suggest the following hypothesis:
H4: Following an influence attempt, approach support behaviors are considered
preferable to avoidance support behaviors.
Source selection
Middle-person sources and transitivity
The middle-person position—centered between, in this case, two antagonistic members of a social network—is an attractive source of social support, promoting the
creation of transitive relationships. Transitive relationships are common and occur
when people befriend their friends’ friends. In fact, people rarely maintain indirect
relations with others; they prefer to establish direct relationships with their friends’
friends (Burk, Steglich, & Snijders, 2007). Because there is an impetus for two indirectly connected individuals to develop a friendship, Burk et al. (2008) claimed that,
36
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
based on their research on adolescent school friendships, “reciprocity and transitivity are the most prominent predictors of friendship ties” (p. 517; see also Snijders
& Baerveldt, 2003). Similar ideas follow from balance theory (Heider, 1958). This
research and theory suggest that participants desire to convert indirect network relations into direct ones, thereby establishing mutual friendships.
In some cases, the mutual friend (i.e., the middle person) is the person with the
most knowledge, greatest ability, or strongest likelihood of balancing resources and
restoring harmony after an influence attempt. Middle persons occupy a distinct position with respect to a conflicted dyad because they know both parties; they may be
able to provide comfort or advice from a perspective that represents the views of both
partners (see also Caplow, 1968). Simmel (1950) described one role in a triad as a
mediator, “whose equal distance to both conflicting parties assures his impartiality . . .”
(p. 151), making the mediator a good candidate for providing social support to both
members of a conflicting dyad. Research on supportive communication also emphasizes the importance of relational knowledge in enabling a support provider to tailor
comforting messages (see Burleson, 2003; Burleson & MacGeorge, 2002).
Although Experiment 1 benefited from having agents and targets in positions
varying in network centrality, the sociogram did not allow us to test the attractiveness of middle persons for social support. This idea provides us with another reason
to employ a different sociogram in Experiment 2 (see Figure 2), one with two middle
persons (Fernando and Garth) between the hypothetical agent and target (Diane and
Heather). When an influence attempt occurs between two indirectly linked members,
the middle person may be the most accessible person from whom to seek support and
the most knowledgeable support provider. Thus:
H5: Targets are more likely to select as sources of social support network members who
occupy middle positions between themselves and the influence agent rather than
network members with more distant relations.
Method
Participants
Participants were 105 undergraduate students from a variety of majors at the same
university used for Experiment 1 (54 males, 49 females, and 2 with gender unreported). None of the participants from Experiment 1 participated in Experiment
2. The participants were recruited from basic communication courses and received
extra credit for their participation. Their mean age was 20.76 years old (Mdn = 20.00,
SD = 2.12, Range = 13.00). Participants self-identified as Caucasian (80%), African
American (8%), Asian (3%), Hispanic or Latino (3%), other (2%), or they did not
indicate a racial or ethnic membership (4%).
Procedures
After receiving approval from the university’s IRB, participants completed a
Web-based survey. After agreeing to an online consent form, participants were shown
a simulated social network (Figure 2, the kite formation credited to Krackhardt, as
found in Krebs, 2011).
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
37
Influence and Social Support
E. L. Fink et al.
In a pilot study, a separate set of participants (N = 16) was asked to consider the
names used in this sociogram in combination with a list of eight attributes: power,
reward, punishment, admiration, ability to enforce appropriate behavior, strength,
activity, and goodness. Participants were asked if they thought each name exhibited
more than a typical amount of each different attribute, and we did not observe any
consistent patterns across the names. Specifically, the scores describing power associated with the two names used for the influence inductions (Diane and Heather)
were transformed via an arcsine-root transformation and then compared by using
paired-sample t tests, except when the measure had no variance. There was no statistically significant difference between these two names for their association with any
of the eight attributes. These pilot data suggest that any findings in this experiment
would not be due to stereotypes associated with the names used for the two actors in
the influence attempt.
The procedures and most measures were the same as those used for Experiment 1.
Participants in the main experiment were shown a Web page containing a sociogram,
which had 10 nodes. As in Experiment 1, participants were asked to consider an
influence attempt presented in one of four versions: Diane threatened Heather, Diane
attempted to persuade Heather, Heather threatened Diane, or Heather attempted to persuade Diane. Participants then were asked to indicate the probability that the agent
got what she wanted from the target, and after imagining that they were the target
of the influence attempt, to select network members for support (see Experiment 1
for the exact wording of the questions for compliance and source selection). Unlike
Experiment 1, participants were asked to consider which of Barbee and Cunningham’s
(1995) support strategies they would most like to receive from the network member
they selected for support. We also used an open-ended question to ask participants
to describe the supportive comments, actions, or behaviors they would desire from
members of the social network if they were the target of the influence attempt. Finally,
participants were asked to provide their demographic information.
Instrumentation
Power. Participants reported their perceptions of each hypothetical network
member’s power using the same five items used in Experiment 1 (see Table 1 for αs,
means, and standard deviations). The responses for each person were averaged. CFAs
were conducted to examine each hypothetical network member’s power, and all of
the solutions produced acceptable model fit. For example, the power assessments for
the two main characters, Diane: χ2 (5, N = 105) = 4.91, p = .43; NFI = .98; CFI = 1.00;
RMSEA = 0.00, 90% CI [0.00, 0.13]; and Heather: χ2 (5, N = 105) = 5.30, p = .38;
NFI = .98; CFI = 1.00; RMSEA = 0.02, 90% CI [0.00, 0.14], fit well.
Support strategies. Participants imagined they were the target of the influence
attempt and indicated their preference for different strategies of social support using
the scales developed by Barbee and colleagues (reported in Derlega et al., 2003).
