‘ : I» r :lv ‘

'll

 

 

 

 

LIBRARY
Michigan State
University

This is to certify that the
thesis entitled

The Role of Social Capital in lnterorganizational Alliances

presented by

Branda L. Nowell

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Psychologx

 

 

ZAL((_/ £§er ZQUWCL'A

Major Professor” 5 Signature

Date

 

MSU is an Affinnative Action/Equal Opportunity Institution

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

, NOV 2 5 2007
4““; a; r- tit
‘ “W

_MAY_1L5_ZOJOI
DEC 0 6 ZUIO

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 pdClRC/DateDqudd—pJ

 

THE ROLE OF SOCIAL CAPITAL IN
INTERORGANIZATIONAL ALLIANCES

By

Branda L. Nowell

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Psychology

2006

\V

ah
rel;
has
ana
fon

6X8

 

 

8?

pr;

 

ABSTRACT
THE ROLE OF SOCIAL CAPITAL IN INTERORGANIZATIONAL ALLIANCES
By
Branda L. Nowell

In response to increasing demands for greater coordination and collaboration
among community institutions, interorganizational alliances (IAs; i..e., coalitions,
collaboratives, coordinating councils) have emerged as a popular mechanism for
strengthening the capacity of a community system to respond public and social issues.
While there is ample theory and conceptual arguments present in the literature to suggest
that the nature of relationships among stakeholders may constitute a form of social capital
that can facilitate the functioning and effectiveness of IAs, there is a dearth of knowledge
about what types of relationships are most important in the context of IAs and what role
relationships among members actually play in influencing IA effectiveness. This study
has sought to address this gap by using key informant interviews and social network
analysis techniques to operationalize social capital within the context of one prominent
form of IA — domestic violence coordinating councils (DVCCs) - and empirically
examine its relationship to indicators of DVCC effectiveness.

Social network and survey data were collected from 638 members belonging to 48
different DVCCs located across the state of Michigan. Results found that the overall
amount of social capital among DVCC stakeholders was indeed a significant and positive
predictor of the extent to which a DVCC was perceived to be effective at both improving
the level of coordination among organizations within the existing community system as

well as making needed changes to the infrastructure of the community response system

liC

SU

 

le‘

[3']

SV

inl

bu

m}

ma

 

 

itself. However, results indicate that the role of social capital in the context of DVCCs is
not the same for coordination as it is for system change outcomes. Specifically, findings
suggest that systems change outcomes may be more strongly impacted by the overall
level of social capital among stakeholders relative to coordination outcomes. Further, the
types of relationships that are most important also differ between coordination and
systems change outcomes. In particular, the extent to which stakeholders perceive one
another to share a common philosophy concerning what domestic violence is and how it
should be addressed appears to be uniquely and strongly related to systems change.
Overall, findings suggest that when the work of a DVCC calls for not just improving how
information and resources flow through the existing domestic violence response system,
but rather actually making changes to the infrastructure of the system itself, the DVCC
will likely need to foster stronger relationships among participating stakeholders than
may have been needed for improving coordination. These relationship-building efforts
should focus particular attention to identifying what differences in philosophy may exist
among members about the issue of domestic violence and work toward building more

shared frameworks of understanding.

ACKNOWLEDGEMENTS

I would first like to acknowledge and sincerely thank all of the DVCC leaders and
members who took time away from the incredibly important work that they do to
participate in this study. Their passion and commitment for the work was inspiring to be
around are. I would particularly like to thank my community partners, Mary Keefe,
Tammy Lemmer, and Michelle Bynam for being so generous with their time and
expertise in developing and carrying out the study. Thank you as well to Susan
Ramspacher, Kathy Tarplay, Zoe Nylun and the members of their councils for all their
help and feedback in piloting the survey and to the Michigan Nonprofit Research
Program and the Blue Cross Blue Shield Foundation of Michigan for their financial
support.

I am also deeply grateful for the support of my committee and faculty advisors.
Thank you to my committee chair Pennie F oster-F ishman for the countless hours of
patient mentoring, for her seemingly endless passion and energy for the work, and for
being my role model in what it means to be a scholar who never settles for good enough.
It has been a great honor to have had the opportunity to be her student. Thank you to
Debby Salem for her brilliant critical mind, for always calling it like she sees it, and for
her unwavering support throughout my graduate career. Thank you to Deb Bybee for so
selflessly giving of her time and talents and for being my ever-patient statistical spotter
and teacher. Thank you to Cris Sullivan for her critical role in helping to open the doors
that made this study possible and for sharing with me her knowledge of and passion for

the issue of violence against women. Thank you to Kevin Ford for always challenging

iv

 

 

 

CT.

Ab
bei:
ma

all)

 

me while supporting me 100%. Thank you as well to Ken Frank for introducing me to
the world of social network analysis and to Nicole Allen, whose work was a great source
of inspiration for this study, for being such a willing and reflective sounding board.

This project would not have been possible without the efforts of my fabulous
team of undergraduate research assistants: Courtney Dana, Marvin Valencia, Matt
VanderMeullen, Kelly Holcombe, Cory Mullins, and Jessie Wong. I would especially
like to thank Deborah Satka, Katie Suva, and Brian Miedlar who were with the project
from beginning to end and gave 110% the entire way. Thank you as well to my fellow
graduate students who have been my companions and co-learners on this journey. I am
particularly thankful for my cohort Adrienne Adams, Marisa Sturza, Nate Thomas, and
Aisha Smith and my ever-willing partner in crime Shelby Berkowitz for their gifts of
insight, support, sanity, and most of all, humor.

Finally, I want to thank my family and fiiends outside of academia for all their
encouragement and for tolerating me through all the periods of academic self-absorption
when I wasn’t able to be the daughter, sister, or friend that I would have wanted to be.
Above all things and in all things, I want to thank my husband and best friend Jason for
being my constant unwavering champion; for believing in me with such intensity as to
make it almost impossible for me to not believe in myself even in the darkest moments
along this journey. I also would like to thank him for being my greatest teacher in life’s

most important subjects, the arts of happiness, generosity, and balance.

TABLE OF CONTENTS

OVERVIEW ............................................................................................... 1
The Relationship between Social Capital and IA Effectiveness ........................ 3

The Conditions under which Social Capital is Most Strongly

Related to IA Effectiveness ................................................................. 4
IA Outcomes for which Social Capital Matters Most ................................... 5
Summary ....................................................................................... 6
INTRODUCTION ..................................................................................... 7
DVCC Effectiveness ........................................................................ 8
The Theoretical Relevance of Social Capital to DVCC Effectiveness .............. 14
The Role of Social Capital in Promoting DVCC Effectiveness ..................... 27
Composition Effects on the Role of Social Capital .................................... 33
Outcomes .................................................................................... 37
Summary .................................................................................... 39
RESEARCH DESIGN AND METHODS ........................................................ 41
Overview .................................................................................... 41
Procedures ................................................................................... 41
Measurement Development ............................................................... 51
Measures .................................................................................... 56
Data Analysis ............................................................................... 69

vi

 

 

 

CC

AP

 

 

RESULTS ............................................................................................. 73

Univariate Analysis ......................................................................... 73
Modeling Perceptions of DVCC Effectiveness ........................................ 86
Exploring Interactions between Composition Diversity and Social Capital ....... 89

Exploring Relative Importance of Different Indicators of Social Capital..... . . ...100

DISCUSSION ...................................................................................... 107

The Role of Social Capital to DVCC Effectiveness at Promoting

Community Change ...................................................................... 107
Organizational Capacity .................................................................. 112
The Interaction of Social Capital with Diversity ..................................... 114
Relative Importance of Different Indicators of Social Capital ..................... 117
Limitations ................................................................................. 123
Implications for Future Research ....................................................... 125
CONCLUSION ...................................................................................... 128
APPENDICES ......................................................................................... 132
Appendix A: Leader Recruitment letter ................................................ 133
Appendix B: Leader Interview Handout ............................................... 135
Appendix C: Leader interview protocol ............................................... 138
Appendix D: Stakeholder Categories .................................................. 147
Appendix E: Member recruitment letter ............................................... 148

vii

Appendix F: DVCC Member survey ................................................... 150
Appendix G: Follow up reminder ...................................................... 165
Appendix H: DVCC Effectiveness at promoting community change ............. 166
Appendix I: Qualitative Analysis for Strengthened Organizational

Capacity Scale ............................................................................ 166

REFERENCES ..................................................................................... 1 73

viii

Tab

Tab

 

 

Tat

Tat

Tat

Tat

Tat

Tat

Iat‘

 

 

LIST OF TABLES

Table 1: Response Rates for Levels I and II ..................................................... 46
Table 2: Correlation Matrix: Indicators of Social Capital ..................................... 63
Table 3: One-way ANOVAs for Mail versus Web Survey Respondents .................... 70
Table 4: Summary of Model Constructs ......................................................... 72
Table 5: Descriptives on Aggregated Dependant Variables ................................... 76
Table 6: Descriptives on Independent Variables ................................................ 77
Table 7: Descriptives on Independent Variables (cont) ....................................... 78
Table 8: Descriptives on Transformed Independent Variables ................................ 79

Table 9: Number of Activities Council Spent Between 1% and 20% of
time (out of 10 possible) ................................................................. 81

Table 10: Number of Activities Council Spent 20% or More of Time

(out of 10 possible) ..................................................................... 81
Table 11: Estimate of Covariance Parameters for Coordination Effectiveness. . . . . . . . .....82
Table 12: Estimate of Covariance Parameters for Systems Change Effectiveness. . . . .84
Table 13: Estimate of Covariance Parameters for Organizational Learning ................ 85
Table 14: Estimate of Covariance Parameters for Enhanced Opportunity .................. 86
Table 15: Relationship of Social Capital to Coordination Effectiveness .................... 88
Table 16: Relationship of Social Capital to Systems Change Effectiveness ................ 89
Table 17: Correlations of Diversity Indicators with Social Capital and Outcomes... . .....90

Table 18: Analysis of Gender Diversity Interaction for Coordination Effectiveness ...... 92

Table 19: Analysis of Gender Diversity Interaction for Systems Change

ix

 

 

I2

Te
Ta
Ta
Ta

Ta

Ta]

Ta]

Tu

 

 

 

 

Table 20:

Table 21 :

Table 22:

Table 23:

Table 24:

Table 25:

Table 26:

Table 27:

Table 28:

Table 29:

Table 30:

Table 31:

Effectiveness ............................................................................. 93

Analysis of Tenure Diversity Interaction for Coordination Effectiveness ...... 94
Analysis of Tenure Diversity Interaction for Systems Change
Effectiveness ............................................................................. 94
Analysis of Management Diversity Interaction for Coordination
Effectiveness ............................................................................. 95
Analysis of Management Diversity Interaction for Systems Change
Effectiveness ............................................................................. 96
Analysis of Sector Diversity Interaction for Coordination Effectiveness ....... 97
Analysis of Sector Diversity Interaction for Systems Change Effectiveness. . .98
Analysis of Breadth of Stakeholders for Coordination Effectiveness ............ 99
Analysis of Breadth of Stakeholders for Systems Change Effectiveness ........ 99
Independently Modeled Indicators of Social Capital Predicting
Coordination Effectiveness ........................................................... 101
Combined Model Predicting Coordination Effectiveness ....................... 103
Independently Modeled Indicators of Social Capital Predicting
Systems Change Effectiveness ...................................................... 104
Combined Model Predicting Systems Change Effectiveness ................... 105

 

 

 

SCH

inn

hea

firs

UIIC

red

Co

hur

bui

fur

 

C0

COI

 

(1':

 

 

OVERVIEW

There is a clarion call for greater coordination and collaboration among human
service organizations and agencies — particularly for those who share a common
investment in a given social issue. The appeal of collaborative approaches for tackling
health and human service issues appears to stem primarily from two perspectives. The
first is a demand for collaboration as a means to improve what is commonly viewed as an
uncoordinated and fragmented service delivery system fraught with expensive
redundancies and ineffective programming (Hoge & Howenstine, 1997; McLaughlin &
Covert, 1984). The second perspective is concerned with promoting the stability of the
human service sector and looks to interorganizational collaboration as a key strategy for
building the capacity of service organizations to deal with environments characterized by
high turbulence, greater demands from consumers and communities, and shrinking
funding sources (Bailey & Koney, 2000; Huxham, 1996).

lnterorganizational alliances (IAs) have become a prominent response to this call
in communities throughout the United States. Referred to by many names (e. g.,
coalitions, coordinating councils, collaboratives), IAs are community-based groups
comprised of leaders and staff representing both nonprofit and for profit organizations as
well as public agencies who share a common issue domain (e. g., mental health services,
domestic violence, HIV/AIDS). These groups meet regularly for the purpose of
identifying and implementing strategies for improving the community’s response to their
targeted social issue (Foster Fishman, Salem, Allen, & Fahrbach, 2001). They are
formed based on the premises that adequately addressing the social issue for which they

are convened is beyond the scope of any one organization or agency and that an effective

 

 

 

 

 

 

COII

Hac

Alli

org

CXl:

all:

up"

 

Clll

0U

 

sci

LC

SC.

— —(j

ll'L'

 

community response to the issue will therefore require coordination and collaboration
among community stakeholders (Allen, Watt, & Hess, 2003; Fowler, Donegan, Lueke,
Hadden, & Phillips, 2000; Gamache & Asmus, 1999; Glisson & James, 1992; Sullivan &
Allen, 2001; Wischnowski & McCollum, 1995). However, because community
organizations and agencies have at least some degree of autonomy in determining the
extent to which - and manner in which - they will interact with other organizations or
agencies, such collaborative approaches depend upon participating organizations to
develop cooperative interorganizational relationships. Cooperation as it is used in the
present study refers to actions that involve actors working together toward a common
goal or in one another’s interests (Alter & Hage, 1993). The presence of cooperative
interorganizational relationships has commonly been identified as critical to the success
of interorganizational collaboration and subsequently a key factor in predicting IA
effectiveness at addressing its targeted social issue (Alter & Hage, 1993; Foster-Fishman,
Berkowitz, Lounsbury, Jacobson, & Allen, 2001; Mattessich, Murray-Close, & Monsey,
2001).

Within a community service system, the presence of strong interorganizational
cooperative relationships that can be used as a resource for accomplishing productive
outcomes such as service coordination is referred to by sociologist and organizational
scientists as social capital (Baker, 2000; Cohen & Prusak, 2001; Coleman, 1988;
Leenders & Gabbay, 1999). Coleman (1990) — one of the most prominent social capital
theorists (Schuller, Baron, & Field, 2001)- defines social capital as the value within
social-structural relationships that an “actor” such as an individual, an organization, or a

network of organizations (e.g., an IA) can mobilize to make possible the ‘achievement of

certain ends that would not be attainable in its absence’ (p. 302). While it has received
limited attention by researchers interested in IAs, social capital has become a concept of
increasing interest to scholars and practitioners interested in interorganizational
collaboration within the for-profit sector as it directly concerns how relationships within
an interorganizational network impact the ability of that network to utilize cooperation
among members as a resource for accomplishing its goals (Baker, 2000; Cohen &
Prusak, 2001; Coleman, 1988; Leenders & Gabbay, 1999). IAs are formed on the
premise that the development of cooperative relationships among community agencies
and organizations invested in a given social issue will improve the community’s response
to that issue (Hart, 1995; Morgan, Guetzloe, & Swan, 1991; Rosenkoetter et al., 1995).
As such there is good reason to expect that social capital has great relevance to the
effectiveness of IAs though this has never been systematically examined. This study
addresses this gap by examining social capital’s role within IAs including (1) how social
capital relates to IA effectiveness, (2) under what conditions it is most strongly related to
IA effectiveness, and (3) for which outcomes of effectiveness social capital is most
critical.
The Relationship Between Social Capital and IA Effectiveness

The first question of interest in this study is whether the level of social capital
among members within an IA is positively related to the effectiveness of the IA at both
creating positive community change and strengthening the organizational capacity of IA
members. Effectiveness for this study is defined as the extent to which an IA is
perceived as effective by its members at both improving its community’s response to the

targeted social issue as well as strengthening the capacity of its organizational members.

The premise underlying this examination posits that the nature of work for which IAs are
convened requires strong interorganizational relationships. Specifically, IAs are formed
to address complex social issues that are beyond the scope, skills and resources of any
single organization or agency working alone. As such, improving the community’s
response to such social issues is argued to depend upon a substantial amount of
interorganizational coordination and collaboration (Bryson & Crosby, 1992; Bailey &
Koney, 2001). Social capital is posited to be highly facilitative of efforts requiring
coordination and collaboration on the part of multiple organizations due to its ability to
reduce the risk associated with interorganizational c00peration (Adler & Kwon, 2002).
This study provides an empirical investigation of this premise by directly examining the
relationship between social capital and IA effectiveness.
The Conditions Under Which Social Capital is Most Strongly Related
To IA Effectiveness
A second area of interest in understanding the role of social capital in IAs
concerns the conditions under which the relationship between social capital and IA
effectiveness is strengthened or weakened. One such condition that has gained particular
attention in theory and research within the organizational science and social capital
literatures is the diversity of the membership composition. Scholars have posited that
social capital, due to its ability to promote cooperation, can positively impact the
effectiveness of groups such as IAs, in part, by increasing the accessibility of resources
(e.g., knowledge, skills, connections, supplies) that can be applied to any given problem
(Adler & Kwon, 2001; Coleman, 1988). Researchers have also pointed out and found

some support for the notion that the benefits of social capital within a problem solving

group are augmented when different members have access to different types of resources
(Burt, 2000; Reagans & Zuckerman, 2001). Conversely, they argue that social capital
among members of a group who each have access to redundant types of resources will be
less beneficial than social capital within a diverse group where each member brings to the
table resources not available or possessed by the other members. This would suggest
that diversity in the composition of an IA would interact with the amount of IA social
capital such that the positive relationship of social capital to IA effectiveness will be
strengthened in the context of greater diversity. This study investigates this relationship
by examining how diversity in IA composition impacts the relationship between IA social
capital and IA effectiveness.
IA Outcomes for Which Social Capital Matters Most

IAs are designed to be comprehensive strategies capable of improving the
community’s response to a given social issue at multiple levels through various
community change efforts (Wolff, 2001). Examples of such efforts include increasing
coordination among community organizations and agencies, addressing service gaps by
facilitating the development of new or the improvement of existing programs and
services, influencing positive changes in public policies, and promoting prevention
efforts through public awareness and education (Wolff, 2001). In recognition of this,
scholars have theorized that different types of community change efforts may require
different types of capacities (Himmelman, 2001). However, despite this recognition at a
theoretical level, there has been little empirical investigation aimed at building our
understanding of what IA conditions are most important for what types of outcomes.

This study begins to address this gap by exploring whether social capital within the IA is

 

 

 

 

mil?

OUR

uni

of:

the

 

more important for IA effectiveness at accomplishing certain types of community change
outcomes over others.
Summary
In summary, the purpose of this study is to contribute to our current
understanding of IAs by providing one of the first targeted, comprehensive examinations
of the role of social capital within IAs. In order to accomplish this, this study addresses
the following research questions:
A. What is the relationship between social capital and IA effectiveness?
B. How does diversity within the membership affect the role of social capital as it
relates to IA effectiveness?

C. For which outcomes does social capital matter most?

 

 

ODE

\‘io

SCI

COI

 

INTRODUCTION

This study investigates the role of social capital in promoting the effectiveness of
one prominent type of IA formed to improve a community’s response to domestic
violence. With an estimated 4.8 million intimate partner rapes and physical assaults
perpetrated against women every year (Tjaden & Thoennes, 2000), domestic violence is a
serious issue negatively affecting the health and well being of women, families, and
communities throughout the United States. Like most significant social issues,
addressing domestic violence has been recognized as beyond the capacity of any single
agency or institution to address (Clark, Burt, Schulte, & Maguire, 1996; Hart, 1995;
Shepard, Falk, & Elliott, 2002). As a result, the past 20 years has seen an influx of
change efforts aimed at assisting communities to develop and implement a more
coordinated community response to domestic violence (Shepard et al., 2002). In general,
a coordinated community response has been described as a “community wide effort to
bring together all relevant stakeholders to respond to domestic violence comprehensively
and ultimately to end that violence” (p. 270, Sullivan & Allen, 2001).

One increasingly common approach for implementing such a response has been
the development of IAs commonly referred to as domestic violence coordinating councils
(DVCCs; Clark et a1, 1996). DVCCs are collaborative groups made up of
representatives and leaders from an array of criminal justice and human service
agencies/organizations (e.g., police, prosecutors, judges, advocates, health providers,
shelter providers) as well as concerned citizens and community groups invested in
addressing domestic violence within their community (Shepard et al., 2002; Sullivan &

Allen, 2001). Their intended function is to reduce the prevalence of as well as improve

thc'

CO‘

nal

the

An

C0

mt

C0

DI

 

\‘i

d:

Ci.

 

the community’s response to incidences of domestic violence by increasing the level of
coordination and collaboration among community agencies and organizations (Hart,
1995; Sullivan & Allen, 2001). With their popularity fueled by the endorsements of
national organizations such as the National Council of Juvenile and Family Court Judges,
the American Prosecutor’s Research Institute, the American Bar Association, and the
American Medical Association, DVCCs have become a prevalent form of IA found in
communities throughout the United States (Gamache & Asmus, 1996).
DVCC Effectiveness

In order to develop a model for understanding the role of social capital as a
mechanism through which DVCCs positively impact their members and their
community, it is important to first define what DVCC impacts might look like if the
DVCC were to be effective in its efforts. The current literature on evaluating the
effectiveness of IAs has begun to recognize that outcomes of collaborative groups like
DVCCs occur at multiple levels (Provan & Milward, 2001; Stevenson & Mitchell, 2003;
Sullivan & Allen, 2001). Indeed, an area of criticism within existing research on IAs has
been the tendency for researchers to define effectiveness criteria too narrowly, thereby
failing to adequately model the impact of such interventions (Provan & Milward, 2001 ).
In order to overcome this weakness, this study tested a multi—dimensional model of
DVCC effectiveness in contributing to a coordinated community response to domestic
violence. Building upon current literature, I proposed that DVCCs contribute to the
development of a more coordinated community response by both promoting community
change and strengthening the capacity of member organizations. Each of these are

discussed below.

 

 

 

 

Ct

I'll

C0

pt]

8?

Ci}

 

SC

 

Promoting Community Change
Most often the principal reason given for convening a DVCC has been to create

systems changes within a community (Sullivan & Allen, 2001). For this study,
considerations of DVCC effectiveness at promoting community change takes into
account two components. The first is the effectiveness of the DVCC at making
comprehensive change across multiple areas believed to ultimately reduce the prevalence
of domestic violence in the community, increase the safety, health, and well-being of
those impacted by domestic violence, and increase the accountability of those who
perpetrate domestic violence (Hart, 1995; Pence & Shepard, 1999; Sullivan & Allen,
2001). The second component is DVCC effectiveness at accomplishing its goals and
making headway at fixing the most significant problems with the community’s response
to domestic violence. Each of these are discussed in turn.
Comprehensiveness

Based on the premise that improving the community’s response to domestic
violence requires change at multiple levels, evaluations of comprehensiveness are
concerned with whether DVCCs are serving as effective mechanisms for creating
positive changes in the community’s response to domestic violence across several key
areas (Allen, 2005). These areas include (1) increasing the level of coordination among
community agencies and organizations, (2) shifting policies concerning domestic
violence, (3) addressing service gaps by developing new or improving upon existing
services and programs, and (4) increasing public awareness and knowledge of domestic

violence.

Each of these areas constitutes a specific type of community change argued by
scholars as important for addressing domestic violence within a community. For
example, DVCC effectiveness at increasing the level of coordination among community
organizations and agencies refers to whether the DVCC has caused organizations to see
themselves as part of a larger community system and, as such, led to them working
together more effectively and efficiently. Such coordination has been identified by
experts as critical to creating a community system capable of responding to incidences of
domestic violence in a way that is coherent and maximizes the potential for victim safety
and offender accountability (Pence & Shepard, 1999). DVCC effectiveness at shifting
policies refers to whether the DVCC has been successful at making needed changes in the
practices and protocols of community organizations and institutions and/or local
legislation governing the community’s response to the crime of domestic violence. For
example, shifting policies to require pro-arrest or mandatory arrest procedures is often
advocated for as an effective vehicle for increasing batterer accountability (Shepard,
Falk, & Elliott, 2002). DVCC effectiveness at addressing service gaps refers to whether
the DVCC has served to impact the community service system by either improving
existing programs or developing new programs to better meet the needs of the
community. Experts have argued that service systems constantly need to be critically
examining programs and services to ensure that they are accessible to and meeting the
needs of the community they serve (Jones, 2003). Lastly, DVCC effectiveness at
increasing public awareness and knowledge of domestic violence refers to whether the
DVCC has served to educate its community concerning domestic violence. This area of

change is important as increasing public awareness is commonly viewed as a key vehicle

10

for reducing the incidences and severity of domestic violence in the community (Sullivan
& Allen, 2001) .

