Organizational restructuring and social network changes implications on work performance, work satisfaction and satisfaction with changes

ORGANIZATIONAL RESTRUCTURING AND SOCIAL NETWORK
CHANGES:
IMPLICATIONS ON WORK SATISFACTION AND SATISFACTION WITH
CHANGES






GRACE LEE CHAU CHIN
(BBA (Hons), MSc (Management), NUS)





A THESIS SUBMITTED
FOR THE DEGREE OF DOCTORATE OF PHILOSOPHY (MANAGEMENT)
DEPARTMENT OF MANAGEMENT AND ORGANISATION
SCHOOL OF BUSINESS
NATIONAL UNIVERSITY OF SINGAPORE
2004

ii

Acknowledgements
This is dedicated to my beloved grandmother who did not live to see the completion
of this academic endeavor, but whose life and example provided the impetus and
motivation for me to prod on and finish the journey. This endeavor would not have
been possible without the encouragement and cheer of family and friends.

To Kelvin, my husband, whose unstinting and unwavering support gave me the
confidence to finish this monumental task.
To Bryan, Brandon and Beverly, whose winning smiles and shining faces never fail to
cheer me up after a long day.
To Usa and Yuen Kay, my cheerleaders, intellectual partners and friends. The race
may not have been completed without you both.
To Associate Professor Albert Teo for supervising the thesis amidst his very heavy
commitments and Associate Professor Quek Ser Aik for selflessly giving of his time
to guide me in the statistical analysis.
To God be the ultimate glory for all the precious lessons that I have learnt about
myself, my own limitations and most of all, His bountiful grace and faithfulness in
sustaining me in this academic journey.

Grace Lee
April 2004

iii
TABLE OF CONTENTS
Acknowledgements …………………………………………………… ii
Table of Contents……………………………………………………… iii
Summary ……………………………………………………………… vi
List of Tables ……………………………………………………….… viii
List of Figures ………………………………………………………… x


Table of Contents

CHAPTER ONE: INTRODUCTION 1

1.1 A Social Network Analysis of Organizations 4

1.2 Research Questions 6
1.3 Aims and Contributions 9
1.4 Outline of Dissertation 10


CHAPTER TWO: LITERATURE REVIEW 11

2.1 The Social Network Structure 11
2.1.1 Network Centrality 14
2.1.2 Network Structural Perspective of Social Capital 16
2.1.3 Structural Holes 19
2.1.4 Types of Network 21
2.1.4.1 Trust Network 25
2.1.4.2 Advice Network 28
2.2 Organizational Restructuring 29
2.3 Work Satisfaction 33
2.4 Employees’ Satisfaction with Organizational Changes 36
2.5 Work Performance 39


CHAPTER THREE: HYPOTHESES DEVELOPMENT 42

3.1 Trust Networks following Organizational Restructuring 42
3.2 Advice Networks following Organizational Restructuring 45
3.3 Structural Context and Work Satisfaction 47
3.4 Network Turnover and Satisfaction with Changes 51
3.5 Minor Hypotheses 56
3.6 Control Variables 61
3.7 The Research Model 63

iv


CHAPTER FOUR: METHODS AND MEASURES 65

4.1 Research Setting 66
4.2 Respondents and Response Rates 70
4.3 Sociometric Measures 71
4.3.1 Pre-restructuring data collection (Phase 1: baseline) 71
4.3.2 Phase 2 data collection (3 months after restructuring) 72
4.3.3 Phase 3 data collection (9 months after restructuring) 74
4.4 Dependent Variables 75
4.4.1 Work Satisfaction and Satisfaction with Changes 75
4.4.2 Work Performance Ratings 76
4.4.3 Peer Assessment Ratings 78
4.5 Network Measures 80
4.5.1 Network Centrality Measures 80
4.5.2 Structural Holes 82
4.5.3 Network Turnover 84
4.6 Coding of Control Variables 84
4.7 Data Analysis 85