Participants used 5-point scales (1 = not at all; 5 = very much) to complete a total of
20 statements that correspond to approach strategies (solve and solace) and avoidance
38
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
Table 1 Perceived Power, Network Centrality, and Support Measures of Network Members
(N = 105), Experiment 2
Network
Member
Andre
Beverly
Carol
Diane
Ed
Fernando
Garth
Heather
Ike
Jane
Perceived Power
M (SD)
5.84 (1.61)
5.76 (1.58)
5.45 (1.59)
8.08 (1.91)
5.39 (1.56)
6.45 (1.71)
6.50 (1.72)
5.07 (1.73)
3.79 (1.96)
3.14 (2.61)
Network Centrality
Alpha NEigen NBetween Nominate
.88
.87
.88
.80
.86
.86
.86
.90
.88
.91
49.81
49.81
40.42
68.03
40.42
56.24
56.24
27.70
6.80
1.58
2.32
2.32
0.00
10.19
0.00
23.15
23.15
38.89
22.22
0.00
Likelihood
95% Confidence
Interval Around
Likelihood
40.07 (29.49)
39.88 (29.47)
42.87 (26.93)
—
41.08 (26.03)
60.67 (26.84)
58.66 (27.40)
—
40.72 (37.83)
28.30 (33.61)
[34.43, 45.71]
[34.24, 45.52]
[37.72, 48.02]
—
[36.10, 46.06]
[55.54, 65.80]
[53.42, 63.90]
—
[33.48, 47.96]
[21.87, 34.73]
Support
5.61%
15.89%
12.15%
—
4.67%
16.82%
12.15%
—
12.15%
16.82%
Note: Perceived power = participants’ estimates of hypothetical classmate’s power; NEigen
= normalized eigenvector centrality × 100; NBetween = normalized betweenness centrality × 100; Nominate = percentage of participants who nominated the classmate as their top
choice for support; Likelihood = participants’ estimates (means and standard deviations) of
how likely it is that the target would go to the given classmate for support after the influence
attempt.
strategies (dismiss and escape), with five statements devoted to each of the four
strategies.
We found strong, positive associations between the measures of solve and solace
(r = .80, p = .01) and between dismiss and escape (r = .68, p = .01). We also examined
the results of two CFAs. The items measuring preferences for approach support strategies exhibited unidimensionality, with adequate model fit, χ2 (35, N = 103) = 53.27,
χ2 /df = 1.52, p < .05; NFI = .89; CFI = 0.96; RMSEA = 0.07, 90%, CI [0.03, 0.11];
and acceptable reliability (α = .89). Similarly, the items measuring preferences for
avoidant support exhibited unidimensionality, with adequate model fit, χ2 (35,
N = 103) = 66.01, χ2 /df = 1.89, p < .05; NFI = .85; CFI = 0.92; RMSEA = 0.08, 90%
CI [0.04, 0.11]; and acceptable reliability (α = .85). Finally, composite measures of
approach support (M = 3.90, SD = 0.68) and avoidant support (M = 2.27, SD = 0.71)
were created by averaging the relevant items, with higher scores indicating greater
preferences for the support strategies (see also Derlega et al., 2003). Preferences for
approach and avoidant support strategies exhibited a significant negative association,
r = −.50, p < .001, which indicates that the two scales share 25% of their variance.
Results
Preliminary analyses
We first examined participants’ responses to being the target of an influence attempt
according to the options presented by Bochner and Insko (1966). We conducted these
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
39
Influence and Social Support
E. L. Fink et al.
Table 2 Participant Responses to Being the Target of an Influence Attempt From a Member
of a Social Network, Experiment 2
Type of Response
Not seeking social support
Conform
Disparage the source
Persuade the agent
Total
Seeking social support
Informational support
Emotional support
Esteem support
Network support
Tangible support
Total
Other
Number
of
Responses
Percent
of Total
Responses
(N = 452)
Adjusted
Average
Percent of
Responses
(N = 414)
Adjusted Average
Percent of Responses
per Category (Seeking
social support vs. not
seeking support)
14
54
65
133
3.10%
11.95%
14.38%
29.43%
3.38%
13.04%
15.70%
32.12%
10.53%
40.60%
48.12%
99.25%
71
2
1
176
31
281
38
15.71%
0.44%
0.22%
38.94%
6.86%
62.17%
8.40%
17.15%
0.48%
0.24%
42.51%
7.49%
67.87%
—
25.27%
0.71%
0.36%
62.63%
11.03%
100.00%
—
Note: N = 452 responses; the number of respondents = 105. Percent of total responses indicates
the proportion of each category as it occurs in the total sample of responses. Each specific type
of social support is included under the larger category of seeking social support. Adjusted average percent of responses utilizes only those responses in the sample that fit this coding scheme
(38 responses from the total sample did not fit this coding scheme). Adjusted average percent of
responses per category indicates the frequency of each code’s occurrence within its larger category, seeking or not seeking social support. The values for conform, disparage the source, and
persuade the agent are computed from the responses that did not mention social support as
a possibility. The percentages for the specific types of social support are computed from the
subset of the responses that indicated seeking social support was an option. Not all totals equal
100% due to rounding error.
analyses to determine how participants understood their position in the sociogram
and how frequently they would choose to seek social support from individuals within
the simulated network. As indicated in Table 2, 3% of our responses indicated that
they comformity to the influence attempt, 12% reported that they would disparage the
source, and 14% indicated attempts to persuade the agent. In contrast, 62% indicated
seeking social support from other members of the social network.
Those who expressed a desire to seek social support were asked to write out the
supportive comments, actions, or behaviors that they would desire from members
of the social network if they were the target of the influence attempt. These written responses were categorized according to a widely used typology of supportive
communication that differentiates informational, emotional, esteem, network, and
40
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
tangible support (Xu & Burleson, 2001). These types of support are distinguished by
the functional content of the messages or behaviors that are communicated to the
person in distress.
This categorization revealed that 63% of the responses indicated seeking network
support, which involves bolstering people’s social contacts or initiating interpersonal
connections; 25% sought informational support; 11% indicated seeking tangible
support; and fewer than 1% of responses expressed an interest in emotional or esteem
support (see Table 2). These preliminary analyses suggest that our sample members
not only understood their role in the hypothetical social network well enough to
appreciate the benefits of seeking support, but that they also differentiated and
desired distinct types of support.
To assess whether our judgment that the sociograms differed in the way we proposed, we had six communication doctoral students, who were blind to our hypotheses and judgment regarding the sociograms, evaluate the sociograms on a 5-item scale
(with 1 = not at all; 5 = very much for each item), based on items used to measure entitativity in prior research (e.g., Carpenter & Radhakrishnan, 2002; Spencer-Rodgers,
Williams, Hamilton, Peng, & Wang, 2007). Group entitativity measures the extent to
which people perceive a group as containing a unified, cohesive collection of people rather than an independent association of individuals (e.g., “How cohesive is this
group?”; “How unified is this group?”).