Thus, while it is recognized that increasing interagency coordination, shifting
policies, developing new or improving upon existing services and programs, and
increasing public awareness are not the ultimate outcomes sought by a DVCC, changes in
these areas are referenced as important intermediate outcomes through which the
DVCC’s goals of violence reduction, victim safety and well-being, and offender
accountability can be achieved (Allen, 2001). Examining DVCC effectiveness at
creating community change in multiple areas is important as DVCCs are explicitly
intended to be comprehensive strategies capable of impacting the community’s response to
domestic violence at multiple levels and through multiple approaches (Sullivan & Allen,
2001). Based on this criterion of comprehensiveness, it can be argued that a DVCC which
has improved service coordination, reduced service gaps, made productive changes to policy,
and educated their community about domestic violence has been more effective at promoting
community change than a DVCC who has made headway on only one of those outcomes.
Addressing Community Gaps

It is essential that conceptions of DVCC effectiveness at promoting community
change also balance evaluations of comprehensiveness with measures that take into account
the extent to which the changes that have been accomplished by the DVCC are leading the
DVCC closer to its goals and addressing the issues of greatest concern. Important questions
in this area include the extent to which the DVCC has made good headway at fixing the
most significant problems with the community’s response to domestic violence and the

extent to which the DVCC has engaged in activities that have consistently moved it

11

cl(

6“

In

ch

St

Cl

 

 

closer to achieving its goals. The necessity of such questions results from the reality that
even though a DVCC may be effective at creating community changes across an array of
areas, it may not necessarily be effective at creating the kinds of changes acknowledged by
its members as being those most needed in the community to accomplish its goals.

As such, this study examined DVCC effectiveness at promoting community
change defined as its effectiveness across the areas of increasing coordination, shifting
policies, developing/improving services and programs, and increasing public awareness.
In addition, the operationalization of DVCC effectiveness at promoting community
change incorporated a measure of the general effectiveness of the DVCC at
accomplishing its goals and making the changes believed by members as most needed in
the community.

Strengthening the Capacity of Member Organizations

A less commonly investigated but nonetheless important impact of a DVCC
concerns the extent to which the DVCC has been effective at strengthening the
organizational capacity of its members. The term capacity as it is used in this study
refers to an attribute of an organization or agency that facilitates the accomplishment of
its mission and/or helps the organization to sustain itself over time (Letts, Ryan, &
Grossman, 1999). At this level, the focus is not on how the existence of the DVCC has
impacted the community as a whole, but rather, on the extent to which the DVCC has
served to benefit the organizations and agencies who participate as members. However,
conceptually, it is important that this not be confused with an organizational member
level outcome. The premise of this study was that, on average, DVCCs differ in their

effectiveness at strengthening the organizational capacity of their members. As such, it is

12

II it I I I
.L

Iht‘
eff
or:
Gr
m:
5U":
fur
.0
in:
is '.

7

in.

Or;

(i-t

(B

SUE

pal

Sui

the variance between DVCCs— not between members- that is of interest in this area of
effectiveness.

Theorists have long argued that IAs can play a critical role in building
organizational capacity and promoting organizational stability (Bailey & Koney, 2000;
Gray, 1989). While the question of how participation in a DVCC impacts the capacity of
members has received scant attention, research on interorganizational relationships has
suggested that involvement in an IA can strengthen organizational capacity in a number
of ways. Figuring prominently among such organizational benefits are (1) increased
access to information that can support organizational learning (Ahuja, 1992; Bailey &
Koney, 2000; Burt, 1992), (2) increased capacity for solving problems that arise (Alter &
Hage, 1993; Ahuja, 1992), (3) increased access to resources (e.g., expertise, supplies,
funding) that can promote the stability and sustainability of the organization (Austin,
2000; F oster-Fishman et a1, 2001; Ireland, Hitt, & Vaidyanath, 2002) and (4) greater
influence and visibility within the community (Provan & Milward, 2001).

Focusing on the effectiveness of DVCCs at strengthening member organizations
is an important addition to considerations of DVCC effectiveness as it recognizes that
involvement in an IA, such as a DVCC, has substantial implications for how an
organization or agency does its work, often requiring a significant diversion of resources
(i.e., time, funding, identity, and organizational priorities) to support the partnership
(Bailey & Koney, 2000). Given these costs, DVCCs must be effective mechanisms for
supporting and building the capacity of their organizational members in order for active
participation to be a productive and viable option over time. Failure to provide

substantive benefits to members may have negative consequences both in decreasing the

13

internal commitment to, and subsequently the efficacy of, the DVCC (Butterfoss,
Goodman, & Wandersman, 1993), as well as potentially weakening member institutions
by diverting scarce resources away from an organization’s primary programming (Bailey
& Koney, 2000). As such, this study examined DVCC effectiveness at strengthening the
organizational capacity of members in the areas of information attainment, problem
solving knowledge and abilities, sustainability (resources attainment), and community
influence and visibility.
The Theoretical Relevance of Social Capital to DVCC Effectiveness

A second important task in developing a model for understanding the role of
social capital in promoting the effectiveness of DVCCs is to operationalize social capital
and more fully articulate its theoretical relevance to IAs. IAs are designed to provide the
opportunity and structure for an array of prevention and response organizations and
agencies to regularly come together to increase the level of coordination and
collaboration within their community (Bryson & Crosby, 1992; Foster Fishman et al.,
2001; Gray, 1989; Sullivan & Allen, 2001). One of the key characteristics that
distinguish collaborative or coordinated strategies from other approaches to community
change is their reliance on cooperative relationships (Foster-Fishman et al, 2001; Alter &
Hage, 1993). Cooperation as it is used in the present study refers to actions that involve
actors working together toward a common goal or in one another’s interests (Alter &
Hage, 1993). This would include a range of interorganizational activities common to IAs
such as information sharing, coordinating services, and collaborating on joint projects
(Himmelman, 1996). Experts argue that strategies involving coordination and

collaboration rely upon cooperative relationships characterized by qualities such as a high

14

 

 

 

 

degre
anofl‘

(Era)

coagi
prod!
nece_

on 11‘.

Open

(llafl

 

effor‘
SUCCI
othe
\Nho

”lak

degree of trust, legitimacy, frequent communication, and willingness to respond to
another’s concerns for their success (Alter & Hage, 1993; F oster-Fishman et al, 2001;
Gray, 1989; Mattessich, Murray-Close, & Monsey, 2001).

According to the literature, such relationships signify social capital for
cooperative action as they represent a resource that can be drawn upon to produce a
productive outcome (Gabbay & Leenders, 2001; Coleman, 1988). The theoretical
necessity of relationships high in social capital for supporting cooperative action is based
on the premise that cooperation with others has costs, and therefore risk, associated with
it (Alter & Hage, 1999; Bailey & Koney, 2001; Coleman, 1988; Gray, 1989). These
costs, such as loss of autonomy and commitment of time and resources, are a key
consideration for stakeholders in deciding whether to cooperate with others (Butterfoss et
al., 1993; Huxham, 1996). This is particularly true for organizations and agencies
operating in the resource scarce environment of the health and human services sector
(Bailey & Koney, 2001). The premise is that if each stakeholder within a cooperative
effort effectively does his or her part, then their collective effort is more likely to be
successful and the resource investment worthwhile. However, if some follow through but
others fail or, worse yet, take advantage of the situation for their own gain, then those
who committed their resources stand to loose. Therefore, cooperation entails that parties
make themselves vulnerable to a certain level of risk stemming from uncertainty
regarding whether the other parties are both capable and willing to follow through on
what they promise. Even if the cost is only the possibility of wasting time and energy on
an unproductive effort (Alter & Hage, 1994), it is this very vulnerability that can make

collaboration and coordination so tenuous. In order to minimize the risk to themselves,

15

 

 

 

 

sake
(Cka}

cannr

firrc
orgr
FOSlt
(I10
SUQS
conu
Art)

denr

 

 

nay
so.i
furtl
und
bet
res
\rh.
rek

Iha

fin

stakeholders may hold back on what they are willing to commit to a collective effort
(Gray, 1989). However, unfortunately, outcomes requiring collective action frequently
cannot be achieved in a context of such restraint (Mizrahi & Rosenthal, 1992).

Given this bleak prescription - how, then, does collaboration occur? For, indeed,
there is an abundance of examples that give testimony to the. fact that people,
organizations, and even communities work together cooperatively all the time (Bartunek,
Foster F ishman, & Keys, 1996; Coe, 1988; Hardy & Phillips, 1998; Mulroy, 1997;
O'Looney, 1994; Reilly, 2001). In answer to this question, social capital literature
suggests that over time and through interaction, actors (e.g., people, organizations,
communities) can develop relationships with one another that have value (Lin, 2001).
Actors learn what they can expect from one another and begin to trust in one another’s
demonstrated capabilities. Each party to the relationship can be counted upon to act in a
way that takes into consideration the best interests of the other party(s) because, by doing
so, it strengthens the relationship. Among multiple actors, norms begin to develop that
further reinforce the value of the relationships and sanction behaviors that threaten to
undermine them. Thus, the relationships themselves become a commodity of value
because the stronger the relationships, the greater chance that parties will have access to
resources made available through those relationships (Coleman, 1988). This value is
what is referred to as social capital - the value which exists within social-structural
relationships that an actor can mobilize to make possible the achievement of certain ends
that would not be attainable in its absence (Coleman, 1990).

Theorists argue that social capital is a highly facilitative if not necessary condition

for networks of individuals or organizations to collaboratively work together due to its

16

 

 

 

abilit
mem
SUPP‘
have
Lcen
D\'C
“1th
orga
(Har

com

impr-

will

arm ,1

 

 

ability to reduce the risk associated with cooperation and increase the accessibility of
member resources (Coleman, 1988). The organizational science literature has provided
support for this claim, finding teams high in social capital to be more innovative and to
have greater capacity for achieving common goals (Baker, 2000; Cohen & Prusak, 2001;
Leenders & Gabbay, 1999). Herein lies the relevance of social capital to DVCCs.
DVCCs are formed based on the explicit recognition that addressing domestic violence
within a community is beyond the scope, resources, and expertise of any one agency or
organization, and therefore will require a coordinated effort from multiple organizations
(Hart, 1995). However, in order for a coordinated community response to occur,
community organizations and agencies must have the relationships necessary to work
together cooperatively. As such, one can argue that while the objective of a DVCC is to
improve the community’s response to domestic violence, an implicit premise is that it
will achieve this, at least in part, through fostering the development of social capital
among community organizations and agencies.
Conceptualizing Social Capital

As discussed thus far, social capital is embedded within the characteristics of social
relations such as the degree to which two actors trust, communicate, and are responsive to
one another’s concerns (Knoke, 1999). Within the social capital literature, each of these
aspects (e.g., the extent to which one actor trusts another actor) is referred to as a tie
characteristic. A tie is defined as a connection or relationship between two actors
(Borgatti & Foster, 2003; Wasserrnan & Faust, 1994). A tie characteristic refers to a

quality of that relationship.

17

 

 

 

com,
to re

tie c

EXp:
era;

and

50C r

“€12

 

A tie characteristic becomes representative of social capital within the relationship
between two actors when there is reason to believe that it could facilitate the achievement
of certain ends that would not be attainable in its absence (Coleman, 1988). For example,
it has been demonstrated that two actors who trust one another to follow through on what
they say they will do have greater potential for engaging in cooperative action than those
who have a relationship characterized by mistrust (Tsai & Ghoshal, 1998). It is
important to note, however, that social capital is not a unidimensional construct — it is
comprised of multiple tie characteristics (e. g., trust, level of communication, willingness
to respond to concerns) that are both instrumental and expressive in nature. Instrumental
tie characteristics are qualities of a relationship that relate to the types of resources that
can be obtained via that relationship (Ibarra, 1993). For example, within
interorganizational research, instrumental ties examined have included advice giving,
information sharing, and assistance (Knoke, 1999). Expressive tie characteristics tend to
reflect the more social-emotional qualities of relationships among actors within the
network that may be facilitative of cooperative action (Ibarra, 1993; Lincoln & Miller,
1979). For example, researchers have sought to understand network functioning based on
expressive ties such as fiiendship, trust, and respect. The beliéf is that how one actor
evaluates another actor on such characteristics will have bearing on the degree to which
and manner in which they will interact with one another.

Because the importance of different types of tie characteristics varies by context
(Adler & Kwon, 2002), it is important to articulate what types of ties are relevant to the
social network under investigation. For this study, the context of interest was the

network of ties among DVCC members. Within this network, five types of tie

l8

 

 

that

inflrf

infr

law t

 

bane

Willi

indk

 

abou

not

3110‘

the

C01

\V

CC

ht

.‘Sa

characteristics were examined: communication frequency, responsiveness to concerns,
trust, recognized expertise, and shared philosophy. The identification of these ties was
informed by both current literature as well as qualitative interviews with DVCC key
informants representing each of the major stakeholder groups participating DVCCs (e. g.,

law enforcement, prosecuting attorneys, shelters/victim service providers, courts, and

batterers intervention)’. Each tie characteristics is described below.

Qualitative interviews revealed that instrumental ties around fiequent interaction and
willingness to respond to concerns were particularly valued social relations and therefore
indicative of social capital within a DVCC context. For example, key informants talked
about DVCCs being most effective when members had developed relationships with one
another in which members felt confident that if they brought up a concern or issue to
another member, that member would be responsive in taking the concern seriously and do
their best to address the issue. The belief that another actor would be responsive to
concerns can be thought of as an instrumental tie as it directly relates to a resource
believed to be available to an actor through a relationship (Ibarra, 1993). Furthermore,
key informants reflected on the importance of communication and frequent interaction
among members. Specifically, they suggested that DVCCs worked well when members
went out of their way to share relevant information with one another and felt comfortable
contacting another member to ask a question or request information. Frequent
communication and a willingness to respond to one another’s concerns represent both
how accessible members are to one another as well as how likely they are to assist one

another. Other researchers have similarly identified these as valuable resources for

 

' Sampling and methods for the preliminary qualitative interviews are described in detail on page 54

19

 

 

 

 

 

' —

DIOF
allifl
C00}
in ll
CXCII

Fen

inll

 

am}

c0“

 

cooperative action. For example, Rivard & Morrissey (2003) found that greater service
coordination was more likely to occur in dyads of service providers who also helped each
other attain individual agency goals. Reagans and Zuckerman (2001) looked at
communication frequency among dyads within organizational teams and found that teams
whose members communicated with one another more frequently outperformed teams
characterized by less frequent communication.

One of the most commonly investigated expressive ties within the literature on
social capital is trust. Consistent with research in social psychology and group dynamics
(e. g., Parks & Hulbert, 1995), there appears to be a general consensus among social
capital scholars that trust is a highly facilitative - if not necessary - condition for actors
within a network to work cooperatively together (Burt, 1992; Nahapiet & Ghoshal, 1998;
Putnam, 2000; Schuller et al., 2001; Tsai & Ghoshal, 1998; Tsui, Egan, & O'Reilly,
1992). As commented by Knoke (1999), as an actor builds a reputation for keeping
promises, ‘that reputation itself becomes a prized asset useful for sustaining its current
alliances and forming future ones” (p. 33). The importance of trust for facilitating
cooperative problem solving and information sharing has also been strongly recognized
in the organizational literature. For example, based on a review of the literature
examining trust in organizational settings conducted over the past 40 years, Dirks and
Ferrin (2001) found evidence that trust can impact the likelihood of cooperative action
(i.e., cooperative problem solving, information sharing) in two ways. First, trust
influences how an individual assesses the future behavior of another. For example, the
authors suggest that under conditions of high trust, individuals are more likely to

cooperate with others because they judge others as being willing to cooperate with them;

20

thereby reducing the degree of risk involved. Second, they suggest that trust influences
cooperation by affecting how a person interprets past or present actions of another actor.
This proposition is based on the idea that individuals assess risk in a cooperative
undertaking based on what they think they can expect from the other party(s).
Knowledge of a person’s past or present actions typically form the basis for these
expectations. However, any one action can be interpreted a number of ways. The
authors argue that under conditions of high trust, individuals are more likely to believe
another’s explanation for their actions as well as be more likely to evaluate actions as
benevolent in their intention.

Consistent with this, previously mentioned qualitative interviews with DVCC key
informants similarly identified trust as an important characteristic of DVCC member
relationships. Specifically, informants indicated that trust in another member’s
willingness to follow through on what they say they will do is a valued resource within
the DVCC. Consistent with Dirk & Ferrin’s (2001) propositions, key informants talked
about the value of such trust for reducing their sense of risk when cooperating with
another whom they assessed as being trustworthy.

Two other types of expressive ties that have not yet received a great deal of
consideration within the social capital literature but may be highly relevant within IA
networks like DVCCs are ties of recognized expertise and shared philosophy. Building
off of the work of Gray (1989; 1991), for the purpose of this study, recognized expertise
is defined as the belief that another actor offers a unique perspective or form of expertise
that is valuable to understanding and/or addressing a shared problem or concern.

Particularly when collective problem solving efforts require the involvement and

21

 

 

inve.

 

stak

 

impl
the c

200s.

grou
gror.~
gm
info

Pail

 

 

info
(\Vi
“it
is b.
198
11116
pro
of l
(Br

19

tel

investment of a diverse range of stakeholders (Gray, 1989), the legitimacy of each
stakeholders’ expertise in the eyes of the others has been found to have important
implications for the ability and willingness of a collaborative group to make full use of
the diverse perspectives and forms of expertise represented at the table (Bond & Keys,
2000; Clark, Baker, Chawla, & Maru, 1993).

One explanation for the importance of recognized expertise within collaborative
groups comes from related research on factors that facilitate information sharing in
groups. Laboratory research on decision-making groups has found that even diverse
groups prefer to discuss shared information that all members know instead of unshared
information known only to one or a few individuals within the group (Wittenbaum &
Park, 2001). Through a process of mutual enhancement, discussing commonly shared
information is experienced as validating to both the discussant and the group
(Wittenbaum, Hubbell, & Zuckerman, 1999). Further, cognitive heuristics associated
with consensus on shared knowledge leads group members to infer that such knowledge
is both accurate and important to the decision making task at hand (Chaiken & Stangor,
1987; Karneda, Ohtsubo, & Takezawa, 1997; Wittenbaum et al., 1999). In the context of
inter-organizational problem solving groups such as DVCCs, this effect may be highly
problematic in that, frequently, stakeholders are characterized as holding different types
of expertise and experiential knowledge relevant to addressing domestic violence
(Butterfoss et al., 1993; Coe, 1988; de Jong, 1996; Kegler, Steckler, Malek, & McLeroy,
1998). Information sharing theory suggests that groups like DVCCs may struggle to
benefit from the diverse range of perspectives represented at the table because of the

tendency of members to preference discussions concerning information and ideas shared

22

 

 

 

 

 

offe

notl
kno

3V3

Col

Cap

by all. However, research on information sharing has also found that the bias toward
discussing shared information is mitigated when members are positively evaluated by
other group members as being capable and having expertise into the task at hand
(Wittenbaum, 2000; Wittenbaum & Park, 2001). Such evaluations are congruent with
Gray’s (1985) conceptualization of legitimacy suggesting that DVCCs characterized by a
higher degree of recognized expertise among members may be more likely to have access
to perspectives, experiences and insights unique to a given member or subset of
members.

Consistent with Gray’s and Wittembaum’s propositions, key informant interviews
revealed that the legitimacy of a given DVCC member’s expertise in the eyes of their co-
members was viewed as an important resource to that member. Specifically, informants
commented that effective DVCCs were characterized by a membership that
acknowledged and valued the expertise that each member brought to the table. Such
acknowledgements were important as they were believed to increase member
commitment to the group as well as the group’s utilization of what each member had to
offer. Scholars of interorganizational collaboration have made similar observations,
noting that legitimacy of a member in the eyes of group makes it more likely that the
knowledge, perspectives, and resources available to that member will become a resource
available to the group (Alter & Hage, 1993). As such, member relationships
characterized by perceptions of a high degree of recognized expertise can become a
collective resource to the DVCC and therefore represent an important form of social

capital.

23

 

 

IllSl\
by .l

dom

 

addr
agai
and
Sire
ml
per

lnlr

int
fra
\\‘(

C0

A shared philosophy as it is used in this study concerns the extent to which members
of a DVCC perceive one another to share a similar philosophy regarding what domestic
violence is and how it should be addressed within the community. While not commonly
present in operationalizations of social capital, relationships characterized by at least
some degree of “sharedness” in beliefs, vision, and/or frameworks of understanding has
been well-recognized scholars and practitioners as an important factor in promoting
collaboration (e.g., Gray, 2005). This form of social capital is particularly germane to the
history and context of the Domestic Violence Movement in the United States. As argued
by Jenkins and Davidson (1999), assumptions and philosophies about the nature of
domestic violence have served to define what practices, policies, and laws are enacted to
address it. Traditionally in the United States, intimate partner violence perpetrated
against women was viewed as originating from person level circumstances, inadequacies
and/or pathologies such as abuse of alcohol, poor relationship skills, or inability to handle
stress, anger, or frustration (Pence, 1999). These frameworks led practitioners and policy
makers to focus largely on addressing domestic violence through encouraging
perpetrators, and often their partners, to seek help in dealing with their intra and
interpersonal deficiencies (Jenkins & Davidson, 1999).

However, over the course of the past 30 years, women’s advocates have fought to
introduce new frameworks for understanding intimate partner violence against women —
frameworks that give audience to the social and cultural facilitators of violence against
women such as assumptions of the legitimacy of male authority and a national history of
condoning the use of violence as a means for husbands to control their wives (Pence,

1999; Pence & Gallaway, 1985). Out of this movement have emerged counter

24

 

 

 

 

that
to ei

h0\\

 

prac
offir
the
rest
lacl
org
age
Cor
(B;

1h:

philosophies which shift the diagnosis of domestic violence from a pathology or
individual deficiency to one of coercion in which men’s use of violence against women is
understood as a means for obtaining and maintaining power and control in the
relationship (Pence, 1999). In conjunction with this philosophy has been the strong
avocation that domestic violence be viewed as a crime of no lesser significance than any
other form of physical assault and that community efiorts to address domestic violence
focus their attention on domestic violence as a crime rather than a family issue (Hart,
1995)

As such, the history of the field of domestic violence over the past 30 years has been
marked by heated struggles as policy makers, practitioners, and advocates have all sought
to either shift or maintain the dominant philosophy for what domestic violence is and
how it should be addressed so as to define the basis for influencing the policies and
practices of community agencies such as the police, courts, prosecutors, and probation
offices. These struggles exist to this day and have been blamed as a key contributor to
the very lack of collaboration and coordination among domestic violence prevention and
response agencies that DVCCs are convened to address. In light of this diagnosis that a
lack of shared philosophy constitutes a major barrier to cooperative action among
organizations or agencies, it is unsurprising that a shared philosophy across community
agencies and organizations has been described as facilitative if not necessary to a
community’s ability to create a comprehensive and unified response to domestic violence
(Building a coordinated community response seminar, 2000; Hart, 1995). Thus, because
the extent to which member relations within a DVCC are characterized by a common

philosophy concerning the nature and appropriate response to domestic violence is

25

 

 

Sr

ft’.‘

(20‘

lhr

 

 

premised to have such significant implications for the ability of DVCC members to work
together cooperatively, it is presented in this study as an additional form of social capital.

Up to this point, social capital has been described in terms of a collection of various
tie characteristics (i.e., willingness to respond to concerns, communication frequency,
trust, recognized expertise, and shared philosophy) that characterize a relationship at the
dyadic level. However, while social capital is embedded within dyadic ties, the level of
interest for social capital in the proposed study does not lie at the tie level but rather at the
level of the DVCC. In order to model social capital as a characteristic of an
interorganizational network such as a DVCC, a social network approach was utilized. A
social network approach to conceptualizing social capital is concerned with the patterns
and configurations of ties among the actors in the network. This is referred to as network
structure (Wasserman & Faust, 1994). Network structure has been argued by scholars to
be important in examining social capital because it determines how and to what extent
resources move through a social network such as an IA (Burt, 2000). Specifically, in
considering social capital as a network-level property as is the case in this study, one of
the most commonly investigated attributes of network structure is the interconnectedness
of the network or the network density (Balkundi & Harrison, 2004).

Greater density within a network such as a DVCC means that there are more and
stronger ties among actors of the network. For example, in the context of the present
study, it would indicate that more members are more trusting of one another,
communicate more frequently, perceive one another to be legitimate contributors to the
group, and are more willing to go out of their way to help each other out. Because a

dense network of instrumental ties indicates that actors are actively engaged with one

26

another and dense networks of expressive ties indicate a high level of emotional
closeness between actors (Reagans & Zuckerman, 2001), theorists argue that groups with
high network density have greater access to the full resources represented within the
network, and therefore have an increased capacity to leverage these resources toward
achieving their collective goals (Coleman, 1988). As such, network density across
stakeholder dyads is suggested to be indicative of social capital due to its potential for
facilitating the achievement of common goals within groups like DVCCs (Aldrich &
Zimmer, 1986; Coleman, 1988; Portes, 1998).