CHAPTER FIVE: ANALYSES AND RESULTS 87

5.1 Sample Characteristics 87
5.2 Network Measures and Control Variables 88
5.3 Network Changes after Restructuring 95

5.3.1 Changes in Trust Networks after Restructuring 95
5.3.2 Analysis of Changes in the Trust Networks after Restructuring 98
5.3.3 Changes in Advice Networks after Restructuring 104
5.3.4 Analysis of Changes in the Advice Networks after Restructuring 107
5.4 Changes in Network Positions 110
5.5 Multiple Regressions 113
5.6 Network Determinants of Work Satisfaction 114
5.6.1 Satisfaction with Working in the Organization 114
5.6.2 Satisfaction with Job 118
5.6.3 Satisfaction with Immediate Supervisors 120
5.6.4 Satisfaction with Colleagues 123
5.6.5 Satisfaction with Changes in the Organization 125
5.7 Impact of Network Turnover 128
5.8 Network Determinants of Work Performance 129
5.8.1 Advice Network Centrality and Work Performance 129
5.8.2 Structural Holes and Work Performance 132

CHAPTER SIX: DISCUSSIONS AND CONCLUSIONS 138

6.1 Changes in Social Networks after Restructuring 138
6.2 Network Variables, Work Satisfaction and Satisfaction with Changes 141
6.3 Minor Hypotheses: Network Variables and Work Performance 143
6.4 Limitations of the Study 145
6.5 Implications 146

v

6.6 Theoretical Contributions 148
6.7 Future Directions 150

References 151
Appendices
Appendix 1 168
Appendix 2 171
Appendix 3 175
Appendix 4 178





vi

Summary
While organizational scholars have acknowledged the importance of
understanding how employee attitude, reactions and morale are impacted following
organizational restructuring (Mishra and Spreitzer, 1998), there has been little
attention paid to the mediating effects of social network structure in explaining
employee outcomes. This study seeks to fill this gap in the literature by examining
how changes in employees’ network centrality and structural holes induced by
organizational restructuring impacted employees’ work satisfaction, satisfaction with
changes and work performance.
Network centrality and structural holes of employees’ advice and trust
networks were tracked at three points in time: before, three months after, and nine
months after organizational restructuring. Results indicated that survivors’ trust
network centrality was significantly lower nine months after organizational
restructuring when compared to pre-restructuring levels. In contrast, advice network
centrality and structural holes in survivors’ advice networks were significantly higher
after restructuring compared to pre-restructuring levels.
Work satisfaction comprising four dimensions of employee satisfaction with

respect to: working in organization, job, immediate supervisor and colleagues, were
tracked over three separate points in time: before restructuring, three months and nine
months after. Advice betweenness centrality and structural holes in survivors’ advice
networks were found to be positively associated with satisfaction with working in the
organization three months after restructuring. Results indicated that advice network
centrality and effective size of advice networks had no strong associations with
satisfaction with immediate supervisor and colleagues.

vii

Additionally, employee satisfaction with changes in the organizational
restructuring was tracked at two points in time: three months and nine months after
restructuring. Individuals with higher advice network centrality and whose networks
were richer in structural holes were more satisfied with changes in the organization
three months post-restructuring. However, the relationship between the network
variables and satisfaction with changes were not apparent nine months after
restructuring.
Finally, the effects of trust and advice network turnover on satisfaction with
changes relating to the organizational restructuring were examined. The findings
showed that advice network turnover had a negative impact on satisfaction with
organizational changes. The findings also provided empirical support for the
contribution of network variables to work performance. Specifically, advice network
centrality and structural holes were associated with higher work performance ratings
by supervisors in the post-restructured context respectively.