There were significant differences in the entitativity of the two sociograms, F(1,
4) = 144.64, p < .001, η2 = .97. In particular, participants perceived the sociogram
used in Experiment 1 to have less entitativity (M = 2.37, SD = 0.46) than the
sociogram used in Experiment 2 (M = 3.87, SD = 0.16). These perceptions of entitativity were not influenced by the order in which people viewed the two sociograms,
F(1, 4) = 3.56, ns, or by the interaction between the sociogram and the order in
which they were viewed, F(1, 4) = 3.50, ns. These results confirm that the sociogram
used in Experiment 2 represents a more unified, cohesive, and intact group than the
sociogram used in Experiment 1.
Likelihood of seeking social support
On average, targets were perceived to have less than a 50% likelihood of seeking social
support from the other network members (M = 44%, SD = 18%); this likelihood replicates that found in Experiment 1 (M = 43%, SD = 29%). The ANCOVA in Experiment
1 was replicated with perceived probability of seeking social support as the dependent variable, tactic (threaten vs. attempt to persuade) as the independent variable,
and the target’s perceived power and relative influence as covariates. In addition, perceived likelihood of compliance was included as a covariate to explore the relationship
between compliance and seeking support.
The model was statistically significant, F(4, 101) = 3.14, p < .05, R2 = .11. The
relationship between the likelihood of compliance and seeking support was positive,
unstandardized b = 0.25, SE = 0.08, p < .01, which suggests that support was sought
to cope with compliance, not as a form of resistance as was found in Experiment 1.
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
41
Influence and Social Support
E. L. Fink et al.
Other findings from Experiment 1 were not replicated: Tactic, the target’s perceived
power, and relative influence did not predict the likelihood of seeking support in
Experiment 2.
Selecting members for social support
A mixed-model ANCOVA for the perceived likelihood of selecting a network member
as a support provider was conducted, with network member as a within-subjects factor. Member’s similarity of eigenvector centrality and betweenness centrality, sociometric distance, and perceived power were included as covariates, and the network
member’s name was the independent variable.
The absolute difference between the member’s and target’s network centrality
was not significantly related to the likelihood of being considered a source of social
support, eigenvector centrality coefficient = −0.01, SE = 0.16, t(497.34) = −0.06, and
betweenness centrality coefficient = 0.17, SE = 0.13, t(497.55) = 1.32. Therefore, the
support for AH3a found in Experiment 1 was not replicated. In addition, contrary
to Experiment 1, AH3b was supported, coefficient for sociometric distance = −21.27,
SE = 2.68, t(491.99) = −7.83, p < .001, R2 = .11: Sociometric distance was negatively
related to source selection. In both experiments, sociometric distance significantly
predicted selecting a network member for support; however, Experiment 1 found
a positive relationship between sociometric distance and source selection, whereas
Experiment 2 found a negative relationship between these variables.
Network members with greater perceived power had a greater likelihood of being
considered a source of social support, coefficient = 1.46, SE = 0.51, t(666.23) = 2.85,
p < .01, R2 = .01. Thus, the support we found for AH3c in Experiment 1 was replicated
in Experiment 2.
Strategies of support
Hypothesis 4 examined whether approach support behaviors were preferable to
avoidance support behaviors following an influence attempt. To test this hypothesis,
a mixed-model ANOVA was conducted with perceptions of the target’s need for
approach and avoidance support as within-subject variables and the top choice for
social support (the person in the sociogram who received the highest likelihood of
being selected for support) as a between-subjects variable. The within-subject effect
was statistically significant, F(1, 95) = 143.07, p < .001, partial η2 = .60: Participants
reported that the target would prefer approach support (M = 3.83, SE = 0.07) more
than avoidance support (M = 2.35, SE = 0.07). These findings support H4. The
between-subjects effect was not statistically significant, F(7, 95) = 0.67, ns, partial
η2 = .05, indicating that the selected support source did not affect these preferences.
In an exploratory analysis, a mixed-model ANCOVA for support preferences
(approach scores minus avoidance scores) was calculated with network member as
a repeated factor within each participant and condition. Tactic was the independent
variable and the covariates were the target-member similarity in perceived power,
target closeness, member’s perceived power, member-agent discrepancy in perceived
power, and relative influence.
42
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
The results showed that tactic, relative influence, target-member similarity, and
member-agent differences in perceived power were significantly associated with
participant’s preferences for approach support over avoidant support. Preferences
for approach support over avoidant support was higher after influence attempts
using threats (M = 1.74, SE = 0.06) rather than persuasion (M = 1.53, SE = 0.06),
t(808.30) = 2.47, p < .05, R2 = .01, and when relative influence was positive, coefficient = 0.06, SE = 0.01, t(748.12) = 4.31, p < .001, R2 = .02. This pair of results
suggests that approach support is especially valued relative to avoidant support in
intense situations that may be difficult to resist, those in which targets are threatened
by agents who are more powerful than they are.
Preference for approach versus avoidance support was also higher when
target-member similarity was higher, coefficient = 0.10, SE = 0.03, t(772.40) = 3.79,
p < .001, R2 = .02; and when the selected network member had greater perceived
power than the agent, coefficient = 0.10, SE = 0.03, t(801.64) = 3.82, p < .001, R2 = .02.
These findings suggest that approach support is preferred when selecting sources that
are socially close and relatively powerful. Overall, the findings from these exploratory
analyses suggest that there is a strategic basis for people’s preference for approach
support over avoidant support.
Middle-person position
Hypothesis 5 predicted that targets would be more likely to select middle people
as sources of social support than people in other positions. To investigate Hypothesis 5, we constructed 95% confidence intervals around participant’s scores for the
likelihood that they would seek each network member for support following the
influence attempt. In particular, we investigated whether participants were more
likely to choose the middle people (i.e., Garth and Fernando) as support providers
than the other members of the social network. An examination of the means and
confidence intervals in Table 1 revealed that participants chose the middle people to
be support providers significantly more often than they chose other members of the
social network. Overall, participants were more likely to rely on network members
who occupy middle positions to be support providers (combined M for Garth and
Fernando = 59.67, SD = 27.12) than members who occupied other positions in the
social network (combined M for other network members = 38.82, SD = 30.56),
unstandardized b = 20.20, t(384.08) = 9.02, p < .001, R2 = .18. These results support
Hypothesis 5 and add nuance to the previously reported finding of Experiment 2 that
participants would seek support from sociometrically close members of their social
network.