In sum, in an attempt to capture both the structure as well as the multi-dimensional
nature of social capital within an interorganizational network, this study modeled social
capital for a given tie as comprised of (1) communication frequency, (2) willingness to
respond to concerns, (3) trust, (4) recognized expertise, and (5) shared philosophy. The
stronger the relational tie, the more likely a member was to respond to the other’s
concerns, communicate regularly, trust the other member, consider them to have
legitimate expertise, and consider them to share a similar philosophy as them self. Social
capital of the DVCC was modeled as the density of those ties.

The Role of Social Capital in Promoting DVCC Effectiveness

Social capital literature would suggest that DVCCs with higher levels of social capital
have greater advantage and therefore will be more effective at positively impacting their
members and their community. While this relationship has received very little attention
within the context of DVCCs, evidence suggestive of its existence can be found in

research from the broader field of IAs and interorganizational collaboration.

27

 

 

 

 

com

fort

War

 

The Role of Social Capital in Promoting Community Change

Given that DVCCs are principally formed for the purpose of improving a
community’s response to domestic violence (Hart, 1995; Post, Klevens, Maxwell,
Shelley, & Ingram, 2004; Sullivan & Allen, 2001), the link between social capital and
DVCC effectiveness at promoting systems change is of key interest to the proposed
study. This proposed link is based on the premise that community change efforts are
facilitated by the development of social capital among those stakeholders invested in the
change issue (Coleman, 1988). Qualitative case studies of community-based
interorganizational collaborations lend support to this premise through their identification
of the different indicators of social capital as key factors in the success of collaborative
efforts (Bond & Keys, 2000; Campbell, Dienemann, Kub, Wurmser, & Loy, 1999; Gray,
1996; Mulroy, 1997; Tapper, Klienman, & Nakashian, 1997). For example, a case study
by Campbell and colleagues (Campbell et al, 1999) described the development of a
collaborative partnership between a battered women’s shelter and a university school of
nursing. The authors noted that the development of trust between partners was a key
factor in the success of the partnership. Specifically, this relationship made possible the
development of new programs that expanded youth services and increased the
accessibility of health care to women and children living in emergency housing. Another
case study by Tapper and colleagues (Tapper et al., 1997), which reports on a multi-
agency collaboration among a school, several social service agencies, and a local police
force made a similar observation. The authors note that mutual trust among stakeholders

was a critical factor in their ability to coordinate service efforts which in turn led to the

28

 

 

impt

inner

plOt‘.
[lite

and

netti

 

 

 

 

improvement of programs aimed at preventing drug use and delinquency among youth in
inner city neighborhoods (Tapper et al., 1997).

Additional support for the link between social capital and DVCC effectiveness at
promoting community change is found in the writings of several prominent social capital
theorists who have argued for the value of dense networks in a group’s ability to pursue
and achieve common goals. As mentioned, these theorists argue that groups with high
network density have greater access to the full resources represented within the network,
and therefore have an increased capacity to leverage these resources toward achieving
their collective goals (Aldrich & Zimmer, 1986; Coleman, 1988; Portes, 1998). Despite
the fact that this premise has not been systematically studied within the context of IAs,
this argument is consistent with mounting evidence within the, organizational literature
which identifies social capital as a key factor in strengthening the performance and
effectiveness of teams or work groups who have interdependent goals. For example, in a
meta analysis of 31 studies examining the impact of social networks on team
performance, dense network structures of both instrumental and expressive ties among
members were found to be positively associated with team task performance (Balkundi &
Harrison; 2004). Using the culmination of these findings as a foundation, this study
explicitly examined whether higher levels of social capital within a DVCC are positively
related to DVCC effectiveness at promoting community change.

H1: HIGHER LEVELS OF SOCIAL CAPITAL AMONG DVCC MEMBERS

WILL BE POSITIVELY RELATED TO THE EFFECTIVENESS OF THE
DVCC IN F OSTERING COMMUNITY CHANGE

29

 

Wilt
11107

pre

intc
ben
al.

int

wit

0f '\

the:

Pit;

 

Shi

bet

 

The Role of Social Capital in Building Organizational Capacity

Of additional interest to the current study was whether organizations and agencies
who are members of DVCCs characterized by higher levels of social capital benefited
more from their participation in the DVCC. This speaks to the role of social capital in
predicting the overall effectiveness of the DVCC at building the organizational capacity
of its members.

Many scholars have documented, often through case studies, that involvement in
interorganizational collaborative partnerships, such as DVCCs, can provide substantial
benefits to the participating agencies and organizations (Alter & Hage, 1993; Campbell et
al., 1999; Gray, 1996; O'Looney, 1994). As mentioned, commonly these benefits come
in the form of increased access to information that can be used to problem solve issues
and improve/expand services (Alter & Hage, 1993; Austin, 2000), and increased access
to physical and financial resources (Kohm, 1998; Neilsen, 1988; Tsai & Ghoshal, 1998).
However, despite this widespread recognition that such benefits can emerge from
involvement in interorganizational alliances and partnerships, there is far less attention
within the field of IAs directed at systematically investigating the more practical concern
of when such benefits are likely to result.

Social capital theorists posit that such benefits occur when organizations, through
their representatives, foster active cooperative relationships with one another (Cohen &
Prusak, 2001). Within the context of these relationships, organizations are motivated to
share information and resources that they feel might be valuable to the other partners
because in doing so they strengthen the relationship and increase the chances that the

other partners will act the same on their behalf (Coleman, 1988). Further, to the extent

3O

 

 

that
join
that

ben

ei’fe
coll
sup

C88;

rela
the

hen

IEpl
uni'

IECT

No

Cap

 

Olg;
evi-g

alt?

 

that an organization develops active cooperative partnerships, it may be able to pursue
joint efforts that make available additional resources that could not have been attained by
that organization working independently (Huxham, 1996). In short, these organizational
benefits are attained through the development of social capital among members.

While there is no known research on the relationship of social capital to the
effectiveness of [As at increasing member organizational capacity, research on
collaborative partnerships within the field of domestic violence and elsewhere provides
support for the existence of this relationship. For example, in the previously described
case study of a partnership that developed between a battered women’s shelter and a
university nursing school, Campbell and colleagues (1999) noted that the trusting
relationships that developed between the university and shelter staff facilitated not only
the accomplishment of their collective goals, but also provided individual organizational
benefits as well. Specifically, the partnership led to the mutual sharing of information on
best practices which helped to improve shelter services. In addition, Campbell (1999)
reported that these cooperative relationships led to increased access to resources for the
university as the shelter has now become a willing setting for gaining access to and
recruiting domestic violence survivors as participants in faculty research.

While the focus of Campbell’s (1999) study was on a dyadic partnership between
two organizations, social network researchers have examined the relationship of social
capital to increased organizational capacity within more complex networks of multiple
organizations or units. The dominant theory is that settings high in social capital as
evidenced by densely embedded networks, with many connections linking organizations,

are advantageous to organizations due to their ability to facilitate resources sharing and

31

 

 

Cha;

 

 

information exchange which in turn serves to benefit organizations (Ahuja, 2000;
Coleman, 1988; Shan, Walker, & Kogut, 1994). Indeed, linkages to other organizations
through collaborative networks have been found to be the key vehicles through which
firms access external knowledge and resources (Powell, Koput, & Smith-Door, 1996).
Specifically, researchers have found social capital to be a powerful factor in explaining
an organization’s success in areas such the extent to which their staff have opportunities
to develop additional knowledge, skills, and abilities relevant to their work (Hargadon &
Sutton, 1997; Nahapiet & Ghoshal, 1998), the extent to which they have access to
resources (Gabbay & Zuckerman, 1998; Tsai & Ghoshal, 1998); and the extent to which
they have opportunities for interfirm learning (Kraatz, 1998). Consistent with this, this
study proposed that DVCCs characterized by greater social capital would be more
effective at positively impacting their members.

H2: HIGHER LEVELS OF SOCIAL CAPITAL AMONG DVCC

MEMBERS WILL BE POSITIVELY RELATED TO DVCC

EFFECTIVENESS AT BUILDING THE ORGANIZATIONAL

CAPACITY OF MEMBERS

In sum, this study posited that DVCCs characterized by greater social capital
among members would be more effective as defined by their ability to foster community
change and strengthen the capacity of their members (see Figure a).
’ ’ DVCC EFFECTIVENESS? . ‘

Community Change

 

 

 

DVCC Social
Capital

 

 

 

 

Strengthened Member
Capacity

 

 

 

 

 

 

 

32

 

 

 

Composition Effects On The Role Of Social Capital

Theory and research on social capital strongly advocate that considerations of
social capital in its relationship to the effectiveness of a defined network of actors such as
a DVCC must also attend to the diversity of resources represented within the network.
The reason for this is that while social capital speaks to the likelihood that members in a
network will work together cooperatively and be willing to share the resources they have
available to them such as their knowledge, skills, influence, and assets (Coleman, 1988),
social capital scholars have argued that the advantages of such relationships may be
constrained if all members have access to similar types of resources (Burt, 2000; Reagans
& Zuckerman, 2001).

Scholars interested in interorganizational collaboration have made similar
arguments for the value of multi-stakeholder collaboration based on its promise for
uniting the diverse resources available within a community so as to better address its
most intractable issues. As stated by Lasker, Wiess, and Miller (2001), “there is great
potential in partnerships that enable different people and organizations to support each
other by leveraging, combining, and capitalizing on their complementary strengths and
capabilities” (p. 180). This assertion envelops a twofold proposal that partnerships like
DVCCs are most valuable when (1) they are comprised of members willing to work
cooperatively together, and (2) they are comprised of members with different strengths
and capabilities. In other words, this would suggests that social capital and diversity
interact with one another in their relationship to DVCC effectiveness as diversity speaks

to what kinds of information and resources can be gained from an DVCC and social

capital speaks to the likelihood that such resources will be shared.

33

div]
diw
that
can
& (

TESL

 

b€ll

O'R

 

ran.

 

' DEC

sch.

coll

COI‘.

 

and
SOC
liter
56F.

Car

 

Existing research provides a level of support both for the value of composition
diversity for group effectiveness as well as the interaction between social capital and
diversity. With respect to the former, research has provided some evidence indicating
that heterogeneity in group composition (differences in tenure, organizational roles, etc)
can be positively related to work group innovation and creative problem solving (Ancona
& Caldwell, 1992; Bantel & Jackson, 1989; Jackson, 1992). However, reviews of
research in this area has tended to find this relationship to be somewhat inconsistent,
being present in some studies while other studies found either no relationship or even a
negative relationship between work group diversity and group performance (Williams &
O'Reilly, 1998).

One prominent explanation for the inconsistency in these findings has been to
point out that while diversity creates greater potential for effectiveness by expanding the
range of resources available to the group, diverse groups often lack the social capital
necessary to realize that potential. When groups have a high level of social capital,
scholars argue they are more likely to have access to and subsequently benefit from the
collective resources of the group (Coleman, 1988; 1990). When that same group is also
comprised of a diverse membership having access to a range of nonredundant strengths
and capabilities, they have a wider array of resources from which members can benefit
and which can be applied to any given undertaking (Burt, 2000). However, without
social capital, the resources of individual members are less likely to be shared and
therefore have lesser consequence for the group. In this way, social capital is believed to
serve as a mechanism through which the diverse resources held by individual members

can become the collective resources held by the group.

34

 

 

 

COI’r

pfii

neg

ind:

ratlf

cap

 

err

for
Strc
a di
lint
COII
SUS'
Slit

quit

 

 

Empirically, recent research using social network analysis has begun to provide
support for this premise. Specifically, Reagans and Zuckerman (2001) examined the
network structure of 224 corporate research and development teams with regard to their
communication ties. They found that greater communication frequency among dyads of
differing organizational tenure within a team was significantly related to team
performance but that overall tenure diversity within the team was neither positively nor
negatively related to performance. Put simply, it was not merely the fact that there were
individuals of differing tenure status on the same team that predicted performance but
rather it was when younger and older tenured employees engaged with one another that
groups performed better. This indicates that it is the interaction of diversity with social
capital — operationalized in this study as higher levels of communication frequency - that
explains the role of diversity within teams — not the overall level of diversity.

This would suggest that diversity in DVCC composition may have implications
for the extent to which the DVCC is effective at promoting community change and
strengthening the organizational capacity of members. Certainly, it stands to reason that
a diverse composition of members represents a correspondingly diverse array of
knowledge, skills, expertise, and resources that can be brought to bear upon a given
community change effort (Reagans & Zuckerman, 2001). Further, there is reason to
suspect this theory will also have implications for the effectiveness of the DVCC at
strengthening the organizational capacity of members based on the same premise. It is
quite plausible that development of trusting, cooperative relationships with organizations

and agencies that have access to information and resources otherwise unavailable would

35

 

 

yiel

0W;

di'

C3

 

yield greater benefits to an organization than fostering such relationships with
organizations who can offer only redundant knowledge and resources.

However, in order to examine this premise, it is important to articulate the type of
diversity that is of interest. The literature on diversity within organizational settings often
categorizes the many forms of diversity into two primary types: task related and social
(Webber & Donahue, 2001). Task related forms of diversity encompass differences in
knowledge, experience, skills, and abilities that typically correspond to a person’s
professional position in the workgroup. For example, forms of task related diversity
commonly examined by researchers include organizational tenure, role within an
organization, and professional background or level of education (Jackson & Joshi, 2004;
Webber & Donahue, 2001). Social forms of diversity describe those forms of diversity
that represent salient social categories within our society and are ofien readily detectable
based on physical characteristics (Weber & Donohue, 2001). Commonly investigated
forms of social diversity include race, age, and gender (Pelled et al, 1999). Research on
both forms of diversity have found similarly inconsistent results in relation to work group
effectiveness (Weber & Donohue, 2001). In light of this, this study defined diversity as
both task related and social diversity. Task related forms of diversity examined included
DVCC membership tenure, sector (e. g., law enforcement, victim services), and
professional role (i.e., executive, middle manager, direct line staff). Social forms of
diversity examined included diversity in the racial as well as gender composition of the
DVCC. Using these, this study investigated the role of diversity in DVCC composition
as a moderator of the relationship between DVCC social capital and DVCC effectiveness

with the following hypothesis (figure b):

36

H3: HIGHER LEVELS OF SOCIAL CAPITAL AMONG MEMBERS WILL BE
MORE STRONGLY RELATED TO DVCC EFFECTIVENESS WHEN
THERE IS ALSO GREATER DIVERSITY IN DVCC COMPOSITION.

 

 

 

 

 

 

 

 

 

 

 

 

" ’DVC‘C‘EF’FECTIVENESS
DVCC Social ’ ‘ [ Community Change
Capital . '
Strengthened Member
Capacity

 

 

 

 

 

Composition Diversity

 

 

 

Relative Importance of Social Capital to Different Types of
Community Change Outcomes

Another research question of interest in describing the role of social capital in
DVCC effectiveness is to better understand when social capital is most important, or
more specifically, if social capital holds greater importance for certain DVCC outcomes
over others. The context for this question comes from understanding that IAs like
DVCCs often engage in a range of activities and may vary greatly in the proportion of
time dedicated to any one type of activity (Himmelman, 2001). For example, several
collaboration scholars have developed typologies of activities or forms of interaction
engaged in by collaborative groups (e.g., Bryson & Crosby, 1992; Chrislip & Larson,
1994; Himmelman, 2001; Hogue, 1993; Gajda, 2003; Reilly, 2003). Commonly
identified interaction types have incorporated networking or information sharing,
engaging in formal or informal forms of coordination, and/or developing collaborative
planning or programming efforts (Gajda, 2003; Himmelman, 1996; Hogue). Given this, a

pertinent question becomes: Is social capital more important for the accomplishment of

37

lit

rel

out

 

 

 

 

 

chat

SOC

del

 

this

CCF

 

 

certain types of outcomes than others? The utility of this question lies in its contribution
to our ability to define what conditions are most important for what types of outcomes in
order to be more strategic in our implementation and support of DVCCs. Theoretical
support for this question is found in the writing of scholars who have argued that
outcomes vary in the conditions necessary to achieve them — particularly with regards to
the qualities of relationships. For example, Hogue (1993) and Himmelman (2001) both
argue that outcomes that require greater sharing of resources, shared decision making,
and altering of activities for a common purpose require greater trust among members of
an IA. However, despite considerable recognition that different outcomes commonly
sought by IAs may vary in the conditions necessary to achieve them, there has been very
little research conducted to systematically examine this issue. .

The current study conducted an exploratory examination of this question as it
pertains to the condition of social capital by investigating whether social capital is more
important to certain types of outcomes than it is to others. This was done primarily in
two ways. The first approach was to examine whether social capital had a stronger
relationship to the effectiveness of a DVCC at achieving certain community change
outcomes compared to others. Specifically, in accordance with the four community
change outcomes chosen for this study, this study examined whether the strength of
social capital’s influence varied across service coordination, program
development/enhancement, policy change, and public education outcomes. The basis for
this approach was to suggest that if social capital is a more necessary condition for

certain types of outcomes, the strength of the relationship between social capital and

38

 

 

 

 

 

effectiveness should vary significantly depending on what type of community change
outcome is considered.

The second approach was to examine whether DVCCs with higher levels of social
capital tend to focus their time and energies on qualitatively different types of
activities/outcomes than those with less social capital. The basis for this examination
stems from a capacity-outcome ‘fit’ argument which proposes that if social capital is a
more necessary condition for certain activities/outcomes, then DVCCs lacking this
capacity may be less likely to spend their energies focused on those outcomes. For
example, if the assertions of Himmelman (2001) and others (e.g., Gajda, 2003; Hogue,
1993; Reilly, 2001) are accurate and DVCC activities such as developing new programs
require greater relational capacity, then DVCCs lacking in such capital may choose to
spend their time and energies on other activities more fitting to their existing relational
capacity such as information sharing. These questions were examined by exploring the
following propositions:

P l: CERTAIN TYPES OF SOCIAL CAPITAL WILL BE MORE
IMPORTANT DEPENDING ON THE OUTCOME OF INTEREST

P2: THE TYPE OF OUTCOME A DVCC DEDICATES THE MOST TIME TO
WILL BE RELATED TO THE LEVEL OF SOCIAL CAPITAL WITHIN THE
DVCC
Summary
In summary, IAs like DVCCs are called upon to improve their community’s
response to their targeted social issue, in part, through fostering the development of social

capital among community organizations and agencies who share a common investment in

that issue. The rationale for such strategies is based on the premise that addressing

39

complex social issues such as domestic violence is beyond the scope of any single agency
or organization and will therefore require interorganizational cooperation. This study
sought to understand the role of social capital in one prominent type of IA (i.e., DVCCs)
convened to improve the community’s response to domestic violence. Specifically, this
study sought to (I) understand the relationship between social capital and overall DVCC
effectiveness, (2) investigate how the composition of the DVCC membership influences
the relationship between social capital and DVCC effectiveness, lastly, (3) explore
whether the strength of the relationship between social capital and DVCC effectiveness
varies depending on the type of effectiveness outcome under consideration. In order to
accomplish this, the following hypotheses/propositions were investigated:

H1: Higher levels of social capital among DVCC members will be positively
related to the effectiveness of the DVCC in fostering community change

H2: Higher levels of social capital among DVCC members will be positively
related to DVCC effectiveness at building the organizational capacity of
members

H3: Higher levels of social capital among members will be more strongly related to
DVCC effectiveness when there is also greater diversity in DVCC

composition.

P1: Certain types of social capital will be more important depending on the
outcome of interest

P2: The type of outcome a DVCC dedicates the most time to will be related to
the level of social capital within the DVCC

4O

 

 

 

 

pc

 

 

 

RESEARCH DESIGN AND METHODS
Overview

In order to investigate the proposed hypotheses and propositions, this study
conducted a cross sectional survey and social network analysis of DVCC members and
leaders. Because members and leaders are nested within DVCCs, a multi-level approach
was utilized. In multi-level designs, participants are clustered into groups and therefore
there is variance both within and between groups (i.e., DVCCs). In this study, multi-
level models were used to test whether differences among DVCCs with regards to
members’ average perceptions of effectiveness could be accounted for by DVCC level
characteristics. Variables in which there can be variance both within (e. g., among DVCC
members) as well as between DVCCs are referred to as Level I variables. All of the
outcome variables proposed in this study are Level I variables as they were measured and
entered into analyses as member reports. However, variables in which there can only be
variance between groups are referred to as Level 11 variables. For Level 11 variables, the
variable is either measured or calculated such that there is only one data point per group
that represents a characteristic of that group as a whole. All predictors in this study were
operationalized as Level 11 variables as they are all operationalized as characteristics‘of
DVCCs. Accordingly, the terms Level I and Level 11 variables are referred to in the
following descriptions of measurement and analysis designs for this study.

Procedures
Setting
The sample population for this study was Michigan DVCCs. Michigan offers a

particularly appropriate setting for conducting this research for a couple of reasons. First,

41

DVCCs exist in adequate numbers within the state of Michigan (11 = 57) to afford the
statistical power necessary to detect differences with regards to the proposed
relationships. This is a substantial advantage as the inclusion of DVCCs from multiple
states would constitute a third level of analysis (i.e., state context) that would
significantly increase the required sample size to account for such nesting without greatly
increasing the scope of the knowledge to be gained.

In addition, this study benefited from the active involvement and support of two
state level partners: Michigan Coalition Against Domestic and Sexual Violence
(MCADSV) and the Michigan Family Independence Agency Domestic Violence
Prevention and Treatment Board (MDVPTB). MCADSV is a non-profit organization
that provides policy level advocacy for, and support to, domestic and sexual violence
programs in the state of Michigan. The MDVPTB is the state agency responsible for
administering state and federal funding for domestic violence programs and services,
developing and recommending policy, and developing and providing technical assistance
and training to domestic violence programs across the state. Both MCADSV and the
MDVPTB have worked closely with Michigan DVCCs and subsequently have extensive
knowledge of, and strong credibility with, the DVCCs. These community partners
played a valuable role in strengthening the feasibility and utility of this study by helping
to inform measurement development, assisting with participant recruitment, and
participating in the interpretation and dissemination of findings.

Sample
The sample for this study was recruited from the population of 57 DVCCs located

within the state of Michigan. Descriptive information concerning the DVCC and its

42

 

 

rh.

or;

or,

II

[t

 

 

membership was obtained from DVCC membership rosters, member surveys, and reports
from DVCC leaders. DVCC leaders were identified by state level partners as the primary
contact person listed for a particular DVCC based on their state databases. Most
frequently the identified DVCC leader was the chairperson or a steering committee/board
member (75%). However some leaders identified themselves as either an informal leader
within the DVCC (19%) or a DVCC staff member/coordinator (6%). All leaders had
been involved with their DVCC a minimum of one year. On average, leaders had been
involved with their DVCC an average of 8 years and had served in a leadership position
an average of 6 years.

Due to the multi-level nature of this study, there are sample characteristics at two
levels that must be considered. At the highest level, Level 11, sample considerations
focus on the number of DVCCs represented in the study. Nested within each of the
DVCCs participating in the study are organizational representatives who attend DVCC
meetings on behalf of a given organization, agency, or group. Therefore, at Level 1,
sample considerations are focused on the number of DVCC members who participated in
the study. The term DVCC member refers to a person who represents a specific
organization, agency, or group. As such, Level I can also be thought of as the number of
organizations represented in the study. Sample characteristics for both Level I and Level
II are described in detail below.

Level II: DVC C Sample Characteristics

All 57 DVCCs currently active2 in the state of Michigan were invited to

participate. However, because this study relied on member level data for operationalizing

 

2 A currently active DVCC was defined as a DVCC comprised of three or more members that convened at
least once in the past year

43

some DVCC level constructs, it was important that the data be representative of the
DVCC membership. Therefore, in order for a DVCC to be included in analysis, a
member response rate of 70% or greater was required. This meant that surveys had to be
received from at least 70% of the organizations and agencies currently represented on the
DVCC in order for a DVCC to be included in the study. Current members were
identified by DVCC leaders and were defined as organizations/agencies that were
represented at at least one DVCC meeting in the past year. F orty-eight DVCCs (84%)
met this criterion for inclusion. According to leaders, these DVCCs on average had been
around for approximately 10 years but ranged in their age from 2 to 21 years old.
DVCCs ranged in size from 6 to 61 organizations with an average size of 16
organizations/agencies per council. Public agencies made up the largest proportion of
these (58%) followed by nonprofit organizations (30%), community based groups (8%)
and public agencies (4 %).
Level I: DVC C Member Sample Characteristics

DVCCs are comprised of member organizations and agencies that are
represented by a leader or staff person who attends DVCC meetings on behalf of their
organization. The individual who served as the representative of a given organization at
DVCC meetings was invited to participate as the key informant representative of their
organization in the study given their requisite knowledge of both their organization and
of the DVCC. However, only one representative per organization or agency was
included in the study. In the event there were multiple individuals attending DVCC
meetings as representatives of a single organization, a key informant was identified by

DVCC leaders. Leaders were asked the level of activity of each representative based on

44

1\ ‘L It

I iii Ir. \
l
f‘

l

 

 

whether they had attended a meeting in the past year and if they had attended over one
third of the meetings. If one representative was indicated to be more active than the
other(s), the most active representative was chosen as the key informant. If there were
two representatives equally active in the council (e.g., both having attended a third of the
meetings or more), leaders were asked to nominate which representative they felt was
most knowledgeable about their organization and the impact of the DVCC on the
organization. For 99% of organizations/agencies, surveys were received from the sole or
key informant representative on the DVCC. For the remaining six cases in which
primary key informant was either unavailable or declined to participate, data was
collected from a secondary informant. All secondary informants were currently involved
with the DVCC, had attended no fewer than one meeting in the past year and had an
average attendance of 70%.