Keywords: Organizational Change, Organizational Restructuring, Social Networks,
Turnover, Work Performance, Work Satisfaction

viii

List of Tables
Table 2.1: Examples of types of networks in previous studies 25
Table 4.1: Response rates for all Phases 70
Table 4.2: Key Performance Criteria 77
Table 4.3: Performance Rating Guide 78
Table 5.1: Zero-order Correlations for Pre-restructuring data (baseline,
Phase 1)
89
Table 5.2: Zero-order Correlations for Post-restructuring data, three months
after restructuring (Phase 2)
91
Table 5.3: Zero-order Correlations for Post-restructuring data, nine months
after restructuring (Phase 3)
93
Table 5.4: Paired Samples T-Tests (Trust Network Measures) 111
Table 5.5: Paired Samples T-Tests (Advice Network Measures) 112
Table 5.6: Regression Analyses of Advice Network Measures on
Satisfaction with working in the organization before
restructuring (Phase 1)
115
Table 5.7: Regression Analyses of Advice Network Measures on
Satisfaction with working in the organization 3 months after
restructuring (Phase 2)
116
Table 5.8: Regression Analyses of Advice Network Measures on
Satisfaction with working in the organization 9 months after
restructuring (Phase 3)
118
Table 5.9: Regression Analyses of Advice Network Measures on
Satisfaction with Job in the organization before restructuring

(Phase 1)
119
Table 5.10: Regression Analyses of Advice Network Measures on
Satisfaction with Job 3 months after restructuring (Phase 2)
119
Table 5.11: Regression Analyses of Advice Network Measures on
Satisfaction with Job 9 months after restructuring (Phase 3)
120
Table 5.12: Regression Analyses of Advice Network Measures on
Satisfaction with Immediate Supervisors in the organization
before restructuring (Phase 1)
121
Table 5.13: Regression Analyses of Advice Network Measures on
Satisfaction with Immediate Supervisors 3 months after
restructuring (Phase 2)
122
Table 5.14: Regression Analyses of Advice Network Measures on
Satisfaction with Immediate Supervisors 9 months after
restructuring (Phase 3)
122
Table 5.15: Regression Analyses of Advice Network Measures on
Satisfaction with Colleagues before restructuring (Phase 1)
123
Table 5.16: Regression Analyses of Advice Network Measures on
Satisfaction with Colleagues 3 months after restructuring (Phase
2)
124
Table 5.17: Regression Analyses of Advice Network Measures on
Satisfaction with Colleagues 9 months after restructuring (Phase
3)

124

ix

Table 5.18: Regression Analyses of Advice Network Measures on
Satisfaction with Changes 3 months after restructuring (Phase 2)
126
Table 5.19: Regression Analyses of Advice Network Measures on
Satisfaction with Changes 9 months after restructuring (Phase 3)
127
Table 5.20: Regression Models for Network Turnover Rate on Satisfaction
with Changes
128
Table 5.21: Regression Analyses of Advice Centrality Measures on Work
Performance Ratings by Supervisors
130
Table 5.22: Regression Analyses of Advice Centrality Measures on Peer
Performance Ratings
132
Table 5.23: Regression Analyses of Effective Size of Advice Networks on
Work Performance Ratings by Supervisors
133
Table 5.24: Regression Analyses of Effective Size of Advice Network on
Peer Performance Ratings
135
Table 5.31: Summary of Results 135


x

List of Figures

Figure 1: Concepts of centrality (adapted from Krackhardt, 1990) 15
Figure 2: Optimizing structural holes (adapted from Burt, 1992) 20
Figure 3: Possible effects of turnover of friend on stayer (adapted from Krackhardt
and Porter, 1985)
53
Figure 4: Turnover in networks (adapted from Krackhardt and Porter, 1985) 54
Figure 5: Four scenarios depicting degrees of impact from leavers (extracted from
Krackhardt and Porter, 1984, p. 247)
55
Figure 6: Research Model 63
Figure 7: Trust networks before organizational restructuring (baseline, Phase 1) 96
Figure 8: Trust networks three months after organizational restructuring (Phase 2) 97
Figure 9: Trust networks nine months after organizational restructuring (Phase 3) 98
Figure 10: Percentage of ties in the trust networks of all respondents within Work
Groups across All Phases
99
Figure 11: Advice networks before organizational restructuring (baseline, Phase 1) 104
Figure 12: Advice networks three months after organizational restructuring (Phase 2) 105
Figure 13: Advice networks nine months after organizational restructuring (Phase 3) 106
Figure 14: Percentage of ties in the advice networks of all respondents within Work
Groups across All Phases
107