Discussion
Experiment 2 extended the investigation of social support after influence attempts
by seeking to replicate and extend the results from Experiment 1 by using a different sociogram and additional measures of social support. In contrast to Experiment
1, relative influence, target’s perceived power, and tactic did not predict likelihood of
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
43
Influence and Social Support
E. L. Fink et al.
seeking support in Experiment 2. One predictor of source selection was replicated:
More powerful network members had a greater likelihood of being considered a support provider. Similarity in network centrality positively predicted supporter selections in Experiment 1, but it was unrelated to selections in Experiment 2. In both
experiments, the sociometric distance between the target and the network member
was a critical predictor in selecting a support provider; however, the association was
positive in Experiment 1 and negative in Experiment 2.
Changing the sociogram in Experiment 2 allowed us to test whether middle persons are attractive sources of support, and they were. By also measuring strategies
of social support, the findings provide some insight into the forms of support that
were sought. Responses from Experiment 2 indicated interest in many types of social
support, with the most interest in network support, which involves bolstering people’s
social contacts. The interest in network support contrasts with other studies reporting
that emotional support is the most common and desired type of support for many situations, including family traumas, academic disappointments, and relationship terminations (Burleson, 2003; Burleson & MacGeorge, 2002). The prominence of network
support in this study may have been influenced by our use of sociograms.
As predicted, targets preferred to receive approach support more than avoidant
support. The relative preference for approach versus avoidant support was stronger
when targets were subjected to threatening (versus persuasive) influence attempts by
relatively more influential agents. These findings suggest that social support may be
sought in order to cope with compliance, and approach support may be most preferred
when the situation seems dire.
Indeed, in Experiment 2 we observed a positive association between the likelihood
of seeking support and complying with the influence attempt. Although speculative,
this pattern of results suggests that by choosing sociometrically close and powerful
network members to provide involving approach support in intense situations, the
participants in this experiment sought to remedy their troubles and ameliorate their
negative affect. The results from Experiment 1 were inconsistent with this view; in
that experiment, social support indicated resistance to influence rather than distress
relief.
General discussion
This study investigated the intersection of two communication processes: responding
to an influence attempt and seeking social support. Five questions were examined:
(a) Who is likely to seek social support?; (b) How do the kinds of influence attempts
(threats vs. persuasion) affect seeking social support?; (b) Who is likely to be considered a support provider?; (d) What are the reasons for seeking support?; and (e) What
kind of support is preferred after an influence attempt?
The two experiments provided answers—some consistent and some inconsistent—to these questions. We surmise that the different sociograms in the two
experiments account for some of the different results that were observed.
44
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
Sociograms as situations: Intergroup and intragroup dynamics
One explanation for the different findings between our two experiments is that the
sociograms represent different structures, and these structures shape people’s needs
and preferences for seeking support (see Goldsmith & Fitch, 1997; Hobfoll & Vaux,
1993). The sociogram used in Experiment 1 represents two cohesive subgroups,
whereas the sociogram from Experiment 2 suggests one highly connected group with
one relatively disconnected member (Jane). Indeed, the density (i.e., total number
of ties divided by the total number of possible ties; Newman, 2010; Wasserman &
Faust, 1994) for the two sociograms differed: It was 25% for Experiment 1 and 40%
for Experiment 2.
Similarly, our pilot study measuring perceptions of entitativity between the
sociograms revealed that the sociogram used in Experiment 2 was perceived to
represent a cohesive, unified group more than the sociogram used in Experiment
1. Thus, the explanation for the discrepancies found between the sociograms may
reflect intergroup versus intragroup dynamics. The likelihood targets would seek
support was almost exactly the same in both experiments (43% in Experiment 1 and
44% in Experiment 2); however, the manner in which they sought support differed
between the experiments.
A network with two possible subgroups (Experiment 1) may evoke notions of
intragroup loyalty and intergroup rivalry, thereby encouraging resistance to persuasion. Sociometric distance was an importance predictor of support provider
selection in both experiments; however, the direction of influence differed in the
two experiments. In Experiment 1, more distant members were more likely to be
chosen; in Experiment 2, closer members were more likely to be chosen. To resist the
influence attempt, participants in one group may have felt that it was in the targets’
best interest to join or establish connections with members of the other coherent
group.
This idea may explain what was observed in Experiment 1: the unexpected positive
association between sociometric distance and the likelihood of being selected to be a
support provider. In four of the six structural pairings that were used to represent the
influence attempt in Experiment 1, the agent and target were positioned in separate
groups, and our participants may have felt compelled to seek support from someone in the same group as the agent to most effectively combat that person’s influence
attempt. Research suggests that people prefer to seek support from more distant associates when they encounter problems that are particularly sensitive or that they are
unable to discuss or resolve with closely connected others (Granovetter, 1973; Wright,
Rains, & Banas, 2010).
The network with a single, tightly knit core (Experiment 2) may evoke notions of
conformity (Schachter, 1951), thereby encouraging compliance and seeking support
to manage compliance-related distress. Tightly knit groups constrain the consideration of alternatives and do not facilitate connections with outsiders (Janis, 1971);
consequently, people may feel compelled to submit to the group and seek support
from sociometrically close members of the network. There may simply be too few
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
45
Influence and Social Support
E. L. Fink et al.
alternatives for support to generate resistance in this tightly knit network; therefore,
the participants opt for conformity.
As predicted, Experiment 2 also found that participants preferred those who
occupy a middle position between the agent and the target of an influence attempt as
a source of social support. Middle people may be in the best position to balance the
interests of two individuals engaged in a dispute and mediate an attempt at interpersonal influence, which allows both parties to save face and continue a relationship.
The results suggest that people seek out their friends’ friends as an extension of their
personal networks and as resources in times of relational strife.
In general, we found that participants prefer approach support behaviors more
than avoidant support behaviors (Derlega et al., 2003). Approach support is an involving form of support that includes finding answers to problems and addressing a person’s affective reaction to stressors (Barbee & Cunningham, 1995; Barbee et al., 1993).
The differences in the types of supportive communication that we categorized in
Experiment 2 also suggest that the sociogram used in this experiment elicits perceptions of conformity. Network support, for example, was the most commonly reported
type of support in this experiment (63% of the support that was listed was for network
support), whereas emotional support was mentioned fairly infrequently (i.e., <1% of
the support that was sought).
In Experiment 2, we also found that informational support and tangible support
were desired. These three types of support—network, informational, and tangible
support—are characterized by an instrumental focus; thus, our results suggest that
participants call on the members of their social network to provide practical support
in an involving, approach-oriented manner subsequent to an interpersonal influence
attempt.