The Level I sample relates to the number of DVCC members who participated in
this study. The sample population at Level I was the total number of organizations
represented by current members of the DVCC. Within the 48 DVCCs included for
analysis, there were members representing 739 different organizations and agencies.
Representatives from six hundred and thirty-eight (86%) of those organizations elected to
participate.

Table 1: Res onse Rates for Levels I and Il

 

 

Level II DVCC 57 48 84%
Level I Organizational // 0
members of DVCCS 739 638 86 /°

 

 

 

 

 

 

45

The amount of time informants reported having been involved with their DVCC
ranged fi'om less than 1 to 21 years but on average informants reported having been
involved for five years and attended an average of 65% of the meetings over the past
year. They reported having been involved with their organization or agency for an
average of 12 years and represented roughly equal distributions in terms of their
management level with 30% of the sample directors/head administrators, 39% mid level
managers/coordinators, and 31% staff members. The sample was somewhat older (age
range from 24 to 76 with mean of 48) and consisted of more women (64%) than men
(36%). Eighty-five percent of the sample was white. Given the high proportion of
DVCCs and members represented in the sample upon which these estimates are based
and the fact that these estimates are consistent with descriptives obtained from previous
research with this population (Allen, 2001), there is evidence to suggest that they are
generally reflective of actual DVCC composition in the State of Michigan at this time.

Data Collection
Data collection for this study was carried out in two consecutive phases: (1) a phone
interview with DVCC leaders, and (2) a survey of DVCC members.
Phase 1: Interviews with DVC C Leaders

Phase 1 consisted of interviews with leaders from each participating DVCC in
order to collect data on DVCC membership and characteristics.

Recruitment. Leaders were first contacted by letter and then by phone to explain
the purpose of the study and to invite their participation. DVCC leaders’ participation
involved several elements including: (1) providing an up-to-date DVCC membership

roster that included the names, organizational affiliation, and contact information of all

46

members, (2) participating in a 45 minute phone interview, and (3) facilitating member
recruitment by sending out an email and/or alerting members that a survey would be
forthcoming. In order to increase the legitimacy of the project in the eyes of the leader
and thus to encourage participation, the recruitment letter was sent out jointly from
MCADSV and the principal investigator (see appendix A). However, leaders were
informed that their participation was voluntary and that their involvement or lack thereof
would be kept confidential from all parties external to the research team including
MCADSV or any other state level partner. At this phase, 523 DVCCs (91%) agreed to
participate.

Phone interviews. For those who elected to participate, a date and time was
arranged for a 45 minute phone interview. The purpose of the leader interviews was
primarily to gain descriptive information about the size, composition, structure,
resources, and activities of the DVCCs. Informed consent for these interviews was
obtained by emailing or mailing a copy of the informed consent to the leader and having
them fax or mail the signed consent form back to the researcher prior to the interview. In
preparation for the interview, leaders were asked to submit to the researcher an updated
membership roster for their DVCC. This roster was then electronically imported into a
standardized Excel database format. In addition, in order to allow leaders the opportunity
to gather any needed information concerning their DVCC, leaders were sent a handout in
advance containing many of the questions asked during the phone interview (See
Appendix B). The principal investigator conducted the phone interviews in two stages.
The first stage involved systematically going through the membership roster database

with the leader. For each member listed on the database, the leader was asked to (1)

 

3 Note: Four DVCC were later excluded fi'om the final sample due to response rates of less than 70%

47

confirm the information provided; (2) indicate the level of activity of the member on a
scale of non-active (hasn’t attended a DVCC meeting in the past year), sporadic (has
attended a meeting in the past year but has attended less than 1/3 of the meetings) and
active (has attended 1/3 or more of the meetings); and (3) categorizing each member into
a specific stakeholder group (e.g., law enforcement). Leader responses to each of these
were entered directly into the Excel member database for that DVCC as the interview
was taking place.

The second stage was to collect further descriptive information concerning the
DVCCs using an interview protocol containing both fixed response and semi-structured
questions (see Appendix C). Responses were recorded directly onto the interview
protocol and then later entered into an SPSS database by undergraduate research
assistants for analysis. In order to ensure accuracy in data entry, the data from each
interview was double entered into two separate SPSS database by two different students.
The two databases were then electronically compared and all discrepancies resolved until
the two databases were identical.

Stakeholder group categorization.

In preparation for both developing the social network measures as well as
operationalizing stakeholder breadth, leaders were also asked to categorize all their
members into stakeholders groups as part of the leader interview. Stakeholder
categorization was accomplished using a couple of techniques. A list of possible
stakeholder group categories (See Appendix D) was initially based off of previous
research with this population (Allen, 2001) and was expanded based on the leader

interviews. In cases in which the stakeholder group appeared obvious based on an

48

organization’s name listed in the membership roster, the principal investigator would
simply confirm the categorization with the leader. For example, if an individual on the
membership roster was listed as a representative of the Laskon County Police
Department, the principal investigator would confirm with the leader that this individual
served on the DVCC as a representative of law enforcement. In cases in which the
stakeholder group was not obvious based on the organization name, leaders were asked
what stakeholder group this individual represented on the council. The concept of
“stakeholder group” was one that appeared familiar to leaders and generally, they were
able to categorize their members into stakeholder groups without hesitation. In those
instances in which there was some uncertainty as to the correct categorization, the leader
was asked to describe what the organization or agency in question does and how it relates
to the work of the DVCC. Then the principal investigator and leader would go over the
list of possible stakeholder groups and come to consensus on what the most appropriate
stakeholder group categorization for that member would be. If the organization did not
fit into any of the existing stakeholder categories, it was added as a new stakeholder
group.
Phase 2: Survey of D VC C Members

Phase 2 consisted of a survey of DVCC members in order to collect data on their
perceptions of DVCC effectiveness at implementing changes in their community, and the
extent to which their participation in the DVCC has strengthened the capacity of their
organization. This phase also included a social network analysis of DVCC members to

gather data on the level of social capital within the DVCC.

49

Recruitment. A representative from each organization involved with the DVCC
was recruited to participate in the present study. This representative received a letter sent
jointly by the DVCC leader, community partners, and the principal investigator
explaining the purpose of the proposed study and inviting their participation (see
Appendix E). Leaders were also encouraged to take a copy of the letter to share at the
next DVCC meeting and/or send out an email alerting members to the forthcoming
survey. In order to encourage participation, participants were informed that participating
in the survey could benefit their DVCC in two ways. First, study findings and
implications for practice were written up into a practitioner report that was distributed to
each of the DVCCs. Second, participating DVCCs with a response rate of 75% or greater
were entered into a lottery to win one of five $1000 awards that were awarded directly to
the DVCC to support planning and programming efforts.

Data collection. Survey data was collected primarily via an online survey
accessible to participants through a secure website. However, an alternative paper/pencil
version was made available for those lacking access to, or comfort with, the web-based
format (see Appendix F). Seventy percent (70%) of respondents completed the online
version of the survey. One week following the initial announcement, DVCC members
with available email contacts were sent an email containing a link to a website hosted on
a secure server and a unique ID number. Those who were willing to participate were
instructed to log onto the survey website using their unique ID number to ensure
confidentiality. Once they were logged on, participants were able to complete and submit
the survey online. The alternative paper version of the survey was mailed directly to

identified members who did not have an email address listed as part of their contact

50

information. Mailed surveys included a postage paid envelope and instructions to mail
the survey directly back to the research team.

F allow up. The initial survey data collection lasted three weeks and participants
were able to log on at any time during the survey period to complete the survey. In order
to increase response rates, a follow up postcard/email was sent week two (see Appendix
G). For those DVCCs who had not met the response goal at week three, the due date for
the survey was extended. Twenty-four percent (24%) of the sample completed the survey
following the initial due date. Because web addressed were sometimes incorrect or emails
were blocked by spam filters, all survey recipients who had not yet responded were sent a
paper version of the survey. Survey recipients (both web and mail) also received a
follow up phone call. The purpose of the follow up phone call was to inform participants
that the due date had been extended and a paper version of the survey had been sent to
them in the mail. Individuals who had previously received the survey via email were
additionally informed that they could fill the survey using either the web or the paper
version.

Measurement Development

Measurement development on scales created specifically for this study was
informed by qualitative interviews conducted with 15 key informants from across the
state of Michigan. The principal focus for these interviews was to inform the conceptual
framework and item construction for both the impact on member organizational capacity

and social capital measures. Each of these is discussed in turn.

51

Strengthened Organizational Capacity.

The dependent variable of primary interest in this study was DVCC effectiveness,
conceptualized as DVCC effectiveness at promoting community change and DVCC
effectiveness at strengthening the organizational capacity of members. While fieldwork
has recently been done to create measures of effectiveness of community change that are
ecologically valid to the context of DVCCs (i.e., Allen, 2005), such fieldwork has not yet
been conducted with regards to developing instruments for measuring the effectiveness of
DVCCs at increasing the organizational capacity of members. Existing knowledge of the
impacts of IAs on organizational capacity comes largely from case studies and
practitioner writing emerging out of other social issue domains (e.g., substance abuse;
AIDS prevention/intervention; youth violence prevention). Therefore, while past
research has demonstrated that participation in IAs, such as DVCCs, has the potential to
result in increased access to information and resources for agencies and organizations
(Bailey & Koney, 2000), this research merely creates the framework for measuring IA
effectiveness at increasing organizational capacity. Given that the information and
resources gained from IAs may take different forms specific to a given context, generic
references to such benefits are less than optimal for measuring the ways in which
organizational capacity is being built within the context of DVCCs (Linney, 2000). As
such, in order to develop a more ecologically valid measure of DVCC effectiveness at
strengthening member capacity, preliminary fieldwork was conducted to both inform the
development of ecologically valid items regarding the nature of information and

resources that are attained through membership in a DVCC, as well as to explore whether

52

there are other areas of organizational capacity that are impacted by DVCC membership
that are not prominent in the existing literature.

Preliminary fieldwork to inform measurement development of the Strengthened
Organizational Capacity scale consisted of qualitative interviews with key informants.
The sample for these interviews was attained by referrals from community partners
(MCADSV and MDVPTB) with the objective of obtaining a diverse sample of
information rich cases of individuals (Patton, 2002) who could speak to the range of
impacts participation in a DVCC can have on the organizational capacity of members.
Community partners were asked to nominate a set of individuals who (1) were deemed to
have extensive membership experience with, and knowledge of, DVCCs; (2) collectively
represent the primary stakeholder groups who are members of DVCCs (e. g., domestic
violence programs, courts, law enforcement, prosecuting attomey’s office, batterers
intervention programs; Allen, 2001); and (3) collectively represent both effective and less
effective DVCCs. Community partners were instructed to make their determination of
effectiveness based on their subjective perceptions of whether the DVCC has been
successful at positively impacting its members and community relative to the other
DVCCs. Interviews were conducted with key informants over the phone by the principal
investigator where they were asked to describe, based on their experience, the different
ways in which involvement in DVCCs can impact the capacity of organizational
members.

Data from these interviews were collected both by audio tape and written notes
taken by the principal investigator during the interview. To the extent possible, verbatim

quotes were recorded. At the conclusion of the interview, the principal investigator

53

listened to the interview tapes and filled in statements and quotes not captured during the
note taking process. The data for the Strengthened Member Capacity scale was then
analyzed using a content theme analysis approach (Patton, 2000) applied in the following
manner. First, all data relevant to the question of how organizations are impacted as a
result of their involvement in DVCCs was tabled together by case. Each statement
conveying a different type of impact was given its own line and assigned a case number
corresponding to the specific participant it belonged to as well as the stakeholder group to
which the participant belonged. Each statement was then sorted by the principal
investigator into 34 themes composed of statements all sharing a common meaning or
idea (e.g., increased knowledge about what other agencies/organizations do and how they
operate). These themes were then further organized into a preliminary set of categories
(e.g., increased knowledge and awareness). Themes were examined for the breadth of
stakeholders (out of a possible five) who made mention of that idea as well as the total
number of stakeholders. In order to enhance generalizability, those themes mentioned by
at least three of the five stakeholders were marked for special consideration. However, it
is important to note that failure on the part of any give stakeholder to attend to any
particular theme does not necessarily imply lack of generalizability to that stakeholder
group (Patton, 2002). Finally, each theme was then crosswalked against existing
literature describing the types of impacts to organizations suggested by scholars and
researchers.

Four impact types (information attainment/organizational learning, increased
capacity for problem solving, sustainability, and enhanced influence/visibility) emerged

from this process as prominently represented in both existing literature and the

54

exploratory interviews. These categories and their corresponding themes served as the
basis for generating scale items (see Appendix I for data analysis table summary).
Social capital.

While the measurement framework for social capital (i.e., the importance of
including both expressive and instrumental ties, considerations of quality as well as
structure) emerged from reviews of current literature on social network constructions of
social capital (e.g., Balkundi & Harrison, 2004; Ibarra, 1993; Reagans & Zuckerman,
2001; Gabbay & Leenders, 1999) - the qualities of ties (e.g., trust) were developed in
consideration of both current literature and qualitative inquiries with key informants.
Specifically, given that researchers have examined different types of expressive and
instrumental ties (e.g., fiiendship, advice giving), it was important to validate the network
qualities that might be most important within the context of this study. As such, key
informants were first asked the degree to which they thought relationships were a critical
factor in the effectiveness of a DVCC in order to pilot the central premise of this study
(e. g., that social capital is relevant to the effectiveness of IAs). Responses from all key
informants were strongly affirrnative in response to this question. Given that, key
informants were then asked to describe why relationships were so important, what
function they served, and what relationships needed to look like in order for the DVCC to
be most effective. The responses to these questions were examined and used to help

inform the development of the social capital measure in the manner similar to that already

described above.

55

Measures
Member Survey
DVCC Community Change Eflectiveness

The effectiveness of the DVCC at creating community change was initially
measured using a l9-item scale adapted from Allen (2001) which measures the
effectiveness of the DVCC at making the kinds of community changes that are described
by experts in the field as the purpose of DVCCs. Each item was measured on a six point
Likert scale ranging from “not at all” to “a great deal.” This scale included items focused
around five types of community change outcomes. These were the extent to which
members perceived their DVCC to be effective at: (l) improving coordination, (2)
developing new or improving existing programs/services, (3) changing policies, (4)
increasing public awareness and education, and (5) overall perceived goal
accomplishment (See Appendix H for full scale and subscale breakdown).

In order to examine dimensionality within the scale, an exploratory principal
components factor analysis with an oblique rotation was conducted. The results of this
analysis indicated that improved coordination was a related but separate outcome from
the other impacts which represent broader systems change types of outcomes (e.g., policy
change, program development, community education). Interestingly, goal
accomplishment items loaded most strongly on systems change items as opposed to the
coordination scale. In order to maximize the independence of these two scales, items
with cross-loadings of greater than .2 were eliminated. The two-factor solution
explained 72% of the variance in the scale. Based on this analysis, DVCC community

change effectiveness was operationalized as two separate but related outcomes that were

56

examined separately in subsequent analysis to provide a more fine-grained understanding
of the relationship of social capital to DVCC community change effectiveness. The
resulting scales are described in more detail below.

(a) Improved Coordination (alpha = .89). This scale consists of four items
pertaining to the effectiveness of the DVCC at increasing the level of
coordination among community organizations and agencies. Example items
include: (1) To what extent do you feel the efforts of your council have
increased the ability of organizations/agencies to coordinate their efforts? and
(2) To what extent do you feel the efforts of your council have resulted in
organizations and agencies working together more efficiently?

(b) Systems change (alpha = .92). This subscale consists of seven items measuring
the extent to which the DVCC is perceived to have been effective at
influencing policy, changing public attitudes, improving existing services,
developing needed services, and generally perceived to have made good
headway at improving how the community responds to domestic violence.
Example items include: (1) To what extent do you feel the efforts of your
council have influenced city, county, or state legislation concerning domestic
violence?, and (2) To what extent do you feel the efforts of your council have
expanded or improved existing programs or services for women and children
affected by domestic violence?

Strengthened Member Capacity
DVCC effectiveness at strengthening member capacity was initially measured on

a 17 item scale developed for this study. For this scale, participants were asked to rate

57

the degree to which their organization had been impacted as a result of its involvement
with the DVCC in a variety of areas. Each item was measured on a five-point Likert
scale ranging from “not at all” to “a great deal.” The firll scale consisted of items
representing four types of impact identified in key informant interviews and current
literature.

In order to examine dimensionality within the scale, an exploratory principal
components factor analysis with an oblique rotation was conducted. Analysis indicated
that the four types of organizational impacts were best represented as a two factor model.
Factor one subsumed both information attainment and problem solving into a construct of
organizational learning. This makes sense as both information attainment and problem
solving were fundamentally about gaining access to knowledge that can help to support
the organization. Factor two subsumed sustainability and legitimacy into a construct
about enhanced opportunity gained through increased access to resources and enhanced
legitimacy within the community. In order to maximize the independence of these two
scales, items with cross-loadings of greater than .2 were eliminated. The two-factor
solution explained 64% of the variance in the scale. Based on this analysis, DVCC
effectiveness at strengthening organizational capacity was operationalized as two
separate but related outcomes that were examined separately in subsequent analysis. The
resulting scales are described in more detail below.

1) Organizational learning (alpha = .92). This scale consists of seven items
related to the degree to which participation in the DVCC has led to
organizational learning opportunities. Example items include: (1) For my

organization, participation in the council has led to increased access to tools,

58

best practices, and/or other information that has informed the work of my
organization, and (2) For my organization, participation in the council has led
to increased knowledge about how to best interact with other organizations in
order to accomplish our objectives.

2) Enhanced opportunity (alpha = .87). This scale consists of six items related
to the degree to which participation in the DVCC has led to enhanced
opportunities for the organization by providing for both greater access to
resources and enhanced legitimacy within the community. Example items
include: (1) For my organization, participation in the council has led to an
improvement in our ability to obtain grants and/or other firnding
opportunities, and (2) For my organization, participation in the council has
increased our ability to affect public policy.

DVCC Social Capital

Social capital was measured using a social network analysis technique called a
roster questionnaire described by Wasserman & Faust (1994) using a scale developed for
this study based on current literature and key informant interviews. In this approach,
participants were asked to rate on a six item Likert type scale the quality of their
relationship with members from each stakeholder group represented on the DVCC (e.g.,
law enforcement, prosecutors, victim services) on five types of tie characteristics:

1) Communication frequency ( i.e., how often does your organization
communicate with the members of this stakeholder group?) This
characteristic was measured on a six point Likert scale ranging from ‘never’

to ‘every week.’

59

2) Shared philosophy (i.e., to what extent to do you feel the members
representing this stakeholder group share a common philosophy with your
organization/agency regarding what domestic violence is and how it should
be addressed?) This characteristic was measured using a six point Likert
scale ranging fi'om ‘not at all’ to ‘entirely.’

3) Responsiveness to concerns (i.e., to what extent do you feel the members
representing this stakeholder group would be responsive if you or someone
from your organization brought a concern or issue to their attention?) This
characteristic was measured using a six point Likert scale ranging from ‘not
at all’ to ‘entirely.’

4) Acknowledged Expertise (i.e., to what extent do the members representing
this stakeholder group offer a unique perspective or form of expertise that has
been valuable in understanding or/addressing domestic violence?) This
characteristic was measured using a six point Likert scale ranging from ‘not
at all’ to ‘entirely.’

5) Trust (i.e., to what extent can the members representing this stakeholder
group be trusted to follow through on what they say they will do?) This
characteristics was measured using a six point Likert scale ranging from ‘not
at all’ to ‘entirely.’

The roster used was created based on the DVCC membership list provided by the
DVCC leader and was tailored to include only those stakeholder groups who were
represented on that DVCC. The network analysis roster focused on stakeholder level (as

opposed to organizational level) relationships for a couple of reasons. First, stakeholder

60

group affiliation is a highly salient categorization in the context of DVCCs. DVCCs are
principally about improving coordination and collaboration among different stakeholder
groups such as law enforcement, prosecution, and victim services. As such, stakeholder
level relationships are of primary concern within DVCCs. Second, several theories of
social capital (e. g., structural holes theory, bridge theory, weak tie theory) propose that
the relationships represent the strongest form of social capital when they link together
diverse stakeholders because such links are more likely to provide access to non-
redundant types of knowledge and resources. This again supports conceptualizing social
capital as strong cross-stakeholder relationships (e.g., relationships between law
enforcement and prosecutors) within the context of DVCCs.

Lastly, on a pragmatic level, social network measures are highly time consuming
as they require participants to individually rate their relationships with each actor in the
network. If the network were operationalized at the organizational rather than the
stakeholder level, this would require certain DVCCs to provide rating for over 60
different organizations for each of the five indicators of social capital (i.e., over 300
network ratings). Given that a member response rate of 70% or higher was required in
order for a DVCC to be included in analysis, it was believed that organizational level
rosters would significantly diminished the Level 11 sample size and bias the sample
toward smaller DVCCs. Thus, the less burdensome stakeholder level roster was deemed,
both conceptually and methodologically, to be the most appropriate strategy.

The data obtained from members of a single DVCC was then aggregated into a
non-symmetrical stakeholder-by-stakeholder matrix that contained each stakeholder

groups’ perceived relationship with every other stakeholder group represented on the

61

DVCC. There was one matrix created for each DVCC for each of the five indicators of
social capital resulting in a total of 240 (48 DVCC * 5 indicators) matrices. Each of these
matrices was separately analyzed using the social network analysis software UCInet for
Windows to calculate 245 network density scores. Network density is the sum of all
scores across stakeholders within the network divided by the number of stakeholders
within the network (Wasserman & Faust, 1994). The resulting number represents the
average tie strength of the network on that indicator of social capital, has a range of l — 6,
and is an index of the degree to which all members have strong relational ties to one
another (Coleman, 1988; Wasserman & Faust, 1994).

It is important to note that for this study, there was no assumption of reciprocity
imbedded within the conceptualization or operationalization of DVCC social capital.
This allowed for differences in how stakeholders perceived their relationships with one
another (i.e., the extent to which stakeholders perceived one another to share a similar
philosophy) to be taken into account. Specifically, by not assuming a symmetrical data
matrix, lack of reciprocity was taken into consideration during analysis as it ultimately
diminishes the overall density score for the network.

The correlation matrix for the different indicators of social capital is presented in
Table 2. Results of this analysis indicated there was between a small to moderate
positive correlation between the different indicators suggesting that, unsurprisingly,
DVCCs with strong relationships on one indicator of social capital were more likely to
have strong relationships on the other indicators as well. This analysis provided support
for combining the five indicators into one overall social capital scale for testing

hypotheses 1 —3. The combined social capital scale had an alpha of .8, indicating good

62

internal consistency. This combined scale provides a measure of the overall strength of
relationships among stakeholders across the five indicators of social capital.

Table 2: Correlation Matrix: Indicators of Social Capital

 

 

 

 

 

 

Recognized
Com Freq Philosophy Responsiveness expertise
1. Philosophy
028*
2. Responsiveness
0.57*** 0.63***
3. Recognized
expertise
0.33* 0.53*** 0.65***
4. Trust in follow
through
0.31* 068*“ 059*" 0.62***

 

 

Member Demographics and Descriptive Information

Lastly, the member survey asked respondents to provide descriptive information
regarding the organizations they represent, their position within the organization, and
basic demographic information about themselves. Specific questions regarding
organizational characteristics included (1) the type of organization (e.g., police, domestic
violence shelter, prosecutor), (2) the length of time their organization had been
represented on the DVCC, and (3) how active their organization had been overall as a
member of the DVCC as measured by the percentage of meetings the organization has
been represented at over the past 12 months. Specific question regarding characteristics
of the respondent included (1) their position within their organization, (2) the duration of
time in that position, (3) the duration of time with the organization, (4) the duration of
involvement with the DVCC, and (5) the level of involvement with the DVCC (e. g.,

frequency of attendance, involvement in leadership roles). Respondents were also asked

63

to report on their age, race/ethnicity, gender, level of education, and socioeconomic
status.
Control Variables

In order to guard against spurious interpretations that could result from potential
confounds to the relationship between social capital and DVCC effectiveness, this study
controlled for three variables: DVCC age, DVCC size, and the familiarity across
stakeholders.

DVCC age. DVCC age is an important covariate as it has theoretical links to both
the development of social capital and to DVCC effectiveness. For example, from a
developmental perspective, scholars (e.g., Gray, 1989) have argued that IAs like DVCCs
tend to develop in predictable stages, with each stage associated with the development of
additional group capacities that can improve the DVCC’s ability to accomplish
productive outcomes. As such, older DVCCs may be viewed as more effective simply
due to their more advanced developmental stage. Similarly, social-relational theories
such as contact theory (Allport, 195 8) highlight the developmental nature of
relationships, viewing relationships as dynamic and subject to evolve over time and
through interaction. As such, to the extent the DVCC is serving as an intervention for
fostering the development of social capital among members, older DVCCs would be
expected to have higher levels of social capital than younger DVCCs. In this way, a false
positive relationship between social capital and effectiveness could be found as an artifact
of the fact that age co-varies with both of them.