1

CHAPTER ONE
INTRODUCTION

Organizations today are reeling from discontinuities created by a growing
level of globalization, heightened volatility, intense competition and demographic
changes (Prange, 1999). In these times of rapid change, organizations routinely
displace and hire large segments of their workforces in a continuous restructuring
process (Fisher and White, 2000).
Organizations restructure even in the best of economic conditions (McKinley
and Scherer, 2000; McKinley et al., 2000). Organizational scholars suggested that
management frequently uses restructuring to improve productivity and to gain the
favor of Wall Street (Worrell et al., 1991; Ellis, 1998). This generates cognitive order
for executives in turbulent environments but contributes to long-term turbulence at the
environmental level (McKinley and Scherer, 2000). Cognitive order is created when
there is a perceived congruence or “fit” (Drazin and Van de Ven, 1985) between the
organization’s internal structure and the state of the environment. The cognitive order
experienced by top executives and the disruption of environmental conditions provide
the stimulus for further restructuring. Restructuring thus becomes a learned and self-
perpetuating phenomenon (McKinley and Scherer, 2000).
In turbulent environments, top management is driven by pressures to take
some action as opposed to clinging to the existing structure. Farjoun (2001)
suggested that when top management takes no action, it may be interpreted as a sign
of passivity by external and internal stakeholders, and by the executives themselves,
as a relinquishing of control. Therefore, organizations restructure because


2

complacency is penalized and compliance with social expectations is rewarded
(Farjoun, 2001).
Given the widespread practice of organizational restructuring, it is not

surprising that research on restructuring has surged in the past ten years (Bowman et
al., 1999; Mishra and Spreitzer, 1998). Previous empirical work on organizational
restructuring has had a few major focuses. The first is primarily descriptive,
presenting reasons why organizations reorganized, the problems they encountered,
and the forms of reorganized structures (Cascio, 1993; Bruten et al., 1996). The
second focus assesses outcomes of restructuring. Within this stream of research, an
approach commonly used by empirical researchers is to assess whether the financial
and economic objectives of organizational restructuring are achieved. These usually
include expectations of productivity improvement, lower expense ratios, higher
profits, increased return-on-investment, and boosted stock prices (Bowman and Singh,
1993; Bowman et al., 1999). The beneficial effects of restructuring on financial
outcomes are however not always clear. Although some companies have reported
that restructuring spurs organizational performance, others have reported no or even
negative effects (Cascio et al., 1997; Healy, Palepu, and Ruback, 1992; Kaplan, 1989;
Ofek, 1994).
Recently, there have also been a significant number of studies examining non-
financial outcomes of organizational restructuring, specifically its impact on
employee perceptions and reactions (Mishra and Spreitzer, 1998; Amabile and Conti,
1999; Burkhardt, 1994). Cascio (1993) observed that survivors in a restructured
organization experience increased job stress and symptoms of burnout. Fisher (1991)
similarly noted that trust and morale are negatively affected as remaining employees