The lack of emphasis on emotional support is inconsistent with prior research that
claims emotional support is the most commonly desired type of support and is suitable for coping with many stressors (see Burleson, 2003, 2008; Burleson & MacGeorge,
2002). The combination of the topic of this experiment (i.e., an attempt at interpersonal influence by a member of the participants’ purported social network) and the
structure of the network (i.e., one with a single cohesive group) appears to have created
a low desire for emotional support.
Differences in influence
The differences in network structure may explain why the perceived likelihood
of seeking social support was related to features of the influence attempt (the
tactic used and the relative influence of the agent and the target) in Experiment
1 but not in Experiment 2. The features of the influence attempt may be more
salient in situations that provoke resistance rather than elicit relief. In Experiment 1, targets that had more supportive impact than their agents had persuasive
impact were found to be more likely to seek support. The likelihood of seeking
support was also higher when agents used threats rather than persuasion. This
power relationship and influence tactic were negatively related to compliance in
46
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
Smith and Fink (2010); in that study, obtaining social support was an alternative to
compliance.
The principles underlying interpersonal influence may differ in intragroup versus intergroup situations. In intragroup situations, compliance and social support are
positively related because people seek support to cope with the stress resulting from
the attempt at interpersonal influence by members of their own group. In intergroup
situations, in which ingroup and outgroup pressures are more common, compliance
and social support are negatively related because people turn to others to resist interpersonal influence from outgroup members.
The findings support these ideas. The structure of the sociograms used in
Experiment 1 and Experiment 2 differed in two main ways: Experiment 1 only
included one condition in which the target and agent were separated by a single
middle person, whereas that was always the case in Experiment 2. Experiment 2
also represented a single group, whereas Experiment 1 pictured a less dense network
with two subgroups. Beyond differences in structure, Experiment 1 used letters to
label the nodes, whereas Experiment 2 used names. It is possible that using names
enhanced perceptions of intimacy within the network and lessened the desire to resist
influence.
Although the two experiments in this study revealed some differences in the way
that supportive communication functions, prior research in this area has also had
conflicting results: Some studies have reported that social support has no effect in
buffering the individual from the effects of social strain, and some have reported a
positive effect of social support. In addition, a reverse buffering effect has been found,
in which social strain is more strongly linked to depression when support is readily
available (see Walen & Lachman, 2000, for a review).
These conflicting results—ours and those found in other studies—lead us to conclude that more research needs to be done to understand why people seek social
support from members of their social network subsequent to an influence attempt.
Social support may function as a source of resistance to compliance (as indicated by
Experiment 1) or as a source of relief from the distress that may accompany compliance (as indicated by Experiment 2). Future research can identify the structural factors
regarding social support, as well as other factors (e.g., personal, relational, and contextual) that might help explain this process (see Burleson, 2009) and most effectively
lead to a positive buffering effect of supportive communication.
Limitations
A few issues limit these findings. First, the sociograms did not include isolates (i.e.,
individuals without ties to other network members). Isolates may be sought for
social support because they are outside of, and possibly irrelevant to, network-based
power dynamics. Second, the samples of participants used in this study are not
diverse in age or ethnicity, which may influence attention to social dynamics, experiences with them, or decision-making about them. However, as Courtright (1996),
Shapiro (2002), and others have reminded us, the nonprobability sample may be less
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
47
Influence and Social Support
E. L. Fink et al.
important to scientific explanation and development than the careful investigation of
communication processes and focused replications.
This study is also limited because we did not examine the intrapersonal
mediators—evaluations and other cognitions—that may help explain the participants’ decisions. In their review of group theories and experiments, Collins and
Raven (1968) pointed out that many experiments have shown that people often look
to others to evaluate their own responses (see also Schachter, 1959). We think it is
important to understand not only how social networks facilitate resistance to and
coping with influence attempts, but also how networks shape targets’ thoughts and
behaviors, which are also shaped by the reactions, behaviors, and suggestions of the
other network members.
Conclusion
Social influence and its associated responses are processes in which social context and
communication are profoundly interrelated. Our findings indicate that position in a
social network shapes perceptions of power and plays a critical role in how participants think about responding to influence (Smith & Fink, 2010). These experiments
document differences in participant’s reactions to interpersonal influence, including
the use of social support—a notion presented by Bochner and Insko almost fifty years
ago but never tested. We showed that social context and compliance dynamics shape
who is sought for support and the kinds of support that are sought.
The findings indicated a preference for approach support and a consistent, strong
preference for powerful support providers; other determinants varied between the
two experiments. These results expand our knowledge about supportive communication by including the structural factors involved in the identification and selection of
support providers. Understanding the intersection of social relationships and social
behavior provides critical insights that relate communication to influence, power,
social support, and group structure.
Bateson (as reported in Watzlawick, Beavin, & Jackson, 1967) wrote:
When one octopus or one nation puts on a threatening gesture, the other might
conclude “he is strong” or “he will fight,” but this was not in the original message. (p.
101)
The goal of our study has been to understand how such conclusions are made—for
example, “he is strong”—and the ongoing effects that these conclusions have for people. Our results show some progress toward this human goal; generalizations about
octopodes and nations are currently beyond our reach.
Acknowledgments
The authors wish to thank Dr. Anita Atwell Seate for her assistance with this article.
48
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
References
Abrahamson, E., & Rosenkopf, L. (1997). Social network effects on the extent of innovation
diffusion: A computer simulation. Organizational Science, 8, 289–309.
doi:10.1287/orsc.8.3.289
Allen, V. L. (1966). Situational factors in conformity. Advances in Experimental Social
Psychology, 2, 133–175. doi:10.1016/S0065-2601(08)60105-7
Anderson, C., & Galinsky, A. D. (2006). Power, optimism, and risk-taking. European Journal
of Social Psychology, 36, 511–536. doi:10.1002/ejsp.324
Barbee, A. P., & Cunningham, M. R. (1995). An experimental approach to social support
communications: Interactive coping in close relationships. Communication Yearbook, 18,
381–413.
Barbee, A. P., Cunningham, M. R., Winstead, B. A., Derlega, V. J., Culley, M. R., Yankeelov, P.