DVCC size. DVCC size for this study was operationalized as the number of

people who are current members of the DVCC. Size has similar potentially confounding

64

links to both social capital and DVCC effectiveness as it is potentially easier for smaller
DVCCs to develop strong ties across all members than it is for larger DVCCs. DVCC
size can also be related to the outcome of DVCC effectiveness as the extent to which an
IA is able to recruit a large membership is often viewed as a positive indicator of
capacity.

Stakeholder familiarity. Stakeholder familiarity for this study was operationalized
as the extent to which all stakeholders within a given DVCC were familiar enough with
one another to rate the nature of their relationship. The necessity of this control results
from the nature of the valued data collected through the social network analysis used to
operationalize the indicators of social capital. For this study, participants were asked to
rate the nature of their relationship with every other stakeholder on a scale of one to six.
So, for example, in the trust to follow through scale, a relationship rated as a one
indicated that the participant did not trust that stakeholder group at all to follow through
on what they said they would do. However, an assumption embedded within this scale is
that the participant is familiar enough with the members representing that stakeholder
group to respond to the question. As such, in addition to the scale choices ranging one
through six, participants were also given the option of responding “I 'm not familiar
enough with them to know ” for each stakeholder group. This added a familiarity
dimension to the social network data that needed to be controlled in order to maximize
the accuracy of the social capital network data.

In order to do this, a separate density score was created for each indicator of
social capital based on the network of familiarity. If a participant was familiar enough to

answer the question, a one was assigned in the matrix; if they were not familiar a zero

65

was assigned. If there were multiple representatives of a given stakeholder group, one
familiar and one not, an average (i.e., .5) was assigned into the familiarity matrix. This
resulted in a separate matrix containing information about the extent to which each
stakeholder group was familiar enough with every other stakeholder group to respond to
the question. Each of these matrices were then entered into UCInet and a familiarity
density score for each indicator was calculated. These familiarity density scores were
then averaged across the five indicators in the same manner as the social capital scale.

The average level of familiarity was .87 with a standard deviation of .07. This

familiarity density scale score was entered as a covariate for all analysis4.
Missing Data

In light of the high degree of internal consistency demonstrated in all the scales
and the advantages of maximizing both power and representation within the sample, a
50% missing data criteria was used. As such, cases in which there was missing data for

more than 50% of the items within any given scale were excluded from all analysis.

Leader Survey
Composition Diversity
As described in the introduction, composition diversity considered both task
related diversity and social diversity. Measures of task related diversity included DVCC
membership tenure, the ratio of public, private, and non-profit organizations represented
on the DVCC, the breadth of sector representation, and the proportion of

executives/middle managers/staff that attend meetings. Social diversity was

 

" Proposition 1 was also tested using the familiarity score specific to the tested indicator of social capital.
Results were the same. The decision was made to report the combined familiarity score as it represents the
overall familiarity.

66

Operationalized as the ratio of men to women on the DVCC. Unfortunately, racial/ethnic
diversity could not be examined in this study as Michigan DVCCs are largely
homogeneous and Caucasian with regards to their racial/ethnic composition and therefore
there was not adequate racial/ethnic diversity to examine its relationship to social capital
or DVCC effectiveness.

In order to obtain composition diversity measures, DVCC leaders were asked to
provide a list and contact information for individuals who are members of the DVCC and
indicate which stakeholder group each belonged to (based on a list generated from
previous research with this population; Allen, 2001) and whether each individual was
active or inactive in their involvement with the DVCC. Leaders were also asked to
provide approximate proportions for the ratio of public to private/non-profit organizations
represented on the DVCC, the number of executives/middle managers/staff that belong to
the DVCC and who attended meetings, the number of men versus women, and number of
minorities to Caucasians. Due to the exploratory nature of this investigation, diversity
scores were calculated for each type of diversity (i.e., gender, breadth of stakeholders,
proportion of nonprofit, forprofit, and public). With the exception of breadth of
stakeholders, all diversity scores were calculated using Blau’s (1977) index of
heterogeneity. This index is used for categorical variables (1 ~2pi2) where p is the
proportion of the group in the particular demographic category and i is the number of
groups represented. Breadth of stakeholders was simply calculated as the number of

stakeholder groups represented by current members of the DVCC.

67

DVCC Allocation of Time and Energy

Because DVCCs work on a variety of different tasks and services based on the

needs of the community they serve, DVCC leaders were asked to describe how their

council allocated its time and energy across nine different activities over the past year.

Specifically, leaders were asked to proportion the time and energy their DVCC allocated

to:

1)

2)

3)

4)

5)

6)

7)

8)

9)

Dealing with issues related to council firnctioning such as recruiting
members, establishing council structures (e.g., bylaws), strategic planning,
and dealing with internal conflicts

Sharing information among members such as introducing members to what
each organization/agency does and sharing tools, resources, or best practices
pertaining to domestic violence response and prevention

Developing protocols to better coordinate the practices of different
organizations/agencies

Identifying gaps and facilitating the development of new programs or
services

Working to improve existing programs and services
Administering/managing council run programs

Working to change policies related to domestic violence

Engaging in efforts designed to increase public knowledge and awareness of
domestic violence including pertinent resources available in the community
Engaging in primary prevention activities such as changing public attitudes

and beliefs on violence, gender norms, and healthy relationships

68

Leaders were asked to allocate percentages such that the total time spent across all
activities equaled 100%. Leaders were also provided with an “other” category if there
was an activity engaged in by the DVCC that could not be categorized in any of the
above nine activities.

Data Analysis
Data Entry
Quantitative data attained through the web-based survey was automatically
entered into a MySQL database at the time the survey respondent filled it out. This
database was then imported into SPSS. Quantitative data attained through phone
interviews as well as mail surveys was entered into the SPSS database manually by
undergraduate research assistants. In order to protect against human error in data entry,
several techniques were utilized. First, a master template of both survey instruments was
developed by the principal investigator which provided detailed data entry instructions
and protocols for each scale or measure. Data enterers were then trained as a group by
the principal investigator on the survey instruments, the data entry protocols for each
scale, and the SPSS program. Data enterers were then required to demonstrate their
competence in applying the protocol in a mock data entry task and any mistakes made
were discussed and clarified in one-on-one sessions with the principal investigator and
corrected by the data enterer. Second, all surveys were entered independently by two
data enterers into two separate SPSS databases. These databases were then merged and
answers were cross-referenced to check for errors. All inconsistencies and errors were
researched and resolved prior to integrating the data received through the web-based

survey.

69

Because the web-based and the paper/pencil versions of the survey collected the
same information using different methods, it was important to examine whether there
were significant modal differences. Separate one-way ANOVAs were conducted on
both outcome variables (systemschange effectiveness and coordination effectiveness).
Consistent with findings from experimental research examining differences between web
and mail survey approaches (Kieman, Kieman, Oyler, & Gilles, 2005), result of this
analysis indicated there was no significant systematic response bias with regards to the
DVCC member survey related to differences in measurement methodology (i.e.,
paper/pencil versus web-based survey). Results of this analysis are summarized in Table
3.

Table 3: One-way AN OVAs for Mai] Versus Web Survey Respondents

 

 

 

 

 

Degrees of F P-value
Freedom
Coordination Between 1 2.02 . 1 56
effectiveness
Within 609
Systems change Between 1 .225 .635
effectiveness
Within 577

 

 

 

General Analysis Strategy
The proposed model is conceptually a model that investigates DVCC
characteristics to explain differences among DVCCs in their effectiveness. However,
statistically, outcomes occur at the member level (Level I) while predictor variables occur
at the DVCC level (Level II). This provides the opportunity to propose and analyze

multi-level models. In the multi-level models, a DVCC’s score on a given outcome is

70

represented by multiple data points which are nested within that DVCC. Multi-level
analyses for this study utilized hierarchical linear modeling (HLM) techniques to account
for the non-independence of data that resulted from the fact that members were nested
within DVCCs. HLM is a particularly powerful data analytic technique for this research
as it allowed the researcher to examine multi-level relationships within nested data
without violating the assumption of uncorrelated error terms as is problematic in OLS
regression analysis (Hofmann, Griffin, & Gavin, 2000). Data collected at the individual
member level (e.g., strengthened organizational capacity of members) can be used to
examine group level effects by testing whether such effects cluster by DVCC. The
purpose of these models within the current study was to determine whether, on average,
members from different DVCCs varied significantly from one another when within group
agreement was taken into consideration. Per Bryk & Raudenbush (1992) all predictors
were grand mean centered prior to analysis for the purposes of interpretability. For a

summary of constructs and their levels of analysis, see Table 4.

71

Table 4: SUMMARY OF MODEL CONSTRUCTS

Outcome variables

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DVCC coordination effectiveness Member Level I
survey

DVCC systems change effectiveness Member Level I
survey

DVCC effectiveness — promoting organizational learning Member Level I
survey

DVCC effectiveness — enhancing organizational opportunities Member Level I
survey

Predictor variables

DVCC social capital Member Level 11
survey

DVCC gender diversity Leader Level 11
survey

DVCC sector diversity Leader Level II
survey

DVCC tenure diversity Leader Level 11
survey

DVCC management diversity Leader Level 11
survey

DVCC Breadth of stakeholders Leader Level 11
survey

Covariates

DVCC age Leader Level 11
survey

DVCC size Leader Level 11
survey

Familiarity of Stakeholders Member Level 11
survey

 

72

 

RESULTS
Univariate Analysis
Prior to analysis, descriptive information on all independent and dependent

variables was examined in order to investigate the distributions of variance and to ensure
that all continuous variables were normally distributed within acceptable limits (i.e.,
skewness of no more than +/- 1) for the HLM analysis. A summary of descriptive
information is provided in Table 6 and Table 7 for all independent variables and in Table
5 for all dependant variables. Generally, all variables fell within range of acceptable
limits of normal distributions with the exception of management diversity, gender
diversity, and DVCC size. These variables were normalized based on the nature of their

skewness and kurtosis using transformations recommended byTabachnek & Fidell

(1996)5. Descriptives on all transformed variables are summarized in Table 8.
Social Capital

On average, the density of relationships across all five indicators of social capital
(i.e., combined social capital scale) indicated that DVCCs generally have moderately
strong relationships across stakeholders. With a possible range of 1 — 6, overall social
capital density ranged from 3.69 to 5.28 with an average density 4.38. Examining the
five indicators of social capital individually, mean densities ranged from 4.14 to 4.58 and
standard deviations ranging from .31 to .5. Communication frequency outside of DVCC
meetings had the lowest density (mean = 4.14); trust in follow through had the highest

(4.54).

 

5 All reported analyses were computed using transformed variables. Analyses were also run using
untransformed variables and results did not change.

73

Diversity.

Gender Diversity

Overall, DVCCs were composed of mostly women (average 68%).
Approximately 25% of DVCCs had roughly equal proportions of men and women. Only
three DVCCs had proportionately more men than women and the greatest of these was
comprised of 65% men. Given that there are only two categories for gender diversity, the
Blau index for gender diversity theoretically can range between 0 and .5. A Blau index
of .5 indicates that there are equal proportions of men and women in the DVCC. Within
the sample, the average Blau index for gender diversity was .39 but ranged from 0 to .5.
Tenure Diversity

According to leaders, on average across DVCCs, 13% of DVCCs members joined
in the past year, 34% had been involved one to three years, and 53% of members had
been involved for four years or more. Given these three categories, the maximum Blau
score theoretically possible was .67. For DVCCs in this study, the average Blau score for
tenure diversity was .40 with a range from 0 to .65. This indicates that on average,
DVCCs tended to be somewhat diverse in their composition with regards to the number
of years members had been in involved with the DVCC but that there was also a great
deal of variability across DVCCs in the extent of tenure diversity present in the
membership.
Sector Diversity

With regards to sector, the majority of members who participated in DVCC
represented public institutions (5 8%), followed by nonprofit organizations (39%).

Members were least likely to represent for profit organizations (3%). With three

74

categories, the highest Blau score theoretically possible was .67. In this study, Blau
scores for sector diversity ranged from .18 to .66 with an average diversity score of .44.
Management Diversity
DVCCs in this study were characterized by a relatively high level of diversity in

their compositions with regards to levels of management, being comprised on average of
roughly balanced proportions of executives/upper managers (28%), middle
managers/coordinators (3 9%), and staff (33%). With three categories, the highest Blau
score theoretically possible was .67. In this study, Blau scores for management diversity
ranged from .00 to .67 with an average diversity score of .56.
Breadth of Stakeholders

While there were 25 different stakeholder groups identified across all DVCCs, the
number of stakeholder groups present within any one DVCC ranged from 3 — 15 with an
average of approximately 8.7 stakeholders per DVCC. Stakeholder groups on average
were comprised of 1.8 organizations/agencies. DVCCs varied quite a bit in their
composition. The most commonly represented stakeholders included domestic violence
shelters/service providers (96%), law enforcement (90%), prosecutors (85%), courts

(71%), and children’s social services (62%).

75

E8 Samoan

:05 E c0235 2d mozatomou .aaeemecoctfiu 33$"ch smokes 05 E mUO>Q 5053 woos—2&6 =0 coma—co.“ beam £5 E mug—mew 24: 05 omzmoom s

 

 

 

 

 

 

 

 

 

 

 

 

ovd we; omd 36 mice:
and mad. Ned- mod- mmogoxm
and Re Se :3 season am
God om...” 05m wme :82
warm Hme mwé .16 Essexez
wed end mod cod 8588:)“
am; 32 “tum mod omen
we we we we Z
_h~mcwmwmwmo weir—«ou— EEfiaNaneo ohhmMW’WHMM—mm MMHWMWHWMW moumzfim

 

 

emu—eats» «ea—Eon»: eoaawouwuae .5 8.53589 "m 93.2.

76

 

 

 

 

 

 

 

 

 

 

mad omd- a; 2.0 $6 a; 3; mod £832
3N hmd and mmd- $6 $6 one wed mmogoxm
. . . conserva—

mm Q 2. w and de Ed emd de 3 o . .
.9m
5. 5 m2: wmé wmé o: owe ové 36 :82
8.2. 8. MN wad Rd arm hmw mm.m chm 83:38:
8.5 cod ac...” cwm ow.m om; v2” 3.». 82852
8.3 8.3 an; S; cm; 5; NM: Ned omcdm
we we xv we we we we we 2

9.2.52: max 00>: 33m SEE 03.22.me m8: Enema—Em 55:52am
«nounsu we a $395 incem cofinweuom ozmaoamom eBay—m Etna—558:0
eon—5:50

 

 

mo_aa_..«> Eugene—Em .5 agree—omen no 035.

77

 

 

 

 

 

 

 

 

 

78

Gd :6. ”we. N2 N2 28ng
:6 So- Rd- ”2. m3- season
m2 :6 :.o o; is census .sm
5w 2.5 as one as as:

8.: 3o 85 So one page:

2:. cod w; 03 cod 8:832

8.2 ed as ”no one swam

we a. a, a. a. z
Eugene—5m he 5.33.5 mam—95G 955,—. hwmuwomn flew—””Muhwwmnuz WWW—“MM”— mocmcfim

 

 

3.83 833.3» Eon—Snow:— :e 8.53.88: K 035.

 

 

 

 

 

 

 

 

 

 

 

 

can ”so- So. ewes
mud Ed- mwd- mmogoxm
end a _ .o vmd :oufi>oQ .Em
om; Inc mud :83:
co. _ 36 oo. 3 BEES—L
mwd de mmd 5:882
00.3 $6 mud away“—
we we we WH—
eofiueumnfih vegan—ugh vegan-.35. 852.com I—
85 00>: £an: .5885: SEE: segue . .

 

833:3» gouache—E— van—hemmeach .8 actuators: "a o3a.—.

79

DVCC Allocation of Time and Energy
Univariate analysis of where and how DVCCs allocated their time and energy found
that, across all DVCCs, leaders reported the greatest amount of time over the past year
allocated to activities designed to increase community awareness of domestic violence
and sharing information among DVCC members. Leaders reported their DVCC spending
the least amount of time and energy creating and maintaining DVCC administered
programs.

One of the principal purposes for collecting information conceming where
DVCCs were dedicating their time and energy was to characterize DVCCs in terms of the
type of activity to which they dedicated the greatest amount of effort in order to explore
Proposition #2. Proposition # 2 was intended to explore whether the level of social
capital within the DVCC was related to the type of activity to which the DVCC dedicated
the most time and energy. However, imbedded within this proposition was the
assumption that DVCCs did, in fact, prioritize certain types of activities over others. In
order to test this assumption, two variables were calculated. The first variable,
representing high diversity of activity, consisted of the sum of activities to which each
DVCC spent 20% or less of their time and energy. The second variable, representing low
diversity of activity, was calculated as the number of activities to which DVCC spent
more than 20% of their time and energy (see Tables 9 and 10). '

Forty-four percent of DVCCs allocated small (between 1% and 20%) amounts of
time across seven or more different types of activities. On average, DVCCs had only one
activity to which they dedicated more than 20% of their time and 21% of DVCC spent no

more than 20% of their time and energy on any one activity. These findings suggests

80

that, contrary to the assumption of Proposition 2, most DVCCs diversify their time and
attention across a variety of activities rather than focusing on a few core activities. Based
on this preliminary analysis, a core assumption of Proposition 2 was deemed unsupported
and no further analysis was done.

Table 9: Number of Activities Council Spent Between 1% and 20% of time (out of
10 possible)

umber of activities with
1 -20% time allocation %
0 4.0%
2.0%
11.0%
11.0%

11.0%
17.0%
25.0%
13.0%
6.0%

'OOOQQUIADJN

 

Table 10: Number of Activities Council Spent 20% or More of Time (out of 10
possible)

 

’ Number of activities with greater than 20% time
llocation %

 

 

 

 

 

 

 

 

 

"' '; ."' " "' ‘ w ,1 ¢ r.;r ' '.".-.-- ."; tr. . r.
'- ’1 , .‘ ‘7 - -316153.‘ 1353...). int-‘1: P332"; 3" 3:19?» -.¢ ‘3 .‘ -.-. ft.s..‘:"..-.'..~..‘:.:< :n.‘ 012's; ..;‘% ’7. 1.}. H.115: ~." $51.? .. I

0
1
2 22.0%
3
4

airy-:4
J~ "
a

r o
crd'm tws'lé 3"! 1.94:. J- i" ' .11,.‘~:“”"- :.v1_i‘:M' " an. i‘rumfilia r“ " 'iflC‘V-u '5 l..:i'!~-'.l‘L‘..s.‘..J-‘.Iu. ;-.:o:~:"«h.imf.ald r n- n‘ #3.; . u 2.. ..

. 2,113.1. . . A‘sa‘u «9.3“: dumb. sank-'1. wen. . .

DVC C Eflectiveness at Improving Coordination
In order to examine if a multi-level model approach was needed, a null model was run in
HLM which modeled only the differences in intercepts between DVCCs (See Table 11).

The results of this analysis indicated that there was significant variation both within

81

members of a single DVCC as well as between DVCCs. This validated the need to
utilize a multi-level analysis strategy that allows both sources of variance to be taken into
account (Hofinann, Griffin, & Gavin, 2000). The interclass correlation (ICCl) for the
coordination effectiveness outcome indicated that 10% of the variance results from
differences in the intercepts while the remaining 90% of variance resulted from
differences between participants. This means that even though there were differences
among members within any one DVCC, on average, DVCCs were also significantly
different from one another on their average perceived effectiveness. This suggests that
DVCC level predictors (e. g., social capital) can be useful in helping to explain these
differences between DVCCs.

It is also important to note, however, that a substantial portion of the differences
in perceived effectiveness occurred between participants regardless of the DVCC to
which they belonged. This indicates that the extent to which a participant viewed their
DVCC to be effective was greatly influenced by who that participant was or by factors
related to that participant or their organization that were not unifome shared by the
other members of the DVCC.

Table 11: Estimate of Covariance Parameters for Coordination Effectiveness

 

 

 

 

Estimate Std. Error Wald Z P-value
Residual .92 .06 16.34 .000
Intercept Variance .1 .04 2.53 .011

 

82

 

DV C C Eflectiveness at Promoting Systems Change

In comparison to coordination effectiveness, participants were somewhat less
affirmative in their assessments of the degree to which they felt their DVCC had been
effective at promoting broader systems changes in the community (mean of 3.82
compared to mean coordination effectiveness of 4.44). A paired sampled t-test indicated
that these means were significantly different from one another (t = 18.7, df = 553, p =
.000) indicating that, on average, members rated their DVCCs significantly better at
improving coordination than promoting systems change. Systems change effectiveness
had a comparable degree of variability among participants relative to coordination
effectiveness with a standard deviation of 1.08 and a range of 5.

As shown in Table 12, HLM analysis of the null model examining the source of
this variation indicated significant between as well as within-DVCCvariance. ICC(1)
indicated 17% of the variation in perceptions of systems change effectiveness was
explained by differences between DVCCs and 83% explained by differences between
participants. This suggests that DVCC level factors may be valuable in helping to
explain differences in participants’ perceptions of DVCC systems change effectiveness.
However, as with coordination effectiveness, it is also important to note that there was
also substantial variation within DVCCs in perceptions of effectiveness indicating that
perceived effectiveness was impacted by both DVCC level as well as individual or

organizational level factors.

83

Table 12: Estimate of Covariance Parameters for Systems Change Effectiveness

 

 

 

 

Estimate Std. Error Wald Z P-value
Residual .99 .06 15.83 .000
Intercept Variance .21 .07 3.11 .002

 

DVC C Eflectiveness at Promoting Organizational Learning

On average, participants reported that involvement in the DVCC had helped them
and their organization to gain access to information and develop relevant knowledge
between “to some extent” and “quite a bit” (mean = 3.58). Responses encompassed the
full 5 point scale range with a standard deviation of .8. Examinations of the variable
distribution for the Organizational Learning scale indicated good variability across the
range of responses and skewness and kurtosis measures indicated a normal distribution of
responses within acceptable limits (skewness = -.44, kurtosis = .05).

However, as shown in Table 13, the HLM analysis of the null model indicated no
significant between-DVCC differences. In other words, while there was good variability
in the degree to which participants felt their DVCC had had a positive impact on them
and their organization with regards to organizational learning outcomes, the source of this
variance was predominantly between participants irregardless of their DVCC affiliation.
A one-way AN OVA, conducted as an additional source of confirmation, also found the
between-DVCC differences - relative to the within-DVCC differences- to be insignificant
(F=1.1, p = .29). This suggests that the extent to which a participant viewed their DVCC
to be effective at promoting organizational learning was fundamentally influenced by
who that participant was or by organizational level factors that were not unifome shared

by the other members of the DVCC. This finding was surprising as it contradicted a core

84

 

assumption embedded within the proposed Hypothesis 2, namely that certain DVCCs
would be viewed by their members as better than others at promoting organizational
learning opportunities. In light of the lack of differences between DVCCs on this
outcome, Hypothesis 2 as it related to DVCC effectiveness at promoting organizational
learning was not tested.

Table 13: Estimate of Covariance Parameters for Organizational Learning

 

 

 

 

Estimate Std. Error Wald Z P-value
Residual .64 .04 16.37 .000
Intercept Variance .001 .01 .125 .9

 

DVC C Eflectiveness at Enhancing Organizational Opportunity for Members

Relative to DVCC effectiveness at promoting organizational learning, participants
reported somewhat lesser impacts related to gaining access to more opportunities such as
increased grant competitiveness, utilization of services, access to resources, or policy
influence. On average, participants indicated that participation in the DVCC increased
the opportunities available to their organization “to some extent” (mean = 3.05). Paired
t-test analysis indicated this was significantly lower than the average organizational
learning outcome (t=20.75, df=565, p=.000). The Enhanced Opportunity Scale showed
good variability with responses normally distributed across all five response options and
a standard deviation of .9.

However, similar to the Organizational Learning scale, HLM analysis of the null
model showed no between-DVCC differences concerning member perceptions of
enhanced opportunities resulting from DVCC membership (See Table 14). One-way

ANOVA similarly indicated no significant between-DVCC differences relative to the

85

 

within-DVCC variation (F =1 .04, p = .4). As all predictors proposed in this study are
Level II DVCC characteristics, Hypothesis 2 was deemed unsupported and organizational
impacts were excluded from further analysis as dependent variables.

Table 14: Estimate of Covariance Parameters for Enhanced Opportunity

 

 

 

 

Estimate Std. Error Wald Z P-value
Residual .81 .05 16.96 .000
Intercept Variance .000* .000

 

* Covariance parameter is redundant. The test statistic and confidence interval cannot be
computed

Modeling Perceptions of DVCC Effectiveness

Testing Hypothesis 1, namely that higher levels of social capital among DVCC
members would be positively related to the effectiveness of the DVCC at fostering
community change through both improving coordination and promoting broader systems
change, was the focus of the first phase of analysis. For the purposes of this analysis,
DVCC effectiveness at coordination and systems change were examined as separate
dependent variables and were modeled independently of one another. The overall
analysis strategy for testing Hypothesis 1 was to utilize HLM to model differences in
intercepts between DVCCs for each dependent variable and to test whether social capital
was a significant and positive predictor of differences in these intercepts. In order to
examine this question, an HLM equivalent to the hierarchical regression approach
proposed by Cohen and Cohen (1983) was used.