3

experience workload increases and job insecurity. Directly contrasting these findings
are other studies that showed that employees are energized and consider restructuring
to be an opportunity for personal growth (Emshoff, 1994; Henkoff, 1994; Isabella,
1989). The different and varying results from organizational studies done on
restructuring show that the effects of organizational restructuring on employees are

far from straightforward. It does suggest that more research needs to be done to
further understand what mediates the effects of organizational restructuring on
employees. What is interesting is the lack of attention paid to the mediating effects of
social network structure in explaining employee outcomes.
When an organization restructures by reorganizing current divisional
structures and reporting relationships, the restructuring has a direct influence not only
on the task-related and job-related aspects of employees, but also on where employees
are located within the organization, as well as on the social interactions of individuals
(Reilly, Brett and Stroh, 1993). Changes to the social and relational dimensions of the
organizational context will ultimately have an effect on employee outcomes such as
commitment, work satisfaction and morale (Shah, 2000). In a recent study, Shah
(2000) examined the impact of social structure on survivors’ reactions to layoffs in a
downsizing organization. He suggested that a longitudinal study is necessary to
investigate network effects over different time frames. In addition, he suggested that
more attention should be paid to the impact of structural changes on survivors’ work
performance given that “positional ramifications of a layoff may also have
performance implications” (p.111). Where Shah’s study (2000) examined structural
implications of a downsizing initiative, no study has focused on the implications of a
restructuring initiative that involves changes in administrative and accountability
structures, rather than downsizing per se. Since longitudinal research examining the


4

social structural implications of restructuring is sparse, this study seeks to fill this gap
in restructuring literature by exploring changes in the trust and advice network
structures following organizational restructuring in a longitudinal study. The changes
in employees’ network centrality and structural holes induced by organizational
restructuring were then examined to determine their effects on employees’ work
satisfaction, satisfaction with changes and work performance.

1.1 A SOCIAL NETWORK ANALYSIS OF ORGANIZATIONS
An organization is made up of social groupings with relatively stable patterns
of interaction over time (Weick, 1969). Networks are constructed when individuals
interact (Tolbert et al., 1995). Social networks are defined as “specific sets of
linkages among a defined set of persons, with the additional property that the
characteristics of the linkages as a whole may be used to interpret the social behavior
of the persons involved” (Mitchell, 1962, p.2). Not all pairs of individuals are directly
joined, and some are joined by a variety of relationships. When many individuals are
involved, the resulting structure can be analyzed to derive information about the
individuals or the organizational groupings. The network structure can reflect much
about the functioning of organizations and individuals (McPherson, Popielarz and
Drobnic, 1992).
Network scholars are interested in the patterning of relationship ties within a
social network structure and seek to establish both their causes and consequences. It
is important to understand why relationships exist in some context and are absent in
another (Tichy, Tushman and Fombrun, 1979). There could be two reasons why
relationships are absent. If either party to a relationship does not participate, the
relationship simply does not exist. The second reason why some interactions do not


5

take place is because all interactions occur in a context of institutions which include
rules and roles (Tolbert et al., 1995). Organizational policies impose some of these
rules and roles. For example, organizational units can be explicitly directed to
interact with one unit but not others, or are instructed to report to one level rather than
another. Others might arise from prejudices or rules that are set in place for other
purposes. Therefore, individuals at different levels of the social structure sometimes
do not mix with each other. In some organizations, for example, high-level managers
may not mix with lower-level employees.

Network scholars are particularly interested in mapping informal relationships
and ties. As Tolbert et al. (1995, p. 343) described the realities in organizations:
“The structural designers of organizations, those who mandate reporting
relationships or memo distribution lists or access to databases, are much like
architects who try to predict where the pedestrian traffic will be or should flow
on a university campus. They lay their cement, install fences and other
obstacles, but inevitably the flows of people and classes carve bare spots in the
grass where the sidewalks need to be.”