A., & Druen, P. B. (1993). Effects of gender role expectations on the social support
process. Journal of Social Issues, 49, 175–190. doi:10.1111/j.1540-4560.1993.tb01175.x
Barbee, A. P., Gulley, M. R., & Cunningham, M. R. (1990). Support seeking in close
relationships. Journal of Social and Personal Relationships, 7, 531–540.
doi:10.1177/0265407590074009
Bearman, P., Moody, J., & Stovel, K. (2004). Chains of affection: The structure of adolescent
romantic and sexual networks. American Journal of Sociology, 110, 44–99.
doi:10.1086/386272
Berdahl, J. L., & Martorana, P. (2006). Effects of power on emotion and expression during a
controversial group discussion. European Journal of Social Psychology, 36, 497–509.
doi:10.1002/ejsp.354
Berger, C. R., & Calabrese, R. J. (1975). Some explorations in initial interaction and beyond:
Toward a developmental theory of interpersonal communication. Human
Communication Research, 1, 99–112. doi:10.1111/j.1468-2958.1975.tb00258.x
Bochner, S., & Insko, C. A. (1966). Communicator discrepancy, source credibility, and
opinion change. Journal of Personality and Social Psychology, 4, 614–621.
doi:10.1037/h0021192
Bodie, G. D., Burleson, B. R., Gill-Rosier, J., McCullough, J. D., Holmstrom, A. J., Rack, J. J., ...
Mincy, J. (2011). Explaining the impact of attachment style on the impact of supportive
messages: A dual-process framework. Communication Research, 38, 228–247.
doi:10.1177/0093650210362678
Bonacich, P. (1972). Factoring and weighting approaches to status scores and clique
identification. Journal of Mathematical Sociology, 2, 113–120.
doi:10.1080/0022250X.1972.9989806
Borgatti, S. P. (2005). Centrality and network flow. Social Networks, 27, 55–71.
doi:10.1016/j.socnet.2004.11.008
Borgatti, S. P., Everett, M. G., & Freeman, L. C. (2002). UCINET for windows: Software for
Social Network Analysis. Harvard, MA: Analytic Technologies.
Burk, W. J., Kerr, M., & Stattin, H. (2008). The co-evolution of early adolescent friendship
networks, school involvement, and delinquent behaviors. Revue Française de Sociologie,
49, 499–522.
Burk, W. J., Steglich, C. E. G., & Snijders, T. A. B. (2007). Beyond dyadic interdependence:
Actor-oriented models for co-evolving social networks and individual behaviors.
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
49
Influence and Social Support
E. L. Fink et al.
International Journal of Behavioral Development, 31, 397–404.
doi:10.1177/0165025407077762
Burleson, B. R. (2003). Emotional support skill. In J. O. Greene & B. R. Burleson (Eds.),
Handbook of communication and social interaction skills (pp. 551–594). Mahwah, NJ:
Lawrence Erlbaum.
Burleson, B. R. (2008). What counts as effective emotional support? Explorations of
situational and individual differences. In M. T. Motley (Ed.), Studies in applied
interpersonal communication (pp. 207–227). Newbury Park, CA: Sage.
Burleson, B. R. (2009). Understanding the outcomes of supportive communication: A
dual-process approach. Journal of Social and Personal Relationships, 26, 21–38.
doi:10.1177/0265407509105519
Burleson, B. R., & Denton, W. H. (1992). A new look at similarity and attraction in marriage:
Similarities in social-cognitive and communication skills as predictors of attraction and
satisfaction. Communication Monographs, 59, 268–287. doi:10.1080/03637759209376269
Burleson, B. R., & Goldsmith, D. J. (1998). How the comforting process works: Alleviating
emotional distress through conversationally induced reappraisals. In P. A. Andersen & L.
K. Guerrero (Eds.), Handbook of communication and emotion: Research, theory,
applications, and contexts (pp. 245–280). San Diego, CA: Academic Press.
Burleson, B. R., & MacGeorge, E. L. (2002). Supportive communication. In M. L. Knapp & J.
A. Daly (Eds.), Handbook of interpersonal communication (3rd ed., pp. 374–424).
Thousand Oaks, CA: Sage.
Burleson, B. R., & Samter, W. (1996). Similarity in the communication skills of young adults:
Foundations of attraction, friendship, and relationship satisfaction. Communication
Reports, 9, 127–139. doi:10.1080/08934219609367645
Burt, R. S. (1992). Structural holes: The social structure of competition. Cambridge, MA:
Harvard University Press.
Cai, D. A., & Fink, E. L. (2011). Intercultural networks. In G. A. Barnett (Ed.), Encyclopedia of
social networks (Vol. 1, pp. 419–421). Los Angeles, CA: Sage.
doi:10.4135/9781412994170.n164
Cai, D. A., Fink, E. L., & Xie, X. (2012). “Brother, can you spare some time, or a dime?”: Time
and money obligations in the U.S. and China. Journal of Cross-Cultural Psychology, 43,
592–613. doi:10.1177/0022022111401394
Callaghan, P., & Morrissey, J. (1993). Social support and health: A review. Journal of
Advanced Nursing, 18, 203–210. doi:10.1046/j.1365-2648.1993.18020203.x
Campbell, D. T. (1958). Common fate, similarity, and other indices of the status of aggregates
of persons as social entities. Behavioral Science, 3, 14–25. doi:10.1002/bs.3830030103
Caplow, T. (1968). Two against one: Coalitions in triads. Upper Saddle River, NJ: Prentice Hall.
Carpenter, S., & Radhakrishnan, P. (2002). The relation between allocentrism and
perceptions of ingroups. Personality and Social Psychology Bulletin, 28, 1528–1537.
doi:10.1177/014616702237580
Collins, B. E., & Raven, B. J. (1968). Group structure: Attraction, coalitions, communication,
and power. In G. Lindzey & E. Aronson (Eds.), The handbook of social psychology (2nd
ed., Vol. 4, pp. 102–204). Reading, MA: Addison-Wesley.
Courtright, J. A. (1996). Rationally thinking about nonprobability. Journal of Broadcasting &
Electronic Media, 40, 414–421. doi:10.1080/08838159609364362
50
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
Cutrona, C. E., & Russell, D. W. (1990). Types of social support and specific stress: Toward a
theory of optimal matching. In B. R. Burleson, I. G. Sarason, & G. R. Pierce (Eds.), Social
support: An interactional view (pp. 319–366). New York, NY: John Wiley.
Cutrona, C. E., & Suhr, J. A. (1994). Social support communication in the context of
marriage: An analysis of couples’ supportive interactions. In B. R. Burleson, T. L.