In this technique, the analysis for each dependent variable consists of two stages.
In the first stage, only the covariates (i.e., familiarity density, DVCC size, DVCC age)

were modeled as Level II predictors of DVCC effectiveness. In the second stage, the

86

 

social capital scale was added to the model to test Hypothesis 1. A comparison of the
difference between the ICC(1) between the first and second models was examined as an
indicator of the magnitude of variance explained by the addition of the aggregated social
capital measure. Further, the difference in the log likelihood ratio between the first and
second stage models relative to the increase in degrees of freedom was examined as an
indicator of improvement in model fit. A chi square test was then used to determine
whether this improvement in model fit was significantly larger than zero. Maximum
likelihood estimation was utilized on all HLM analyses in order to provide the
appropriate chi square distribution necessary for testing differences in Log Likelihood
ratios. Lastly, effect sizes were calculated using the effect size measure developed by
Raudenbush and Lui (2001).
The Relationship of Social Capital to DVC C Coordination Effectiveness

As described above, stage one of the analysis involved creating a model which
included only covariates (i.e., familiarity, age, and size) as Level II fixed effect
predictors. Covariates were not significantly related to coordination effectiveness. As
summarized in Table 15, the second model demonstrated partial support for Hypothesis 1
as the aggregated social capital measure was significantly and positively related to
coordination effectiveness. Based on differences in the variance explained, findings
indicated that social capital explains 23% additional variance in member perceptions of
coordination effectiveness in comparison to the covariate only model. The chi square
difference in the log likelihood ratio based on one degree of freedom difference indicated
a significant improvement in model fit (chi square difference = 8.23; 1 df; p>.005).

Taken as an aggregated scale, social capital had an effect size of 1.06, which according to

87

the standards proposed by Cohen (1988), indicates a strong effect on the outcome

variable of coordination effectiveness.

Table 15: Relationship of Social Capital to Coordination Effectiveness

 

 

 

 

 

 

 

 

 

 

 

 

Model 1 Model 2
Estimate Estimate P-value Effect size
(standard error) (standard error)
Intercept 4.44 (.06) 4. 44(. 06) 0.000
Social
Capital .66(.22) 0.004 1 .06
Familiarity .1 7(. 85) -. 52(. 82) 0.52 7
Age .02(.01) .02(.01) 0.116
Size .466 33) .59; 30) 0.059
Vilma?“ 13.7% 36.8%
explamed
Log
Likelihood
Difference 8.23 (1)**
(change in
DF)

 

 

 

The Relationship of Social Capital to DVCC System Change Eflectiveness

Systems change effectiveness was modeled in the exact same manner as

coordination effectiveness. As shown in Table 16, results found that social capital was
significantly and positively related to DVCC systems change effectiveness. Based on
differences in the variance explained, findings indicate that social capital explains 33%
additional variance in member perceptions of coordination effectiveness in comparison to
the covariate only model. The chi square difference in the log likelihood ratio based on
one degree of freedom difference indicated a significant improvement in model fit (chi

square difference = 14.88; 1 df; p<.000). Per Cohen’s criteria, social capital had a very

strong effect (1.3) on systems change outcomes.

88

 

Table 16: Relationship of Social Capital to Systems Change Effectiveness

 

 

 

 

 

 

 

 

 

 

 

 

Model 1 Model 2
Estimate Estimate P-value Effect size
(standard error) (standard error)
Intercept 3.83(.08) 3.83(.O7)*** 0.000
Social
Capital . l.08(.26)*** 0.000 1.3
Familiarity
-.37(.1.1) -1.52(.96) 0.121
Age .03(.02) 03(01) 0.084
5'“ .6I(.42) .82(.36)* 0.027
vanal‘ce 12.8% 46.3%
explained
Log
Likelihood
Difference 1488(1) mu.
(change in .
DF)

 

 

 

Exploring Interactions Between Composition Diversity and Social Capital

The next phase of analysis was conducted to test Hypothesis 3 which proposed
that the relationship between social capital and DVCC effectiveness would be stronger
under the context of greater diversity. However, even though there was a theoretically
specified directionality to this hypothesis, the diversity type was treated as an exploratory
variable. As such, five different types of DVCC composition diversity were examined.
The first step in this analysis was to examine the correlations between each type of
diversity and the outcome variable to see whether diversity types showed patterned
relationships both with each other and with the outcome variables to justify aggregation
into an overall diversity score. As shown in Table17, the results of this analysis showed

that different forms of diversity varied in the direction of their relationship with the

89

 

outcome variables and with social capital. More specifically, gender diversity was
positively and significantly correlated with coordination outcomes. Tenure diversity and
sector diversity had nonsignificant but negative correlations to coordination effectiveness.
Breadth of stakeholders and management diversity showed very little correlation. For
systems change outcomes, none of the indicators of diversity showed significant
correlation, however, tenure diversity showed a negative directionality while the other
forms had a positive directionality.

Table 17: Correlations of Diversity Indicators with Social Capital and Outcomes
l 2 3 4 5 6 7

 

 

1. Coordination effectiveness

 

2. Systems change effectiveness 0.68”

 

 

 

 

 

 

3. Social capital 0.30* 0.35"

4. Gender Diversity 0.37" 0.17 0.16

5. Tenure Diversity -0.20 -0.15 -0.17 019

I6. Sector Diversity -0.11* 0.05 -0.05 0.02 0.29*

7. Management Diversity 0.08 0.17 0.20 -0.01 0.08 0.42"

8. Breadth of Stakeholders 0.03 0.20 -0.36** 0.15 0.02 0.49“ 0.14

 

 

 

In light of this, each type of diversity was examined separately in its interaction
with social capital in relation to both coordination and systems change effectiveness.
Specifically, an interaction term was created for each diversity index (e. g., tenure, gender,
sector); calculated as the product of the diversity index score and the overall social capital
scale score. Following the procedures for testing interactions proposed by Aiken & West

(1991) this interaction term was then modeled in an HLM equation along with the main

90

effects to test whether the interaction was a significant predictor of either coordination or
systems change effectiveness.

The above analysis approach was justified in light of the exploratory nature of
the investigation and the evidence that different forms of diversity related to the
dependent variables in different ways. However, the limitation to this analysis is that it
inflates the likelihood of Type 1 errors resulting from multiple analyses. While there
are corrections for multiple analyses such as the Bonferonni correction, these approaches
have been highly criticized by statisticians due in part to their corresponding inflation of
Type 2 errors. In light of this, this study chose to follow the recommendations put forth
in the literature (e.g., Pemeger, 1998) to simply treat these analyses as exploratory. All
interaction models included the controls of DVCC age, size, and. familiarity.

Gender Diversity

As shown in Table 18, for coordination effectiveness, main effect analysis found
positive significant effects for both gender diversity and social capital. This indicates that
DVCCs with more balanced proportions of men and women and higher social capital
were viewed by their members as more effective. However, when the interaction term
was entered into the model, findings indicated that there was no significant interaction
between gender diversity and social capital in relation to either coordination or systems
change effectiveness. This means that there is no evidence to suggest that social capital
is more or less important to coordination outcomes when there is greater diversity among

members in terms of the proportional representation of men and women.

91

Table 18: Analysis of Gender Diversity Interaction for Coordination Effectiveness

 

 

 

 

 

 

 

 

 

 

Coordination Model 1: Main Effects Model 2: Interaction Model
Estimate P-value Estimate P—value
(Standard (Standard
error) error)
Intercept 4.44(.05)*** 0.000 4.45(.05)*** 0.000
Gender d‘verS‘W .52(.24)* 0.032 -.5(4.26) 0.907
30°13] cap‘tal .57(.22)** 0.011 .39 (.73) 0.595
Gender * Social
capital .23(.98) 0.816
Size .46(. 26) 0.086 .49(.27) 0.075
F“’"”'“"’y -. 59(. 78) 0.455 -4 7c 79) 0.560
Age .027 01) 0.167 .01701) 0.387

 

 

 

 

 

 

For systems change effectiveness, the main effect analysis found no significant
effect for gender diversity. In other words, while greater gender diversity was positively
related to coordination effectiveness, it appears to have no relationship to whether
members perceived their DVCC to be effective at promoting systems change. When the
interaction term was entered into the model, findings further indicated that there was no
Significant interaction between gender diversity and social capital in relation to systems
Change effectiveness. This means that there is no evidence to suggest that social capital
is more or less important to systems change outcomes when there is greater diversity

among members in terms of the proportional representation of men and women.

92

Table 19: Analysis of Gender Diversity Interaction for Systems Change

Effectiveness

 

Systems change

Model 1: Main Effects

Model 2: Interaction Model

 

 

 

 

 

 

 

 

 

Estimate P-value Estimate P-value

(Standard (Standard

errog error)
Intercept 3,33(,o7)*** 0.000 3.74(.06)*** 0.000
Gender dwe’S‘W .06(.29) 0.849 7.74(5.09) 0.134
50°13“ “PM 1.07(.26)*** 0.000 2.33(.88)** 0.010

,, . .

Gender Social capital _1 .730 .17) 0144
Age .77(. 32) 0.023 .71(.32)* 0.032
5’“ -1574 96) 0.110 -I.43(.96) 0.144
Familiar"? .037. 01) 0.086 .02(. 01) 0-120

 

 

 

 

 

Tenure Diversity

As shown in Table 20, there was no evidence to suggest that tenure diversity was

in any way related to either coordination or systems change effectiveness. Tenure

diversity showed no significant main effect and findings indicated that there was no

significant interaction between tenure diversity and social capital in relation to either

coordination or systems change effectiveness. This means that there is no evidence to

suggest that social capital is more or less important to either coordination or systems

change outcomes when there is greater diversity among members in terms of the number

of years they had been involved with the DVCC.

93

 

Table 20: Analysis of Tenure Diversity Interaction for Coordination Effectiveness

 

 

 

 

 

 

 

 

 

 

Coordination Model 1: Main Effects Model 2: Interaction Model
Estimate P-value Estimate P-value
(Standard (Standard
error) error)
Intercept
4.44(.06)*** 0.000 4.44(.06)*** 0.000
T d' '
em“ “’6th -.18(.36) 0.609 -.93(7.04) 0.895
Social capital
.61(.24)* 0.013 .53(.76) 0.492
Tenure * Social
capital .l6(l.62) 0.920
Size
.54633) 0.107 .51(.34) 0.134
F .1. .
0"" ”my -56(. 86) 0.521 -.42(86( 0.630
Age .
.02601) 0.118 .01(.01) 0.279

 

 

 

 

 

Table 21: Analysis of Tenure Diversity Interaction for Systems Change

Effectiveness

 

Systems change

Model 1: Main Effects

Model 2: Interaction Model

 

 

 

 

 

 

 

 

 

Estimate P-value Estimate P-value

(Standard (Standard

error) error)
Intercept 3..84(07)*** 0000 3,,78(O7)*** 0.000
Tenure drversrty _.13(_14) 0.747 _9.33 (3.15) 0,231
Soc1al capital 1.0427)..." 0.000 .09(.88) 0.920
Tenure * Social
capital 2.22(1.87) 0242
Age

.85638) * 0.031 .6439) 0.104

5'28 -1.49(1.01) 0.148 -1.66(1.01) 0-108
Fam’l‘m’y .03(.02) 0.084 .03(.02) 0-052

 

 

 

 

 

94

 

 

Management Diversity

As shown in Tables 22 and 23, there was also no evidence found to suggest that
management diversity was in any way related to either coordination or systems change
effectiveness. Management diversity showed no significant main effect and findings
indicated that there was no significant interaction between management diversity and
social capital in relation to either coordination or systems change effectiveness. This
means that there was no evidence to suggest that social capital is more or less important
to either coordination or systems change outcomes when there is greater diversity among
members in terms of their level of management within their organization or agency.

Table 22: Analysis of Management Diversity Interaction for Coordination
Effectiveness

 

 

 

 

 

 

 

 

 

 

Coordination Model 1: Main Effects Model2: Interaction Model
Estimate P-value Estimate P-value
(Standard (Standard
error) error)
Intercept 444005?" 0.000 4,46(,05)*** 0.000
Management drverSIty -.76(.45) 0.097 _1 3.4092) 0,096
Socral capital .76(.23)*** 0.001 1.87(.73)* 0.013
Management * Social
capital 2.86(1.81) 0.120
3'29 .61(.28)* 0.034 .69(.27)* 0-015
Fam'l’m’y -.59(.79) 0.459 -.33(.77) 0667
Age .02(.01) 0.051 .02(.01) 0.107

 

 

 

 

 

 

95

Table 23: Analysis of Management Diversity Interaction for Systems Change

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Effectiveness
Systems Change Model 1: Main Effects Model 2: Interaction Model
Estimate P-value Estimate P-value
(Standard (Standard
error) error)
Intercept 3.83(.07)*** 0.000 3.81(.O7)*** 0.000
Management diversrty _.1 6(.54) 0.763 _9.01(9.79) 0,362
Socral capital 1.1(.27)*** 0000 193(39):: 0.036
Management * Social
capital 2.(2.24) 0.375
5'29 .8(. 34) 0.023 .8(. 34) 002"
Famtllarity -1.57(. 96) 0.109 -1.55(.97) 0.117
Age .03(.01) 0.074 .03(.01) 0.087
Sector Diversity

_ As shown in Table 24, main effect analysis indicated that both social capital and
sector diversity are significant predictors of coordination effectiveness. However, while
social capital was positively related to coordination outcomes, greater diversity of sectors
showed a negative relationship. In other words, the more equal DVCCs are in the
proportions of non-profit, for-profit, and public agencies, the less effective they were
perceived to be at improving coordination. This significant main effect was somewhat
surprising as sector diversity was not significantly correlated with coordination
effectiveness. This would suggest that one or more of the other variables in the model
(e. g., social capital, age, size or familiarity) may act as a suppressor to the relationship

between sector diversity and coordination outcomes. For example, one common form of

96

suppression that could explain this effect is that one of the control variables or social

capital is accounting for error variance associated with the relationship between sector

diversity and coordination effectiveness, resulting in a lower p-value.

However, results did not find a significant interaction between sector diversity

and social capital in relation to coordination effectiveness. This indicates that social

capital is not more or less important to coordination when there is greater diversity

among members in terms of institutional sector.

Table 24: Anaysis of Sector Diversity Interaction for Coordination Effectiveness

 

 

 

 

 

 

 

 

 

 

Coordination Model 1: Main Effects Model 2: Interaction Model
Estimate P-value Estimate P-value
(Standard (Standard
error) error)
Intercept 444(35): =1: 4: 0_000 4,45(,05)* * * 0.000
Sector diversrty '1-17(-53)* 0.031 -l.61(7.82) 0,838
30°13“ “PM .67(.21)** 0.003 .63(.77) 0.418
* . .
Sector Social capital .07033) 0,971
5'29 .57(. 26)* 0.038 .6026)* 0.029
Familiarity _1. 06(. 81) 0.198 -. 97( 81) 0.233
Age .02(. 01)* 0.043 .02(. 01) 0.097

 

 

 

 

 

As shown in Table 25, there was no evidence found to suggest that sector

diversity was in any way related to systems change effectiveness. Sector diversity

showed no significant main effect and findings indicated that there was no significant

interaction between sector diversity and social capital in relation to systems change

97

 

effectiveness. This indicates that social capital is not more or less important to systems
change when there is greater diversity among members in terms of institutional sector.

Table 25: Analysis of Sector Diversity Interaction for Systems ChangefiEffectiveness

 

 

 

 

 

 

 

 

 

 

Systems Change Model 1: Main Effects Model 2: Interaction Model
Estimate P-value Estimate P-value
(Standard (Standard
error) error)
Intercept 3.83(.06)*** 0.000 3,8(,O6)*** 0.000
Sector d1versrty -.81 (.63) 0205 -5.08(9.51) 0,595
Socral capital 1.090255)..." 0.000 .75(.94) 0,431
* . .
Sector Social capital .96(2.22) 0.667
Size 492632)" 0,014 .79(.32)* 0.019
Familiarity ,1_ 92‘ 98) 0. 05 7 -2.02(99) * 0.047
A88 .03(.01) * 0.050 .03(.01) 0.058

 

 

 

 

 

Breadth of Stakeholders

As shown in Table 26 and 27, there was no evidence found to suggest that the
breadth of stakeholders was in any way related to either coordination or systems change
effectiveness. Stakeholder breadth showed no significant main effect and findings
indicated that there was no significant interaction between management diversity and
social capital in relation to either coordination or systems change effectiveness. This
means that there is no evidence to suggest that social capital is more or less important to
either coordination or systems change outcomes when there is greater diversity among
members in terms of their level of management within their organization or agency.

However, it is interesting to note that, while not significant, the direction of the

98

 

relationship between breadth of stakeholders and coordination effectiveness was

negative.

Table 26: Analysis of Breadth of Stakeholders for Coordination Effectiveness

 

 

 

 

 

 

 

 

 

 

Coordination Model 1: Main Effects Model 2: Interaction Model
Estimate P-value Estimate P-value
(Standard (Standard
error) 91'1'01')
Intercept 4.44605)": 0000 4.45(.05)*** 0.000
Breadth of Stakeholders _.05(.03) 0.16 .12(.36) 0.743
50012“ “PM .59(.22)** 0.010 .84(.64) 0-191
Breadth * Social -.04(.03) 0,637
capital
5'29 .77(.35)* 0.034 .86(.33)* 09”
Familiar“? -.97(.87) 0.269 -.9(.82) 0274
Age .02(.01) 0.091 .02(.01) 0.182

 

 

 

 

 

Table 27 : Analysis of Breadth of Stakeholders for Systems Change Effectiveness

 

Systems Change

Model 1: Main Effects

Model 2: Interaction Model

 

 

 

 

 

 

 

 

 

Estimate P-value Estimate P-value

(Standard (Standard

error) error)
Intercept 3.83 (.07).... 0000 331(07):" 0.000
Breadth of Stakeholders .01(.04) 0.868 .2044) 0.660
Social capital 1.09(.26)*** (1000 139(77) 0.078
Breadth * Social
capital -.04(.10) 0.664
Size .73642) 0.087 .64(.41) 0.124
Fam'hm’y -1.5(1.04) 0.155 4.361101) 0182
Age .03( 01) 0.080 .02(. 01) 010"

 

 

 

 

 

99

 

 

Exploring Relative Importance Of Different Indicators Of Social Capital

In light of the findings that social capital, operationalized as an aggregate of five
qualities of relationships, was significantly and positively related to perceived DVCC
effectiveness at both improving coordination and promoting broader systems change, the
next important question to investigate was whether certain qualities of relationships are
more important than others. In other words, is it simply that DVCC effectiveness
requires stronger relationships in general, or are certain types of relationships more
important than others in explaining differences in DVCC effectiveness? In order to
explore this question, two stages of analysis were conducted. First, each indicator of
social capital (e. g., trust in follow through, recognized expertise, shared philosophy) was
modeled separately to examine whether, individually, each explained a significant
amount of variance in the outcome. In the second stage, all significant predictors were
entered together to examine their collective relationship to each outcome. In this
combined model, effect sizes were calculated for each indicator to examine its relative
effect. Stage one allowed for better specification of the model for a given outcome while
stage two allowed for examining whether any one type of social capital accounted for
unique variance over and above the variance shared across the different indicators.

Because the directionality of the relationship between social capital and
effectiveness was empirically established in the earlier analysis, one-tailed significance
tests were used to maximize statistical power in the combined models. However, due to
the heightened risk of Type I errors present from repeated tests, the individual analyses

utilized the more conservative two-tailed significance tests.

100

Relative Importance of Social Capital Indicators for Coordination Eflectiveness
When each indicator was modeled independently to coordination effectiveness,
results showed that communication frequency, responsiveness to concerns, recognized
expertise, and trust in follow through were all positively and significantly related to
coordination effectiveness. Shared philosophy appeared to have the weakest relationship
to coordination effectiveness, resulting in only a trend toward significance.

Table 28: Independently Modeled Indicators of Social Capital Predicting
Coordination Effectiveness

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Estimate (std. error) (fin

Model 1 Communication Frequency ,33(,14)* 0,026
Familiarity -85 (, 92) 0. 363

Age .02(. 01) 0.186

Size .57(.3I)’ 0.079

Model 2 Shared Philosophy ,33(,19)A 0,053
Familiarity _ 21(82) 0.794

Age .02( 01) 0.216

Size ,61(,32) 0.067

Model 3 Responsiveness to Concerns ,41(, 13)* 0,028
Familiarity -.16(83) 0.848

Age .02( 01 ) 0.124

Size .55(.31) 0.090

Model 4 Recognized Expertise “12(10):: 0,042
Familiarity -1 6 ( 83) 0.852

Age .02(. 01) 0.072

Size ,4(,31) 0.210

Model 5 Trust in Follow Through ,49(,19)* 0.013
: Familiarity - 15¢ 80) 0.847
‘ Age .02(. 01) 0.138
L Size .58(. 3) 0.066

 

101

 

In the second stage analysis, all indicators were entered into a combined model
predicting coordination effectiveness (See Table 29). In this combined model, results
indicated that no one indicator of social capital accounted for a significant amount of
unique variance over and above the variance that was shared by all four indicators. One
interpretation of this finding is that, for coordination outcomes, what is important is
having generally strong relationships characterized by frequent communication,
responsiveness to concerns, trust in follow through, and recognized expertise. It may be
that no one type of relationship is particularly important over and above the others.
However, in examining the effect sizes, frequent communication and trust in follow
through both had moderate effects on coordination outcomes relative to the weak effects
associated with shared philosophy, responsiveness to concerns, and recognized expertise.
This provides weak evidence that communication frequency and trust in follow through
may be particularly important types of relationship qualities for coordination
effectiveness. As these differences were not statically significant, however, they should

be interpreted with due caution.

102

Table 29: Combined Model Predicting Coordination Effectiveness

 

Estimate P-Value Effect Size

(std. error) gone-tailed)

 

 

 

 

 

 

 

 

 

 

Intercept 4.44(.05) 0.000

Communication

F .22(.16) 0.089 0.66
requency

Shared Philosophy .03(.27) 0.460 0.02

Responsiveness to

C .00(.29) 0.497 0.05
oncerns

Recognized Expertise .11(.26) 0,340 0.21

Trust in Follow

Through .32(.27) 0.1 19 0.53

Familiarity -. 8C 91) 0.191

Age .024 01) 0.069

Size ,6(_31) * 0.033

 

 

 

 

 

Relative Importance of Social Capital Indicators for Systems Change Efiectiveness
When each indicator was modeled independently in its relationship to systems
change effectiveness, results showed that all five indicators of social capital were
positively and significantly related to differences between DVCCs in members’

perceptions of effectiveness.

103

Table 30: Independently Modeled Indicators of Social Capital Predicting Systems

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Change Effectiveness
W W
, . , _ error) (two-tailed)
Model 1 Communication Frequency ,5(,17)** 0,005
Familiarity -1.96(1.14) 0. 091
Age .02( 02) 0.1 75
Size .7 7(. 39) 0.056
Model 2
Shared Philosophy 34(21):.“ 0000
F amiliarity . 3 (. 93) 0. 745
Age .02( 01) 0.201
Size .92(. 36) * 0.01 7
Model 3 Responsiveness to Concerns 57(12):” 0,003
Familiarity -_ 9( 99) 0. 3 71
Age .037. 02) 0.087
Size .766 38) 0. 054
Model 4 Recognized Expertise ,55(,25)* 0,032
Familiarity -.82(1.05) 0.439
Age . 036 02) 0. 05 4
Size .54(.40) 0.180
Model 5 Trust in Follow Through ,75(,23)** 0,002
Familiarity .88( 99) 0.382
Age .02(. 02) 0.115
Size .79(. 38) * 0.046

 

 

 

 

When all indicators of social capital were modeled collectively, two indicators

distinguished themselves as accounting for unique variance above the others. Both

shared philosophy and communication frequency were significantly and positively

related to DVCC systems change. In examining the effect sizes, results showed that both

104

 

shared philosophy and communication frequency had moderately strong effects on
systems change effectiveness. The unique effect of trust in follow through was
moderately weak and non-significant. Responsiveness to concerns appeared to have no
unique effect once the variance it shared with the other five indicators was partialed out.
Interestingly, effect sizes indicated that recognized expertise had a non-significant but
negative relationship to systems change effectiveness once the shared variance was
partialed out.