More often than not, interactions and relationships in organizations consist of
informal ties that link people across formal lines of control. These social interactions
emerge over time, become relatively stable, and may shadow formally prescribed
work-flow and authority relationships. These interactions and relationships form the
social networks within the organization.
Examining organizational dynamics from a network structural perspective has
the advantage of providing understanding and insights into important relationships or


6

ties other than those specified on the organization charts (Krackhardt and Hanson,
1993). Because interactions are often not formally specified, it is important to map
out the presence of informal social networks and trace how they carry information or
ideas from place to place.
A network analysis of organizations not only advances our knowledge of
organizations by the modeling of formal and informal interactions, it provides
additional insights on access to resources within the context of analysis (Brass, 1992).
Coleman (1988) argued that the structural context provides access to and control of
valued resources at any given time.
Over the last decade, network scholars have been keen to understand how

network ties and structural positions that are embedded in both formal and informal
interaction patterns are associated with instrumental outcomes like innovation
opportunities and brokerage opportunities (Burt, 1997; Hargadon, 1997). Given the
merits of utilizing a structural approach to understanding both organizational and
individual phenomena, a study of organization restructuring from a social network
perspective is likely to yield important insights on its impact on the organization and
the consequences on employee-level outcomes.
1.2 RESEARCH QUESTIONS
Social networks are the patterned, repeated interactions among individuals in
the organization (Mintzberg, 1979; Weick, 1969). A major restructuring is likely to
disrupt the existing interaction patterns and relationship ties within the organization as
people leave and newcomers join the organization. Krackhardt and Hanson (1993)
suggested that organizational restructuring creates changes to the informal


7

organization, which they defined as the networks of relationships that employees form
across functions. Social interaction patterns within the organization are likely to be
more fluid than stable in a context such as in an organizational restructuring. Changes
to interaction patterns following organizational restructuring will alter the structures
of social networks within the organization, affecting employees’ centrality within
these social networks and their effective size of networks.
From a social capital perspective, an individual’s centrality in a social network
structure will affect the individual’s ability to draw on that standing to achieve
individual goals (Friedman and Krackhardt, 1997). Social capital theory predicts that
returns to intelligence, education, and seniority depend on a person's location in the
social structure of a market or hierarchy (Granovetter, 1973; Burt, 1997; Lin and
Dumin, 1986; Lin, 1998). Individuals with more social capital get higher returns to
their human capital because they are positioned to identify and develop more

rewarding opportunities (Burt, 1992). What is critical to success is not just individual
attributes like intelligence and charisma, but the way one is embedded in an
organization - that is, one's position in a web of social relations that provide
information and political support (Brass, 1994). Information represents capital that
exists in the very presence of relationships. The more people one knows and the more
informed are that person's contacts, the more information that person has available to
him or her (Burt, 1997).
Based on social capital theory, an individual employee’s centrality in the
social network structure will have implications on individual-level outcomes (Brass,
1984. Ibarra, 1993, Granovetter, 1995; Podolny and Baron, 1997). Despite the
widespread consensus among scholars that informal organizational networks are


8

powerful mechanisms for the control and distribution of a wide array of resources
(Brass, 1992; Krackhardt, 1990), empirical evidence linking resources and outcomes
derived from an individual’s network centrality and structural holes following
organizational restructuring is practically non-existent.
The discussion so far raises interesting questions from both restructuring and
network perspectives. The following summarizes the research questions for this
study:
1. What are the changes within the trust and advice network structures
following organizational restructuring in a longitudinal study?
2. How are employees’ network centrality and structural holes impacted as a
result of changes in the trust and advice network structures?
3. How does organizational restructuring affect employees’ work satisfaction
and satisfaction with changes through the network effects of centrality and
structural holes respectively?
4. How does network turnover affect employees’ satisfaction with

organizational changes?
5. How does organizational restructuring affect employees’ work performance
through the network effects of centrality and structural holes respectively?