Albrecht, & I. G. Sarason (Eds.), Communication of social support: Messages, interactions,
relationships, and community (pp. 113–135). Thousand Oaks, CA: Sage.
DeLongis, A., Capreol, M., Holtzman, S., O’Brien, T., & Campbell, J. (2004). Social support
and social strain among husbands and wives: A multilevel analysis. Journal of Family
Psychology, 18, 470–479. doi:10.1037/0893-3200.18.3.470
Derlega, V., Winstead, B., Oldfield, E. C., & Barbee, A. P. (2003). Close relationships and
social support in coping with HIV: A test of sensitive interaction systems theory. AIDS
and Behavior, 7, 119–129. doi:10.1023/A:1023990107075
Dewe, P. J., & Guest, D. E. (1990). Methods of coping with stress at work: A conceptual
analysis and empirical study of measurement issues. Journal of Organizational Behavior,
11, 135–150. doi:10.1002/job.4030110205
Dutta-Bergman, M. J. (2004). An alternative approach to social capital: Exploring the linkage
between health consciousness and community participation. Health Communication, 16,
393–409. doi:10.1207/s15327027hc1604_1
Ensel, W. M., & Lin, N. (1991). The life stress paradigm and psychological distress. Journal of
Health and Social Behavior, 32, 321–341. doi:10.2307/2137101
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 57, 217–282.
doi:10.1177/0018732675400700202
Fink, E. L., Cai, D. A., Kaplowitz, S. A., Chung, S., Van Dyke, M. A., & Kim, J.-N. (2003). The
semantics of social influence: Threats vs. persuasion. Communication Monographs, 70,
295–316. doi:10.1080/0363775032000179115
Fink, E. L., & Chen, S. S. (1995). A Galileo analysis of organizational climate. Human
Communication Research, 21, 494–521. doi:10.1111/j.1468-2958.1995.tb00356.x
Freeman, L. C. (1979). Centrality in networks: Conceptual clarification. Social Networks, 1,
215–239. doi:10.1016/0378-8733(79)90002-9
French, J. R. P., & Raven, B. (1959). The bases of social power. In D. Cartwright (Ed.), Studies
in social power (pp. 150–167). Ann Arbor: University of Michigan Institute for Social
Research.
Goldsmith, D. J., & Fitch, K. (1997). The normative context of advice as social support.
Human Communication Research, 23, 454–476. doi:10.1111/j.1468-2958.1997.tb00406.x
Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78,
1360–1380. doi:10.1086/225469
Hale, J. L., Tighe, M. R., & Mongeau, P. A. (1997). Effects of event type and sex on comforting
messages. Communication Research Reports, 14, 214–220.
doi:10.1080/08824099709388663
Heider, F. (1958). The psychology of interpersonal relations. New York, NY: John Wiley &
Sons. doi:10.1037/10628-000
Hirschman, A. O. (1970). Exit, voice and loyalty: Responses to declines in firms, organizations,
and states. Cambridge, MA: Harvard University Press.
Hobfoll, S. E., & Vaux, A. (1993). Social support: Social resources and social context. In L.
Goldberger & S. Breznitz (Eds.), Handbook of stress: Theoretical and clinical aspects (2nd
ed., pp. 685–708). New York, NY: Free Press.
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
51
Influence and Social Support
E. L. Fink et al.
Howard, J., & Rothbart, M. (1980). Social categorization and memory for in-group and
out-group behavior. Journal of Personality and Social Psychology, 38, 301–310.
doi:10.1037/0022-3514.38.2.301
Janis, I. L. (1971, November). Groupthink. Psychology Today, 5, 43–46, 74–76.
Jones, S. M. (2004). Putting the person into person-centered and immediate emotional
support: Emotional change and perceived helper competence as outcomes of comforting
in helping situations. Communication Research, 31, 338–360.
doi:10.1177/0093650204263436
Kandel, D. B. (1978). Homophily, selection, and socialization in adolescent friendships.
American Journal of Sociology, 84, 427–463.
Kaplowitz, S. A., Fink, E. L., & Lin, Y. (1998). Speaking loudly and carrying a big stick: The
effect of power tactics and structural power on perceptions of the power user. Research in
Political Sociology, 8, 103–119.
Kelley, H. H. (1992). Common-sense psychology and scientific psychology. Annual Review of
Psychology, 43, 1–23.
Kincaid, D. L. (2004). From innovation to social norm: Bounded normative influence. Journal
of Health Communication, 9, 37–57. doi:10.1080/10810730490271511
Krause, N., Goldenhar, L., Liang, J., Jay, G., & Maeda, D. (1993). Stress and exercise among
the Japanese elderly. Social Science and Medicine, 36, 1429–1441. doi:10.1016/
0277-9536(93)90385-H
Krebs, V. (2011). Graphic illustration of a sociogram developed by David Krackhardt. Kite
Network. Retrieved from http://www.orgnet.com/sna.html
Krueger, J. (2000). The projective perception of the social world. In J. Suls & L. Wheeler
(Eds.), Handbook of social comparison: Theory and research (pp. 323–351). New York,
NY: Kluwer Academic/Plenum.
Latané, B. (1996). Strength from weakness: The fate of opinion minorities in spatially
distributed groups. In E. Witte & J. H. Davis (Eds.), Understanding group behavior:
Consensual action by small groups (pp. 193–219). Mahwah, NJ: Lawrence Erlbaum.
Lewis, M. A., & Rook, C. S. (1999). Social control in personal relationships: Impact on health
behaviors and psychological stress. Health Psychology, 18, 63–71.
doi:10.1037/0278-6133.18.1.63
Litwin, H., & Landau, R. (2000). Social network type and social support among the old-old.
Journal of Aging Studies, 14, 213–228. doi:10.1016/S0890-4065(00)80012-2
Lott, A. J., & Lott, B. E. (1965). Group cohesiveness as interpersonal attraction: A review of
relationships with antecedent and consequent variables. Psychological Bulletin, 64,
259–309. doi:10.1037/h0022386
Marwell, G., & Schmitt, D. R. (1967). Dimensions of compliance-gaining behavior: An
empirical analysis. Sociometry, 30, 350–364. doi:10.2307/2786181
Miller, N. B., Smerglia, V. L., Gaudet, D. S., & Kitson, G. C. (1998). Stressful life events, social
support, and the distress of widowed and divorced women: A counteractive model.