Table 31: Combined Model Predicting Systems Change Effectiveness

 

 

 

 

 

 

 

 

 

 

Meats. H2141: W
W M
Communication
F .32(.18)* 0.044 0.71
requency
Shared Philosorrhy .58(.3)* 0.031 0.81
Responsiveness to
C .01(.33) 0.488 0.01
oncems
Recognized Expertise -.08(.3) 0.399 0.11
Trust in Follow Through .25(.31) 0.214 0.34
Familiarity 14571.05) 0.088
Age .02(. 01) 0.120
Size .991. 36) 0.005

 

 

 

 

Results from this analysis suggest several things. First, findings indicated that
stakeholder relationships characterized by frequent communication outside of DVCC
meetings and a high degree of trust in each other’s willingness and ability to follow
through on what they say they will do appeared to be generally important across both

indicators of effectiveness. Second, it was particularly interesting that shared philosophy

105

 

had the strongest effect on systems change outcomes given that it showed only a trend
toward significance in its relationship to coordination effectiveness when entered alone as
the sole independent variable. This suggests two things. First, it indicates that different
types of outcomes may require different types of relational capacities in order to be
effective. Second, it indicates that members’ viewing one another as sharing a similar
type of philosophy concerning what domestic violence is and how it should be addressed

may be particularly vital for systems change types of outcomes.

106

DISCUSSION
There is ample theory and conceptual arguments present in the literature to
suggest that relationships among stakeholders may constitute a form a social capital that
can facilitate the functioning and effectiveness of IAs such as DVCCs (Bond & Keys,
2000; Campbell et al., 1999; Gray, 1996; Mulroy, 1997; Tapper et al., 1997). However,
despite this, there has been very limited work done focused on systematically
operationalizing and empirically examining the role stakeholder relationships play in IA
effectiveness. This study has sought to address this gap by using key informant
interviews and current literature to operationalize social capital within the context of one
prominent form of IA - DVCCs - and empirically examine the relationship of social
capital to indicators of DVCC effectiveness. In doing so, this study has contributed to
the current literature by transforming implicit assumptions about the importance of
relationships in collaborative contexts into explicit empirical questions concerning what
types of relationships matter, for what types of outcomes relationships matter, and under
what conditions relationships matter.
The Role of Social Capital in DVCCs’ Effectiveness at
Promoting Community Change
Key informant interviews supported a conceptualization of social capital as a
multi-dimensional construct by suggesting that there are several different kinds of
relationships that are important for DVCC functioning and effectiveness. This notion
has been supported in the current literature. For example, social capital has been
conceptualized as embedded within a diverse array of ties including ties based on

friendship (Krackhardt, 1992), collegiality (Frank & Zhao, 2000), communication

107

frequency (Reagans & Zuckerman, 2001), formal collaboration (Foster F ishman et al.,
2001; McDonald, 2003), and trust (Ahuja, 2002). Types of relationships indicative of
social capital highlighted by DVCC informants in this study included relationships
characterized by (I) frequent communication outside of DVCC meetings, (2) a shared
philosophy concerning what domestic violence is and how it should be addressed, (3) a
strong recognition that each stakeholder offers a unique perspective or form of expertise
valuable to addressing domestic violence, (4) a high level of responsiveness to issues or
concerns brought to the attention of one stakeholder by another, and lastly, (5) a high
degree of trust in stakeholders’ willingness to follow through on what they say they will
do. While key informants were asked specifically about what relationships need to look
like in order for DVCCs to be effective, it is noteworthy that each of the identified
relationships have linkages to the collaboration literature as generally important
indicators of social capital in collaborative contexts (e. g., Gray 1989; Foster-Fishman et
al, 2001; Mattessich, Murray-Close, & Monsey, 2001).

Given the general consensus that social capital can be embedded in various types
of relationships, it stands to reason that each of these forms of social capital would have
an additive if not synergistic relationship to the others such that collectively they
represent a more comprehensive inventory of the social capital present within a DVCC.
Modeled together as an aggregated scale of social capital, this study found support for the
hypothesis that social capital is a significant and positive predictor of DVCC
effectiveness at promoting commrmity change. Members perceived their DVCC to be
more effective at both improving coordination and promoting broader systems change

when DVCCs were characterized by more and stronger relationships among stakeholders.

108

While identifying the specific mechanism for this influence was beyond the scope
of the cm'rent study and represents an important area for fiiture research, current literature
does provide several possible mechanisms. For example, Ahuja (2000), Cooke & Wills
(1999), and Shan and colleagues (1994) all found that having strong ties with other
organizations can promote organizational innovation. While their level of analysis was at
the level of the individual organization, it is possible that such effects could be multi-
level in nature. It could be that DVCCs characterized by dense networks provide both
the norms of cooperation and the base of resources necessary for DVCCs to find
innovative pathways for overcoming obstacles and therefore would be evaluated by
members as more effective at promoting community change. It could also be that,
consistent with the findings of Reagans & Zuckerman (2001), stronger relationships
reduce barriers to cooperation and therefore result in higher levels of productivity.

Another important question that has been raised in the literature on collaboration
is whether certain collaborative outcomes require IAs to have different capacities both in
form and in strength. For example, Himmelman (2001) and others (Austin, 2000; Gajda,
2003; Hogue, 1993) have argued that cooperative interactions fall on a continuum
ranging from networking types of interactions to more integrated collaborations. These
scholars have further argued that different types of capacities are required as pre-
conditions for effectiveness at different levels along the continuum. For example,
Himmelman (2001) posits that networking interactions have limited capacity
requirements beyond some investment of time and energy on the part of participants.
However, as one moves up the continuum toward collaboration and collective action,

successful interactions require greater capacity with regards to the level of trust and

109

willingness to share turf. This poses an important question as to whether social capital is
more important for certain outcomes than others.

This study examined two outcomes related to DVCC effectiveness at promoting
community change; namely DVCC effectiveness at improving coordination among
community organizations and agencies and DVCC effectiveness at promoting broader
systems change. Coordination effectiveness, as it was operationalized in this study, can
be thought of as how effective the DVCC has been at improving how things work within
the existing community response system for domestic violence. For example,
participants were asked about the extent to which they felt the DVCC had been effective
at increasing the ability of organizations and agencies to coordinate their efforts and the
extent to which the efforts of the DVCC had resulted in organizations and agencies
working together more efficiently.

On the other hand, systems change effectiveness, as it was operationalized in this
study, can be thought of as the effectiveness of the DVCC at making changes to the
infrastructure of the community system. For example, participants were asked about their
DVCC’s effectiveness at developing needed programs and services, changing policies,
and impacting public attitudes about domestic violence. According to the continuums
proposed by Himmelman (2001), Hogue (1993), and Gajda (2003), coordination types of
outcomes tend to require only a moderate degree of relational capacity. However,
systems change types of outcomes frequently require a high degree of collective action
among stakeholders (Kelly, Ryan, Altman, & Stelzner, 2001). Such collective action
would place systems change outcomes higher on the continua referenced above and

therefore would be require greater relational capacity. In some ways, this is consistent

110

with the conclusions reached by Krackhardt (1992) who, in a case study investigation of
the influence of social networks during an organizational unionization attempt, concluded
that major changes that may threaten the status quo concerning power and decision
making require strong affective relationships. This would suggest that social capital
would have a stronger effect on systems change outcomes relative to coordination
outcomes.

Accordingly, this study found that while social capital was significantly and
positively related to both coordination and systems change effectiveness, it appeared to
have a stronger effect on systems change outcomes based on differences in both effect
sizes and the variance explained. While there were methodological limitations in directly
testing whether these differences in effect sizes were statistically significant, this finding
does provide some empirical evidence in support of the capacity continuum theories
proposed by Himmelman and others. Empirical support for these theories has important
implications for both research and practice as it highlights the developmental nature of IA
capacity and suggests that the level of capacity that is adequate for certain IA outcomes
may not translate into a general capacity for all outcomes. Specifically, this study
provides evidence that DVCCs may need more and stronger relationships in order to
effectively engage in systems change types of activities. This is a particularly important
implication as IAs in general, and DVCCs in particularly, are increasingly being called
upon by outside funders to adopt more systems change oriented goals (e. g.
http://www.cdc.gov/ncipc/DELTA/default.htm; March 2006). Study findings suggest
that strengthening quality and connectedness of relationships may be an important lever

for helping DVCCs to effectively make such a transition.

11]

The Relationship of Social Capital to DVCC Effectiveness at
Strengthening Member Organizational Capacity

A second way of thinking about DVCC effectiveness is to consider the extent to
which the DVCC has been effective at strengthening the organizational capacity of its
members. However, an unexpected yet interesting finding barred exploration of the
relationship of DVCC social capital to this set of outcomes. Imbedded within this
investigation was the assumption that DVCCs would significantly differ from one
another in their effectiveness at increasing member capacity. Contrary to this
assumption, analyses revealed that the DVCC to which a participant was affiliated was
not a significant factor in predicting the extent to which they felt they and their
organization had been positively impacted by the DVCC. In other words, while
participants differed significantly from one another in the extent to which they felt their
organization benefited as a result of their involvement with the DVCC, these differences
appear to be largely a function of individual or organizational level characteristics that
were not shared uniformly by all the members of the DVCC.

This finding leaves important unanswered questions for future research
concerning what types of members are being most strongly impacted by their
involvement in the DVCC and what factors might explain this impact. For example, are
certain stakeholder groups or organizational sectors being impacted more than others?
Are organizations who are represented by higher level managers or executives being
impacted differently than those who are represented on the DVCC by middle managers or

line staff?

112

Another promising direction for future research would be to build on the work of
many social capital scholars (Brass & Burkhardt, 1992; Talmud, 1999; Tsai, 2000) in
examining the social capital of individual DVCC members as a predictor of
organizational impact. This study focused on network density as an operationalization of
social capital at the level of the DVCC. However, one of the limitations of this approach
is that it does not allow the researcher to examine differences among stakeholders in their
individual level of social capital. Indeed, previous research in interorganizational
contexts has found that an organization’s social networks are powerful predictors of
positive organizational outcomes relevant to IAs including capacity for innovation
(Ahuja, 2000); organizational learning (Tsai, 2000), and influence (Brass & Burkhardt;
1992)

One other promising direction for future research related to member impacts
would be to build off of the work of previous scholars in examining characteristics of
DVCC members that could be related to issues of “dosage”. Considerations of
intervention dosage typically concern the relationship between the concentration and
duration of the intervention to the strength of the impact or effect (Cook & Shadish,
1986). The basic idea to this approach is to propose that if DVCCs are effective
mechanisms for increasing organizational capacity, then constructs related to intervention
concentration and duration such as the level of the activity of organizational
representatives, the number of representatives an organization sends to DVCC meetings,
and the length of time in which a member organization has been involved should be
positively related to the extent to which that the DVCC has positively impacted the

Organization. The link between the level of engagement an organization has with an IA

113

and organizational outcomes has received some support in the current literature. For
example, F oster-F ishman and colleagues (2001) found that organizations who had been
involved with human service IAs longer and had more staff members attending meetings
were more likely to be involved in resource exchanges and joint ventures. This provides
evidence that greater involvement in an IA can lead to organizations and agencies
becoming more involved in collaborative efforts. Given this, an interesting direction for
future research would be to expand upon this research to examine how dosage may relate
to the extent to which and ways in IA members feel their organization has been
strengthen as a result of their involvement with an IA.
The Interaction of Social Capital with Diversity

Structural holes theory (Burt, 2000) argues that relationships represent the
greatest form of social capital when they bridge between diverse actors. The premise
underlying structural holes theory is that networks characterized by ties between more
diverse actors are most likely to have access to the widest array of resources including
financial and tangible assets, knowledge, and skills. This would suggest that social
capital would have the strongest relationship to DVCC effectiveness under conditions of
greater diversity.

This study empirically examined this hypothesized interaction in its relationship
to both coordination and systems change effectiveness for a variety of indicators of
DVCC diversity. Specifically, this study investigated the interaction of social capital
with DVCC composition diversity concerning (1) gender, (2) tenure, (3) management
level, (4) organizational sector, and (5) breadth of stakeholder groups represented. Main

effect analyses indicated that both gender diversity and sector diversity significantly

114

impacted the extent to which a DVCC was perceived by its members to have been
effective at improving coordination among community organizations and agencies.
However, while more balanced proportions of men and women appear to enhance
coordination effectiveness, greater diversity among members in terms of their sector
affiliation (for profit, nonprofit, private) appeared to negatively impact coordination
outcomes. Further, while not significant, all forms of diversity with the exception of
gender appear to have a negative directionality in their relationship to both coordination
and systems change outcomes after social capital, age, and size had been accounted for.
This is interesting as it contradicts what one might expect based on much of the
literature on work group diversity which suggests that task relevant types of diversity
(e. g., tenure, sector, management level) are more likely to be positively related to
effectiveness than forms of diversity more tied to social categorizations such as gender
(e.g., Jackson, May, & Whitney; 1995; Jackson & Joshi, 2004; Tsui, Egan, & O’Reilly,
1992, Williams & O’Reilly 1999). One possible explanation for this is the dynamics of
gender diversity within the specific context of DVCCs. Because the issue of domestic
violence has strong links to issues related to gender norms, it may be that a combination
of both male and female perspectives represents a highly task relevant form of diversity
than can improve DVCC effectiveness. However, by itself, this explanation does not
account for why gender diversity was only significantly related to coordination
effectiveness and not systems change. Another possible explanation relates to a potential
response bias related to gender. Specifically, do men generally have lower expectations
for DVCC effectiveness than women and therefore are DVCCs with more men rated as

more effective as a result of this bias? However, one-way AN OVAs found no significant

115

gender differences in participant ratings of either coordination or systems change
effectiveness making this an unlikely explanation.

A final potential explanation is that gender diversity in this study is positively
related to DVCC legitimacy and influence within more male-dominated institutions.
Several of the key stakeholder groups represented on DVCCs (e. g., courts, law
enforcement agencies, prosecuting attorneys) are traditionally comprised of majority men
in both their staff and leadership. For example, in 2001 the National Center for Women
and Policing reported that women accounted for only 12.7% of all sworn law
enforcement positions in large agencies and only 8.1% of smaller more rural agencies in
the United States (Lonsway et al., 2002). This is significant as studies have documented
that women often experience particular challenges gaining influence and obtaining
positions of authority within male-dominated institutions (Floge & Merrill, 1985; Martin,
1980; Yoder, 1991). In this sample, greater gender diversity was synonymous with more
men involved with the DVCC as there were only two instances of men comprising more
than 50% of the DVCCs membership and these had roughly equal proportions of men
and women. As such, the positive impact of gender diversity on coordination
effectiveness may result, not so much from group dynamics related to diversity itself, but
rather from the effect having more male members may have on a DVCC’s capacity to
leverage greater cooperation from male-dominated institutions critical to a coordinated
community response to domestic violence.

Findings from this study did not support diversity as an important variable in
explaining the conditions under which social capital impacts DVCC effectiveness.

Contrary to expectations, results found no evidence of any significant interaction effects

116

concerning the impact of social capital over different levels of composition diversity.
One possible explanation is that differences were unable to be detected due to a range
restriction present in the data. Management, tenure, and sector diversity all had some
degree of a negative skew indicating that, in general, DVCCs tended to be fairly diverse.
This explanation does not, however, explain the lack of a significant interaction for
breadth of stakeholders.

Study results do, however, highlight important areas in need of future research
concerning DVCC composition diversity. Findings indicate that (1) diversity impacts
coordination outcomes more than it impacts systems change outcomes, and (2) different
types of diversity impact coordination effectiveness in different ways. This poses
important questions for future research. For example, what is it about coordination
effectiveness that makes it more strongly influenced by diversity? Second, under what
conditions does diversity have the greatest influence and what is the mechanism through
which diversity impacts effectiveness? And lastly, why does gender diversity appear to
enhance coordination outcomes when all other forms of diversity appear to diminish it?

Relative Importance of Different Indicators of Social Capital

As discussed, a principal finding of this study is that social capital,
conceptualized as the combined network density across five different indicators of social
capital (communication frequency, shared philosophy, recognized expertise,
responsiveness to concerns, and trust in follow through), is positively and significantly
related to DVCC effectiveness. In light of this, an important next step in exploring the
relationship of social capital to DVCC effectiveness was to examine whether certain

types of relationships are more important than others depending on the outcome of

117

interest. Specifically, this study examined the relative importance of the five different
indicators of social capital as they related to both coordination and systems change
effectiveness.

For coordination effectiveness, findings indicated that frequent communication,
recognized expertise, responsiveness to concerns and trust in follow through were all
significantly and positively related to effectiveness when modeled separately. However,
when modeled together, no one indicator distinguished itself as a significant predictor
over and above the shared variance that exists among all them. This could indicate that
for coordination outcomes, what is important is having generally strong relationships in
terms of trust, communication, and responsiveness and that no one type of relationship is
particularly more important than any other. However, this conclusion does not explain
the substantial differences in effect sizes. Effect size calculations indicated that the
unique variance accounted for by communication frequency and trust in follow through
had a moderate effect on coordination effectiveness while responsiveness to concerns had
a weak effect. Recognized expertise and shared philosophy appear to have little
relationship at all to coordination outcomes. Taken together, this provides some evidence
that frequent communication outside of DVCC meetings and a high level of trust among
stakeholders concerning likelihood for follow through may be particularly important
types of relationships for accomplishing coordination outcomes.

Interestingly, examining systems change effectiveness reveals a different pattern.
Modeled individually, all five indicators of social capital were significantly and
positively related to systems change effectiveness. When modeled together, two

indicators of social capital distinguished themselves as particularly important. Findings

118

indicated that dense networks with regards to communication frequency and shared
philosophy had moderate to strong effects on systems change and these effects were
significant.

These findings represent one of the first attempts to make explicit how different
characteristics of relationships impact different outcomes of interorganizational alliances.
While exploratory in nature, results provide the foundation for beginning to develop
some testable theories related to the differences between coordination and systems
change in terms of what characteristics of relationships are most important.

First, consistent with current literature on work teams (i.e., Abuja, 2002; Reagans
& Zuckerman, 2001) findings suggest that frequent interaction among members outside
of meetings is generally important to both coordination and systems change outcomes.
This makes sense as direct communication and interaction provide the vehicle and
opportunity for many network benefits to occur. For example, open channels of
communication can provide the opportunity for earlier identification of problems and
greater information sharing (Ahuja, 2002). However, it is important to acknowledge that
higher communication frequency may also be an outcome of coordination and systems
change. For example, it is possible that greater coordination may require participating
organizations and agencies to communicate with one another on a frequent basis. Given
the cross-sectional nature of the current study, the extent to which communication
frequency is a driver of greater coordination and system change, or an outcome of it,
remains ambiguous. While this is a limitation of the present study in general, this is of
particular concern for communication frequency as there is ample literature to indicate it

is both a driver and an outcome of coordination and systems change. Future research

119

utilizing more longitudinal designs are required in order to more clearly understand the
developmental directionality of this relationship.

Another result of interest in this study is the finding that shared philosophy, while
relatively unimportant for coordination outcomes, is the strongest factor in predicting
systems change outcomes. Shared philosophy, as it was used in this study, directly
relates to how stakeholders think about the issue of domestic violence and what beliefls
and assumptions they hold about the most effective means for addressing it. As argued
by Gamache & Asmus (1999), shared philosophy is not necessarily about shared goals, as
is a critical focus in many scholarly writing on multi-stakeholder coordination and
collaboration. Rather, it is about the underlying, often even unconscious, frameworks
that guide conceptualizations of a problem. The ability to bring together diverse
stakeholders with different perspectives is what makes DVCCs so promising as a vehicle
for systems change. However, findings indicate that fundamental differences in
philosophies concerning what domestic violence is and how it should be addressed
significantly hinders a DVCC’s ability to promote systems change. This finding is
consistent with other recent investigations into factors that facilitate or impede problem
solving in multi-stakeholder context. For example, in a recent case study of a protracted
environmental conflict, Gray (2004) concluded that the most significant factor preventing
a collaborative solution was differences among stakeholders in their frameworks of
understanding related to how they constructed the problem that linked them and how they
felt the problem should be resolved.

It is particularly interesting that shared philosophy appears uniquely important for

systems change; having relatively little importance for coordination types of outcomes.

120

This provides further empirical support to the work of Himmelman (2001) and others
who have argued that different types of collaborative action require different capacities.
Coordination outcomes principally have to do with how efficient the community system
is at channeling information and resources through the system to the stakeholders who
need them. For example, a common coordination outcome is the development of
protocols that proscribe how law enforcement agencies will communicate with victim
service providers when they identify a victim of domestic violence who may be in need
of services (Pence & Sheppard, 1994). Findings from this study suggest that having at
least some level of interorganizational communication and a level of trust that each
stakeholder will act as agreed upon facilitates these types of interorganizational
exchanges. However, it does not appear necessary for these stakeholders to hold similar
beliefs or understandings about the nature of the issue in order to carry out this
coordination function.

Systems change, on the other hand, involves making changes to the infrastructure
of the system itself. Scholars of systems change have argued that community systems
emerge out of and are therefore reflective of the attitudes, values, beliefs, and
understandings of the institutions of which they are made up (Martin, 1992). Given that
effective systems change efforts frequently require the sustained involvement of multiple
stakeholders over time (Markoff, Finkelstein, Kreiner, & Prost, 2005; Nissen, Merrigan,
& Kraft, 2005), it stands to reason that such changes would require that key community
stakeholders be in alignment concerning what that change should look like.

Such alignment may be particularly important in issue domains characterized by

long-standing differences in philosophy. The field of domestic violence has been marked

121

by a history of struggle among key stakeholders such as womens’ advocates, law
enforcement agencies, court systems, and mental health service providers directly related
to philosophical understandings about the nature of domestic violence (Pence &
McDonnel, 1999). For example, over the past several decades, issues related to whether
domestic violence is understood as a relational issue best managed through reconciliation
or a crime best managed through adjudication have been the at the heart of many
struggles in the field of domestic violence (Pence, 1988). If domestic violence is
understood as a crime, further disagreement among stakeholders exists concerning
whether it is viewed as a crime against a specific victim or more broadly thought of as a
crime against the community at large (Sheppard, 1999). Each of these understandings
carry with them significant implications for policy and practice in terms of the extent to
which and the ways in which the community intervenes to protect victims of domestic
violence and hold batterers accountable.

However, it is important to note that struggles over problem definition are not
unique to the field of domestic violence. Similar types of differences in philosophy have
been documented in many areas such community development, community health, and
criminal justice. For example, practitioners, scholars, and policy makers in these fields
have long differed in the extent to which they characterize and understand populations
based on their deficits or assets (Kretzmann, 1995; Trickett, Barone, & Watts, 2001),
whether they view problems from an individual or structural/institutional perspective
(Riessman, 1990), and whether they approach problem intervention from prevention or
remediation paradigm (Felner, Felner, & Silverman, 2001). This suggests that the

alignment of philosophies among stakeholders may have significant implications for

122

collaborative systems change efforts in many arenas beyond the field of domestic
violence.
Limitations

Interpretations of findings from this study need to take into account several
limitations. First, as discussed, the relationships hypothesized were tested using cross-
sectional data. As with all cross-sectional studies, the directionality of the relationships
are statistically ambiguous and statements of directionality are based solely on theory.

Second, DVCC coordination and systems change effectiveness were both
operationalized based on members’ perceptions of effectiveness. Effectiveness was
operationalized in this way in order to promote comparability across DVCCs for
hypothesis testing. Ideally in a study such as this there would be an objective measure of
effectiveness that would have been applicable across DVCCs. However, because
DVCCs go about improving coordination or promoting systems changes in a number of
ways, finding an objective outcome that translates as a measure of effectiveness relevant
across communities poses significant challenges.

In light of this, three strategies were employed to improve the validity of the
employed outcome measures. First, while the data were collected at the individual level,
the analysis operationalized effectiveness for both systems change and coordination
outcomes as the average perceived effectiveness across participants within a given
DVCC. This allows for perceptions across the various stakeholders to be taken into
consideration. Second, the use of HLM allows for differences in average perceived
effectiveness across DVCCs to be modeled in relation to the level of disagreement that

exists among members. Thus, meaningful variance that would be lost using standard

123

regression techniques employing aggregated measures of perceived effectiveness is taken
into consideration in HLM. This multi-level approach was particularly critical in this
study in light of the fact that there was significant variance in members’ perceptions of
effectiveness within any given DVCC. Lastly, items were based on measures published
in previous research with this population (Allen, 2005) and feedback on the validity of
the items was solicited from both state partners and DVCC members during piloting.
A related limitation is the possible inflation of correlations resulting from mono-
method bias. The key independent (e. g., social capital) and dependent variables (e.g.,
DVCC effectiveness) in this study were operationalized using data from DVCC
members. Such designs can be problematic as participants’ general affect (positive or
negative) can lead to generally more positive or negative responses on both independent
and dependent variables resulting in inflated correlations. However, in this study,
independent and dependent variables were operationalized in a way that minimizes the
likelihood of Type II errors resulting from a mono-method bias. Social capital was
operationalized using a social network approach as the average network density of
stakeholder relationships. This creates a DVCC level variable comprised of information
about how each stakeholder is perceived by every other stakeholder. Social capital was
then modeled in its relationship to differences in the average perceived effectiveness
between DVCCs. Given these operationalizations, a DVCCs score on social capital is
relatively independent from any one participants’ perceptions of their relationships with
the other members. It is theoretically possible that mono-method bias could result if an
entire DVCC was characterized by a generalized affect resulting from some factor

independent of any actual differences related to social capital or effectiveness. While

124

there is no research that would highlight this as a likelihood, it will be important in future
research to identify more objective community outcomes in order to validate the
relationship of social capital to effectiveness.