9

1.3 AIMS AND CONTRIBUTIONS
This study aims to focus on an organizational restructuring context that
encompasses work reorganization and changes to administrative structures and work
processes. Through this study, I seek to contribute to the network literature by
examining how social networks within the organization are altered following
organizational restructuring in a longitudinal study. A good network theory of
organization should include a longitudinal perspective of how relationships and
interactions change over time (Tolbert et al., 1995). A network analysis taken in a
snapshot of time might miss the organizing that is going on and the stable system that
eventually evolves. A longitudinal study following organizational restructuring will
provide us with a better understanding of changes in social relations within the work
context.
Restructuring provides a context in which to examine how social networks are
affected by formal structural changes. Given that social network studies have
traditionally focused on intact and stable networks (Shah, 2000), an investigation of
how social networks evolve and are changed after organizational restructuring should
enrich the existing network literature that is primarily focused on intact networks.
Previous studies in organizational restructuring have focused on the social-
psychological implications on employees by directly examining changes in employee
outcomes (Brockner et al., 1986, 1987, 1993, 1994; Mishra and Spreitzer, 1998). This
study aims to add to the existing literature by examining the impact of organizational
restructuring on employees through the mediating effects of social network changes.

10

1.4 OUTLINE OF DISSERTATION
I begin Chapter Two with the literature review. In Chapter Three, I will
present the arguments for the hypotheses that will be tested in this study. The
measures and methodology will be presented in Chapter Four. In Chapter Five,
results of the data analyses for each of the proposed hypotheses will be presented. In
Chapter Six, the results and implications, limitations, and suggestions for future
research will be discussed.


11

CHAPTER TWO
LITERATURE REVIEW
This chapter begins with an introduction on features of social networks that
are foundational concepts in the network literature. The concepts of network
centrality and structural holes are given further elaboration because these network
variables are the focal points of interest in this study. Next, the characteristics of the
two types of networks to be examined in this study — trust and advice networks, are
described. The later sections pertain to the review of literature on research that has
been done on organizational restructuring, work performance, work satisfaction and
employees’ satisfaction with organizational changes.
2.1 THE SOCIAL NETWORK STRUCTURE
The social network structure comprises patterned, repeated social interactions
among individuals (Ibarra and Andrews, 1993). In social network theory, the pattern
of relations among a set of actors in a network can explain outcomes related to either
the individuals or the set (Wasserman and Faust, 1994). The fundamental assumption

in social network methods is the idea that individual-level outcomes and perceptions
may be determined by the social network structure.
Network theorists posit that individuals are embedded in social structures that
influence their perceptions of organizational reality and regulate their access to or
control over valued resources (Granovetter, 1973, Burt, 1992). Social influences on
perceptions can stem either from the larger structural context in which individuals are
embedded or from their more immediate social environment (Salancik and Pfeffer,
1978; Contractor and Eisenberg, 1990).


12

Three particular structural features of social networks — structural cohesion,
similarity and centrality are foundational concepts in the network literature (Friedkin,
1991). Formal properties of networks like connectivity and density indicate network
structures that foster social cohesion. For dyads, the structural cohesion of actors A
and B is a function of their joint membership in a structurally cohesive group, not of
their feelings about each other. For example, in a structurally cohesive group where
the members are mutually reachable via communication paths, then actors A and B in
the group are structurally cohesive if they are joint members of such a group.
Members of a cohesive group are likely to be more aware of each other’s views than
are actors who are not members of a cohesive group. Opinion is likely to be more
visible and salient in a cohesive group because the actors are embedded in a context
that values reciprocity and compromise (Friedkin, 1991). The greater the structural
cohesion of A and B, the more likely they will be in frequent communication on a
particular issue. The pressure toward uniformity of opinions that arises from a
comparison of opinions (Festinger, 1954) is likely to occur in a structurally cohesive
group.
Another property of network structure is structural similarity. Actors who
share similar patterns of relationships with others in the network can be grouped

together. A critical assumption in this approach is that roles and positions are
determined by the patterns of relationships actors share with other actors in the
network. An actor A may not be cohesively joined to B in a network but actor B’s
opinions will be salient to actor A if the latter recognizes that they share similar social
positions. Structural similarity may be defined by similar normative circumstances in
the social structure. Actor A will be attentive to structurally similar B’s opinion or
behavior that will have a bearing on A’s status and interests (Burt, 1987). Friedkin