Journal of Family Issues, 19, 181–203. doi:10.1177/019251398019002004
Mugny, G. (1984). The influence of minorities: Ten years later. In H. Tajfel (Ed.), The social
dimension: European developments in social psychology (Vol. 2, pp. 498–517). Cambridge,
England: Cambridge University Press.
Newman, M. E. J. (2010). Networks: An introduction. New York, NY: Oxford University Press.
52
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
E. L. Fink et al.
Influence and Social Support
Nowak, A., Szamrej, J., & Latané, B. (1990). From private attitude to public opinion: A
dynamic theory of social impact. Psychological Review, 97, 362–376.
doi:10.1037/0033-295X.97.3.362
Rook, K. S. (1984). The negative side of social interaction. Journal of Personality and Social
Psychology, 46, 1097–1108. doi:10.1037/0022-3514.46.5.1097
Rook, K. S. (1990). Parallels in the study of social support and social strain. Journal of Social
and Clinical Psychology, 9, 118–132. doi:10.1521/jscp.1990.9.1.118
Rook, K. S., & Pietromonaco, P. (1987). Close relationships: Ties that heal or ties that bind?.
In W. H. Jones & D. Perlman (Eds.), Advances in personal relationships (Vol. 1, pp. 1–35).
Greenwich, CT: JAI.
Roustit, C., Campoy, E., Renahy, E., King, G., Parizot, I., & Chauvin, P. (2011). Family social
environment in childhood and self-rated health in young adulthood. BMC Public Health,
11, 949–957. doi:10.1186/1471-2458-11-949
Schachter, S. (1951). Deviation, rejection, and communication. Journal of Abnormal and
Social Psychology, 46, 190–207. doi:10.1037/h0062326
Schachter, S. (1959). The psychology of affiliation. Stanford, CA: Stanford University Press.
Shapiro, M. A. (2002). Generalizability in communication research. Human Communication
Research, 28, 491–500. doi:10.1111/j.1468-2958.2002.tb00819.x
Simmel, G. (1950). The sociology of Georg Simmel. Glencoe, IL: The Free Press.
Smith, R. A., & Fink, E. L. (2010). Compliance dynamics within a simulated friendship
network I: The effects of agency, tactic, and node centrality. Human Communication
Research, 36, 232–260. doi:10.1111/j.1468-2958.2010.01375.x
Snijders, T. A. B., & Baerveldt, C. (2003). A multilevel network study of the effects delinquent
behavior on friendship evolution. Journal of Mathematical Sociology, 27, 123–151.
doi:10.1080/00222500305892
Spencer-Rodgers, J., Williams, M. J., Hamilton, D. L., Peng, K., & Wang, L. (2007). Culture
and group perception: Dispositional and stereotypic inferences about novel and national
groups. Journal of Personality and Social Psychology, 93, 525–543.
doi:10.1037/0022-3514.93.4.525
Stasser, G. (1988). Computer simulation as a research tool: The DISCUSS model of group
decision making. Journal of Experimental Social Psychology, 24, 393–422.
doi:10.1016/0022-1031(88)90028-5
Sullivan, C. F. (1996). Recipients’ perceptions of support attempts across various stressful life
events. Communication Research Reports, 13, 183–190. doi:10.1080/08824099609362085
Tajfel, H. (1981). Human groups and social categories. Cambridge, England: Cambridge
University Press.
Taylor, S. E. (2006). Tend and befriend: Biobehavioral bases of affiliation under stress. Current
Directions in Psychological Science, 15, 273–277. doi:10.1111/j.1467-8721.2006.00451.x
Thrasher, J. F., Campbell, M. K., & Oates, V. (2004). Behavior-specific social support for
healthy behaviors among African American church members: Applying optimal matching
theory. Health Education and Behavior, 31, 193–205. doi:10.1177/1090198103259184
Toegel, G., Anand, N., & Kilduff, M. (2007). Emotion helpers: The role of high positive
affectivity and high self-monitoring managers. Personnel Psychology, 60, 337–365.
doi:10.1111/j.1744-6570.2007.00076.x
Turner, J. C. (1982). Towards a cognitive redefinition of the social group. In H. Tajfel (Ed.),
Social identity and intergroup relations (pp. 15–40). Cambridge, England: Cambridge
University Press.
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association
53
Influence and Social Support
E. L. Fink et al.
Turner, J. C. (1985). Social categorization and the self-concept: A social cognitive theory of
group behavior. In E. J. Lawler (Ed.), Advances in group processes (Vol. 2, pp. 77–121).
Greenwich, CT: JAI Press.
Walen, H. R., & Lachman, M. E. (2000). Social support and strain from partner, family, and
friends: Costs and benefits for men and women in adulthood. Journal of Social and
Personal Relationships, 17, 5–30. doi:10.1177/0265407500171001
Wang, Q., Fink, E. L., & Cai, D. A. (2012). The effect of conflict goals on avoidance strategies:
What does not communicating communicate? Human Communication Research, 38,
222–252. doi:10.1111/j.1468-2958.2011.01421.x
Wasserman, S., & Faust, K. (1994). Social network analysis: Methods and applications.
Cambridge, England: Cambridge University Press.
Watzlawick, P., Beavin, J. H., & Jackson, D. D. (1967). Pragmatics of human communication: A
study of interactional patterns, pathologies, and paradoxes. New York, NY: W. W. Norton.
Winstead, B. A., & Derlega, V. J. (1991). An experimental approach to studying social
interaction and coping with stress among friends. In W. H. Jones & D. Perlman (Eds.),
Advances in personal relationships (Vol. 2, pp. 107–131). Greenwich, CT: JAI.
Wright, K. B., Rains, S., & Banas, J. (2010). Weak-tie support network preference and
perceived life stress among participants in health-related, computer-mediated support
groups. Journal of Computer-Mediated Communication, 15, 606–624.
doi:10.1111/j.1083-6101.2009.01505.x
Xu, Y., & Burleson, B. R. (2001). Effects of sex, culture, and support type on perceptions of
spousal social support: An assessment of the “support gap” hypothesis in early marriage.
Human Communication Research, 24, 535–566. doi:10.1111/j.1468-2958.2001.tb00792.x
Zipf, G. K. (1949). Human behavior and the principle of least-effort. Oxford, England:
Addison-Wesley.
54
Human Communication Research 41 (2015) 21–54 © 2014 International Communication Association