13

(1991) observed that the greater the similarity between two actors’ structural
positions, the more similar their initial opinions on issues are likely to be.
Conversely, opinion differences will take on additional social distance corresponding
to structural dissimilarity in actors’ normative circumstances. Hence, actors will
attach more weight to opinions of actors who are in similar circumstances. The
central argument is that position similarity will lead to homogeneity of attitudes and
behavior appropriate for that social position.
A third property of network structure concerns the center-periphery pattern of
a social network. An actor's embeddedness in a social network has important
implications for a number of outcomes. An actor who has high centrality in a network
is likely to have many direct contacts. With the multiple links he or she has with
others in the network, he or she has ready access to many others in the network.
Employees with high network centrality are likely to be regarded as more powerful
and prestigious than more marginal actors (Scott, 1991), and are more likely to have
disproportionate access to important tangible and intangible resources (Brass, 1992).
Central actors tend to be resourceful and cohesively linked to other actors (Friedkin,
1991)
A structural source of control arises from holding a central position within a
network of relationships (Astley and Sachdeva, 1984; Pfeffer and Salancik, 1978), as

opposed to formally prescribed. Like formal authority, network centrality implies a
high position in a status hierarchy and is expected to enjoy different degrees of access
to and control over valued resources (Burt, 1982). The key difference between
formally and informally derived control is that the latter comes from actors' positions
in the actual patterns of interaction that define a social network, rather than from their


14

positions in the formally defined vertical and horizontal division of labor (Monge and
Eisenberg, 1987).
Network research has examined the influence of two types of actors on
individual behavior: cohesive and structurally equivalent referents. Cohesion
researchers state that individuals are influenced by their direct ties because of the
frequency, intensity, and proximity of interaction (Burt, 1987). Structural equivalence
researchers state that individuals are influenced by others occupying the same position
in their social structure that the individuals themselves occupy (structurally equivalent
referents). However, the focal interest in this study is not behavior mimicry. Neither
is it concern with how similarity between structural equivalents provide each other
with information regarding appropriate role behavior. Based on social capital theory,
the two network measures that have been found to be related to instrumental
outcomes are network centrality (Freeman, 1979) and structural holes (Burt, 1992).
These measures are therefore chosen and discussed in the following sections.
2.1.1 NETWORK CENTRALITY
A multitude of measures of centrality, each slightly different from the others,
has evolved in the social network literature (Freeman, 1979; Bonacich, 1987;
Friedkin, 1991; Knoke and Burt, 1983). Much of the thinking concerning network
centrality has been defined by the work of Freeman (1979), whose measures of
network centrality have been widely used by network researchers. On the basis of
their theoretical relevance and previous use, two centrality measures will be discussed

here: degree and betweenness.


15

Degree centrality is measured as the number of direct links connected to a
given position. The degree measure of centrality may represent the number of
alternatives available to an actor (Friedkin, 1991). In-degree measure of centrality is
defined as the number of others in the network who chooses a focal person. An
individual with high centrality will have more available alternatives than another who
is peripheral. Other measures of centrality include out-degree centrality, which is
defined as the number of people chosen by a focal individual. Although out-degree
centralities have been used, it is widely acknowledged that a more reliable indicator is
in-degree centrality which measures the number of people who choose the focal
individual (Freeman, 1979). For example, an employee may choose many others, but
these others may not reciprocate the choice. In-degree centrality is used as a measure
of prestige on the assumption that relations are often asymmetric. Asymmetric
measures of centrality and in particular, in-degree centrality have been shown to be
related to power in diffusion of innovation (Burkhardt and Brass, 1990; Knoke and
Burt, 1983). Figure 1 portrays the concepts of centrality diagrammatically. In both
networks X and Y, position D has the highest degree centrality because it has the most
number of links connected to a given position.




Network X Network Y
Figure 1: Concepts of centrality (adapted from Krackhardt, 1990)
A

F

D

E
A

F

D

E
B

G

H

I J

E