Fourth, issues of response bias are always of concern when relying upon self-
report or key informant types of designs. For example, it could be that certain members
may over or under report actual effectiveness based on a number of factors including
general personality characteristics, lack of knowledge, or a social motivation to present
their DVCC in the best (or worst) light possible. As one effort to minimize this risk,
both highly active and more sporadic members were included in representatives of their
DVCC. It should also be noted that, overall, there was good variability that existed
across DVCCs on all dependent variables.

Implications for Future Research

As with all new explorations, findings from this study have created many new
questions. The most significant findings from this study are that (1) social capital is a
significant predictor of DVCC effectiveness, (2) social capital appears to be most
strongly related to systems change effectiveness, and (3) coordination and systems
change outcomes are most heavily impacted by different qualities of relationships. These
findings provide the foundation for several exciting and promising directions for future
research. While many such implications have been discussed throughout, there are a few
others deserving of comment.

First, this study sought to examine the role of social capital to IA effectiveness by
examining this question within the context of specific type of IA — domestic violence

coordinating councils. However, IAs across issue domains share many similarities with

125

regards to their structure and function. As such, there is good reason to suspect that these
findings can contribute a valuable first step to developing a more generalized theory
concerning the role of social capital within IAs. In order to promote such theory
development, future research should focus on validating and expanding findings from this
study in other issue domains. This would provide an exciting opportunity for not only
building our generalized understanding about the role of social capital in IAs, but also for
developing a clearer understanding of the ways in which the study of social capital in IAs
is influenced by issue context.

Second, the present study conceptualized social capital as the cumulative strength
(density) of relationships across all stakeholders. One important area for future research
would be to expand this conceptualization of social capital to include alternative and
perhaps complementary operationalizations. For example, in examining community
mental health systems, Provan & Milward (1995) found that networks that were
organized around a specific core agency were more effective. While this study was
conducted within the context of an entire community service system and not an IA, it
leads to an important question concerning whether all relationships should be considered
equal. For example, within the context of DVCCs, it is reasonable to suspect that
relationships among certain critical stakeholder groups (e. g., prosecutors, victim services,
law enforcement) may be particularly important for DVCC effectiveness. Thus, an
important area for future research will be identify if certain stakeholder relationships are
more critical to DVCC effectiveness and if so, which ones, so as to improve

operationalizations of DVCC social capital.

126

Third, results from this study provide support for the premise that different types
of outcomes may require different types of capacity. Research on IAs in the field of
domestic violence as well as other issue domains have identified several factors such as
the quality of leadership and decision making practices (Allen, 2005; Butterfoss et al.,
1996; Mizrahi & Rosenthal, 2001), the age or developmental stage of the IA (Kreuter,
Lezin, & Young, 2000), the degree of structure adopted by the IA (Gollieb et al., 1993;
Jasuja et al., 2005; Kegler et al., 1998), the composition of the IA (F oster-F ishman et al.,
2001; Mattessich et al., 2001), and the amount of resources that the IA has available to it
(Mattessich et al., 2001; Weiss et al., 2002) that have been empirically linked to IA
functioning and effectiveness. In light of this, an important area for future research will
be to examine the relative importance of social capital in contrast to these other IA
factors for both coordination and systems change outcomes.

Lastly, given that social capital is a significant predictor of both coordination and
systems change effectiveness in DVCCs, another area for future research is to identify
what promotes the development of social capital within the context of an IA such as a
DVCC. For example, can IAs serve as a vehicle for promoting the development of
social capital among community stakeholders? If so, what characteristics of an IA are
most important for fostering social capital? What characteristics might hinder the
development of social capital? This type of information would be particularly important
to IA practitioners as it could inform the development of technical assistance strategies

designed to improve IA capacity for promoting social capital.

127

CONCLUSION

The notion that relationships are an important part of collaborative efforts such as
IAs is well-represented, if not almost axiomatic, in current literature. However, despite
this, there has been very limited investigation focused on understanding the role
relationships play within these very complex processes. As a result, we know very little
about the types of relationships that matter, the ways in which they matter, the extent to
which they matter, the conditions under which they matter, and the outcomes for which
they matter most. This lack of knowledge comes at a substantial cost as there is growing
evidence to suggest that the work of an IA is extremely challenging and many struggle
greatly in their efforts to accomplish their goals (Halfors, Cho, Livert, & Kadushin, 2002;
Roussos & F awcett, 2000). The theory of social capital asserts that relationships can
constitute an important resource within an IA that can facilitate the accomplishment of
productive outcomes (Coleman, 1988). As such, developing a more sophisticated
understanding of the role stakeholder relationships play in multi-stakeholder contexts can
lead to additional insights in diagnosing current IA capacity and identifying levers for
improving IA effectiveness.

The first step in examining the role relationships play in IA effectiveness is to
identify what types of relationships are indicative of social capital within that context.
This study examined five different characteristics of relationships supported by
qualitative data from DVCC key informants interviews and current literature as important
indicators of social capital. These were the extent to which stakeholders (l)
communicated with one another frequently outside of DVCC meetings, (2) perceived one

another to share a similar philosophy concerning what domestic violence is and how it

128

should be addressed, (3) felt they could count on each other to be responsive if a concern
was brought to their attention, (4) valued the expertise one another brought to the table,
and (5) trusted each other to follow through on what they said they would do. Findings
provide preliminary evidence that these characteristics do, in fact, collectively represent a
form of social capital within DVCCs that is positively associated with greater
effectiveness at both improving coordination among community organizations and
agency as well as promoting broader systems change.

The capacity continuum theories proposed by previous collaboration scholars
(i.e., Himmelman, 2001; Hogue, 1993; Gajda, 2003) have posited that cooperative
actions range along a continuum and that stronger and perhaps different capacities are
needed at different points along the continuum. Results fiom this study provide some
evidence to support these theories, finding that social capital appears to have a stronger
effect on systems change outcomes relative to coordination outcomes. Further, findings
indicate that not only are relationships more important for systems change outcomes, but
systems change requires a different type of relationship. Coordination outcomes appear
to be facilitated by generally strong relationships with potentially a greater emphasis on
frequent communication and trust in follow through among stakeholders. However,
systems change outcomes appear to be unique from coordination outcomes in their
reliance on the presence of a shared philosophy among stakeholders. Specifically, results
from this study suggest that in order for DVCCs to be most effective at changing policies,
addressing service gaps, and shifting public attitudes, stakeholders must believe that a
common understanding exists between them and the other members of the DVCC

regarding what domestic violence is and how it should be addressed.

129

Contrary to expectations, the amount of social capital that is present within a
DVCC does not appear to be an important factor in determining whether DVCC members
perceive their organization or agency to be positively impacted as a result of their
involvement in a DVCC. Rather, findings from this study suggest that it is organizational
or individual level factors that are the principal determinants of the extent to which
members perceive their involvement to have strengthened the capacity of their
organization. Finally, this study found that, while certain aspects of composition
diversity within a DVCC do have implications for DVCC effectiveness, there was no
evidence found to suggest that composition diversity has any impact on the relationship
between social capital and DVCC effectiveness.

In summary, DVCCs, like many IAs, are charged with the challenging assignment
of both improving the level of coordination among organizations within the existing
community system while also making needed changes to the infrastructure of the system
itself. The relationships among stakeholders within a DVCC can represent an important
form of social capital that can have significant implications for what can be accomplished
by a DVCC. However, the role of social capital in the context of DVCCs is not the same
for coordination as it is for system change. When the work of DVCC calls for not just
improving how information and resources flow through the existing domestic violence
response system, but rather actually making changes to the infrastructure of the system
itself, the DVCC will likely need to foster stronger relationships among participating
stakeholders than may have been needed for coordination. These relationship-building

efforts should focus particular attention to identifying what differences in philosophy

130

may exist among members about the issue of domestic violence and work toward

building more shared frameworks of understanding.

131

APPENDICES

132

Appendix A: Leader Recruitment letter

Dear [ leader’s name ]

As you know, efforts to facilitate a coordinated community response (CCR) to domestic
violence are occurring across the state of Michigan. Your name was referred to us by the
Michigan Coalition Against Domestic and Sexual Violence as a leader of the [insert
council name] involved in undertaking this task. With the support of the Prosecuting
Attomey’s Association of Michigan and the Michigan Coalition Against Domestic and
Sexual Violence, we are conducting a survey of domestic violence coordinating council
leaders and members from across the state. We know everyone’s lives are busy but we
hope that you and members of your council (or task force) will be willing to participate.
We have anticipated some questions you may have about this survey and have tried to
answer them below:

Why should I participate?
By participating in this study, you can help your council in two ways:

1. In appreciation for the time and effort involved in participating in this project, all
councils who have involvement from at least 75% of their members will be
entered into a drawing to win one of five $1000 awards that will go directly to
the council. Councils can put this money toward any purpose they see fit such as
helping to host a workshop, bring in a speaker, or assist with printing costs on a
project.

2. The purpose of this survey is to collect information that will directly benefit and
inform the work of Michigan domestic violence coordinating councils. With the
information collected through this survey, we will be able to provide answers to the
following types of questions:

a. What are member organizations and agencies getting out of being involved
with the councils? How are they being impacted? Are diflerent types of
organizations/agencies impacted in diflerent ways? If so, how?

b. What types of community changes are Michigan councils being most effective
at creating? What are the most common activities of Michigan domestic
violence coordinating councils?

c. How does the council structure, leadership, and climate impact its
eflectiveness?

d. Who needs to be at the table? How does the composition of the membership
impact the eflectiveness of a domestic violence coordinating council?

e. What do the relationships among members need to look like for the council to
be more eflective? How much does the eflectiveness of the council at
facilitating a CCR depend on the relationships among members?

Your council will have direct access to this information. At the conclusion of

the survey, the data will be written up in a straight forward, easy to understand

report that will be sent to your council. In addition to providing information that
can inform the strategic development of your council, this information may also
be useful to your council to in applying for grants and contracts.

133

What will participation in this study involve?

As a first step, we ask that you please send us a copy of your member list so that we can
send surveys to members of your council to request their participation. The survey
should take no more than 40 minutes to complete. If you could, we would appreciate you
notifying your members of the survey through an email and/or through an announcement
at your next council meeting. In addition, we would like to conduct a phone interview
with you that will take no more than 40 minutes to complete. Your participation and the
participation of your council members is completely voluntary.

How do I know the information I provide won ’t come back to haunt me?

All information you or any member of your council provides is strictly confidential. You
will provide the information directly to the researchers at Michigan State University and
they will be the only ones who have access to that information. Your information will
not be seen by MCADSV, PAAM, or anyone else not affiliated with the Michigan State
research team. The information you provide will be combined with the information
provided by other domestic violence coordinating councils from across the state of
Michigan. No report made from this data will be connected to a specific council or
council member but rather will consist of numbers that are summarized across councils.

How do I participate?

Within a week, we will be contacting your office to schedule an interview. Please send
your membership list to Brandy Nowell either by email (nowellbr@msu.edu) or fax [517-
432-2945]. Please feel free to contact Brandy Nowell at 517-353-9965 if you have any
questions.

Thank you in advance for your time and contribution.

Sincerely,

Brandy Nowell Mary Keefe
Department of Psychology Michigan Coalition
Michigan State University. Against Domestic and

Sexual Violence

134

Appendix B: Leader Interview Handout

CCR PROJECT: CCR COMPOSTION AND ACTIVITY WORKSHEET

The following set of questions concern the composition of your CCR council or task

force membership. The use of the term ‘active member’ refers to a member who has
attended at least one council meeting in the past year.

1) Since the council was formed:

a) approximately how many organizations or agencies have joined the council?

b) approximately how many organizations or agencies have left the council?
2) Approximately - what % of £9.03. members:

a) Joined in the past year? -————— %

b) Have been involved 1 to 3 years? —— %

c) Have been involved 4 or more years? —— %
3) Approximately, how many members:

(1) Have stopped being active in the past year?

e) Stopped being active 1 to 3 years ago?

f) Stopped being active 4 or more years ago?

4) Approximately - what percentage (%) of total membership is private/non-profit/public?

a) For profit %
b) Non-profit %
c) Public %

5) Approximately - what proportion of the total membership are:

a) Executives %
b) Middle managers/coordinators %
c) Staff %

135

6) Approximately - what percentage (%) of active membership is private/non-

profit/public?
a) For profit %
b) Non-profit %
c) Public %

7) Approximately - what proportion of active members are:

a) Executives %
b) Middle managers/coordinators %
c) Staff %

8) Approximately - what % of active members are men?

9) Approximately - what % of active members are people of colOr?

136

%

%

10) Next I want to ask you about how representative you feel your council is. On a scale of 1 to 5 - 1
being not at all and 5 being a great deal, to what extent does your council have active members from
those stakeholder groups most critical to carrying out the primary goals of your council?

 

Not at all Not much Somewhat Quite a bit A geat extent
1 2 3 4 5

 

 

 

 

 

 

 

 

Councils work on a variety of different things based on their own community needs.
Below is a list of different types of activities that different councils have engaged in.
They might or might not make sense for your council. In thinking about how your
council has focused its energies in the past year, on average, please indicate the % of
time and energy your council has spent on each activity. Just write 0% for those
categories your council has focused no time on:

" Atlvity V

effort spent

 

Dealing with issues related to council functioning such as
recruiting members, establishing council structures (e.g.,
bylaws), strategic planning, and dealing with internal
conflicts

Sharing information among members such as introducing
members to what each organization/agency does and
sharing tools, resources, or best practices pertaining to
domestic violence response and prevention

Developing protocols to better coordinate the practices of
different organizations/agencies

Identifying gaps and facilitating the development of new
programs or services

Working to improve existing programs and services

 

 

 

 

 

Administering/managing council run programs

 

Working to change policies related to domestic violence

 

Engaging in efforts designed to increase public knowledge
and awareness of domestic violence including pertinent
resources available in the community

Engaging in primary prevention activities focused on
changing attitudes and beliefs on violence, gender norms
and healthy relationships

Other? (describe):

 

 

 

 

 

 

Appendix C: Leader Interview Protocol

LEADER PHONE SURVEY

138

Backggound:
1) What is the name of your CCR

 

2) In what year was your CCR formed?

 

3) What is your position on your council?

 

4) How long have you been with this council?

 

5) How long have you held this position?

 

6) What is the name of the organization or agency you represent on the council?

 

7) What is your position with that organization or agency?

 

8) How long have you been in that
position?

 

139

Member List

Go through their membership list: For every member ask:
a) Confirm which stakeholder group they belong to (use list of groups provided to leader):
b) Ask leader to assign them as:

a. Active (attends meeting regularly and participates in council activities)
b. Non-active (rarely attends council meetings)

140

E

CCR PROJECT: CCR COMPOSTION AND ACTIVITY WORKSHEET

The following set of questions concern the composition of your CCR council or task
force membership. The use of the term ‘active member’ refers to a member who has
attended at least one council meeting in the past year.
1) Since the council was formed:

c) approximately how many organizations or agencies have joined the council?

d) approximately how many organizations or agencies have left the council?
2) Approximately - what % of Mg members:

g) Joined in the past year? —— %

h) Have been involved 1 to 3 years? —— %

i) Have been involved 4 or more years? —— %
3) Approximately, how many members:

j) Have stopped being active in the past year?

k) Stopped being active 1 to 3 years ago?

I) Stopped being active 4 or more years ago?

4) Approximately - what percentage (%) of total membership is private/non-profit/public?

d) For profit %
e) Non-profit %
1) Public %

5) Approximately - what proportion of the total membership are:

d) Executives %
e) Middle managers/coordinators %
f) Staff %

6) Approximately - what percentage (%) of active membership is private/non-profit/public?

141

a) For profit %
b) Non-profit %

c) Public %

7) Approximately — what proportion of active members are:

d) Executives %
e) Middle managers/coordinators %
f) Staff %
8) Approximately - what % of active members are men? %

9) Approximately - what % of active members are people of color?

142

%

18) Next I want to ask you about how representative you feel your council is. On a scale
of 1 to 5 — 1 being not at all and 5 being a great deal, to what extent does your council
have active members from those stakeholder groups most critical to carrying out the
primary goals of your council?

 

Not at all Not much Somewhat Quite a bit A great extent
1 2 3 4 5

 

 

 

 

 

 

 

 

19) Councils work on a variety of different things based on their own community needs.
Below is a list of different types of activities that different councils have engaged in.
They might or might not make sense for your council. In thinking about how your council
has focused its energies in the past year, on average, please indicate the % of time and
energy your council has spent on each activity. Just write 0% for those categories your
council has focused no time on:

and effort

 

Dealing with issues related to council functioning such as
recruiting members, establishing council structures (e. g.,
bylaws), strategic planning and dealing with internal conflicts
Sharing information among members about tools, resources,
or best practices pertaining to domestic violence response and
prevention

Developing protocols to better coordinate the practices of
different organizations agencies

 

Identifying gaps and facilitating the development of new
programs or services

 

Working to improve existing programs and services

 

Administering/managing council run programs

 

Working to change policies related to domestic violence

 

Engaging in public education and/or prevention-focused
activities

 

3. Other? (describe):

 

TOTAL

 

 

143

Resources

20) This next set of questions is about resources :

IT'— _T " '_"—' T 5 7h L

1' l I like to ask you about the exte ess your coun
{ 1 it needs to work effectively and to achieve its goals. I’m going to read off a list of
’ ‘ resources. For each of the following types of resources, please tell me on a scale of 1 to
5 —- 1 being not at all and 5 being completely - to what extent does your council currently
l l have what it needs to work effectively and to achieve its goals? If you think your
council does not need a particular resource to work effectively and achieve it’s goals,
you can 82in’ -— not applicable.

Not A To some Quite Completel . Not
at all little extent a bit y Appli . .
le

 

Money

 

Skills and expertise (e.g., leadership, public policy,
administration, evaluation, law, cultural competency,
training, community organizing)

 

Space

 

Equipment (e.g., Computers, books,)

 

Data and information (e.g., statistical data, information
about community perceptions, values, resources, and
politics)

 

Connections to target populations

 

Connections to political decision-makers, government
agencies or other organimtions or groups

 

Endorsements that give the council legitimacy and
credibility in the community

 

Influence and ability to bring people together for
meetings or other activities.

 

How important are the in-kind resources provided by
members, such as expertise and connections, in
supporting the activities of the council?

 

 

144

Organizational Structure

21) What precipitated the council forming (V AW/STOP grant funding, one agency
initiated the collaboration?

22) Which organization took lead in forming the council?

 

23) How often does council meet?

i. Bi-monthly,
ii. Monthly,
iii. Weekly,
iv. Other

 

24) Does council have paid staff: yes no

a) If yes — how many hours funded
for?

b) Which organization or funding source supports this persons
position?

25) Does the council have any funding or financial support? Yes/ no

26) If yes —- what types of financial support?

 

 

 

 

 

 

0 Federal grant $
0 State Grant $
0 County Grant 3
0 City Grant 3
0 Public Donation $
0 Fundraisers $
0 Other

27) What size of population does your council serve?

 

28) How would you characterize the community you serve?
a. Rural community

b. Small city

c. Large urban area

145

The next set of questions is about the structure and processes of your council/task force.
Does your council or task force:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Does your council or task force (circle
one):
29) Have a written agenda for its meetings? Yes No
30) Record and distribute minutes? Yes No
31) Have bylaws/rules of operation? Yes No
32) Have a mission statement in writing Yes No
33) Have goals and objectives in writing? Yes No
34) Have regular meetings? Yes No
35) Have an organization chart? Yes No
36) Have written job/role descriptions? Yes No
37) Have a core planning group? Yes No
38) Have subcommittees or workgroups? Yes No
a) If yes — how many?
b) if yes, how often do they meet. Bi- Monthly Weekly Other
monthly
39) Have established procedures for Yes No
decision making?
40) Have established processes for resource Yes No
allocation?
41) Have established mechanisms for Yes No
evaluating council impacts
42) Have a mechanism established for Yes No
accountability of members completing
asgnments in a timely manner?
43) Have a mechanism for new member Yes No
orientation?
44) Have domestic violence survivors as Yes No
members?
45) Have an advisory group of domestic Yes No
violence survivors?
46) Is your council a 501c3 nonprofit? Yes No
47) Do you have mechanisms for Yes No
communicating with council members
outside of meetings (e. g., list serve,
newsletter, etc)
a) IF yes: what type?

 

 

 

146

Appendix D: Stakeholder Categories

 

 

Stakeholder Group Categories

 

omestic Violence Service Providers

 

[Law enforcement

 

[Prosecuting Attorneys Office

 

[Courts

 

[Children's social service

 

[Health Service Organizations

 

[Mental Health Service Organizations

 

[Detention Probation

 

bommunity Service Organizations

 

[Batterers Intervention Services

 

[Legal Aid

 

[Faith Based Organizations

 

[Pre-School/K-12 Education

 

[Higher Education

 

boncemed Citizens

 

[Substance Abuse Organizations

 

[Youth Programs

 

[Domestic Violence Survivors

 

Eulturally Specific Community Organizations/groups

 

[Defense Attorney

 

[Local Business

 

[Media

kiouncil Staff

 

Eocal Government

 

[Housing Services

 

 

147

Appendix E: Member recruitment letter

Dear [ member’s name ]

As you know, efforts to facilitate a coordinated community response (CCR) to domestic
violence are occurring across the state of Michigan. We know that you are currently
involved with the [insert council name] in undertaking this task. With the support of
[insert council leaders name], the [Prosecuting Attomey’s Association of Michigan] and
the Michigan Coalition Against Domestic and Sexual Violence, we are conducting a
survey of domestic violence coordinating councils/task forces across the state. We know
everyone’s lives are busy but we would like to strongly encourage you to consider
participating in this effort. Even if you have not been an active council member in the
recent past, you represent the perspective of an important stakeholder group and your
perspective is valuable to understanding domestic violence coordinating councils. We

have anticipated some questions you may have about this survey and have tried to answer
them below:

Why should I participate?
By participating in this study, you can help your council in two ways:

1. In appreciation for the time and effort involved in participating in this project, all
councils who have involvement from at least 75% of their members will be
entered into a drawing to win one of five $1000 awards that will go directly to the
council. Councils can put this money toward any purpose they see fit such as helping
to host a workshop, bring in a speaker, or assist with printing costs on a project.

2. The purpose of this survey is to collect information that will directly benefit and
inform the work of Michigan domestic violence coordinating councils. With the
information collected through this survey, we will be able to provide answers to the
following types of questions:

a. What are member organizations and agencies getting out of being involved with
the councils? How are they being impacted? Are diflerent types of
organizations/agencies impacted in diflerent ways? If so, how?

b. What types of community changes are Michigan councils being most eflective at
creating? What are the most common activities of Michigan domestic violence
coordinating councils?

How does the council structure, leadership, and climate impact its eflectiveness?

Who needs to be at the table? How does the composition of the membership

impact the eflectiveness of a domestic violence coordinating council?

e. What do the relationships among members need to look like for the council to be
more efifective? How much does the eflectiveness of the council at facilitating a
CCR depend on the relationships among members?

Your council will have direct access to this information. At the conclusion of

the survey, the data will be written up in a straight forward, easy to understand

report that will be sent to your council. In addition to providing information that
can inform the strategic development of your council, this information may also
be useful to your council to in applying for grants and contracts.

9.9

148

What will participation in this study involve?

Participation in this study will involve filling out a survey that asks questions about your
experiences with your council. The survey will be sent directly to you in the form of an

email that contains a link to an web based survey which can be filled out on-line at your

convenience. The survey should take no longer than 40 minutes to complete.

How do I know the information I provide won ’t come back to haunt me?

All information you or any member of your council provides is strictly confidential. You
will provide the information directly to the researchers at Michigan State University and
they will be the only ones who have access to that information. Your information will
not be seen by MCADSV, PAAM, or anyone else not affiliated with the Michigan State
University research team. The information you provide will be combined with the
information provided by other councils from across the state of Michigan. No report
made from this data will be connected to a specific council or council member but rather
will consist of numbers that are summarized across councils.

How do I participate?
Simply look for an email from the CCR Project in your email within the next week. The
email will contain a link that will take you directly to the survey.

Thank you in advance for your time and contribution.

Sincerely,

Brandy Nowell Mary Keefe

Department of Psychology Michigan Coalition Against
Michigan State University Domestic and Sexual

Violence

149

Appendix F: DVCC Member survey

COUNCIL MEMBER SURVEY

Thank you for taking time to fill out the CCR project survey. By taking the time
to fill out this survey, you provide valuable information for informing the work of
domestic violence coordinating councils across Michigan. Even if you are not
very active with your council currently, you represent the perspective of an
important stakeholder group. As you know, councils can be powerful
strategies for improving the community’s response to domestic violence but
sometimes they struggle to accomplish their goals. This is a research study
designed to help us better understand how domestic violence councils can be
more effective.

 

Instructions:

Please respond to this survey regarding your perceptions and experience with
your council as a whole rather than only one subcommittee or work group. In
this survey, the use of the term ‘council’ refers to domestic violence coordinating
councils, task forces, and coordinating boards. For part of this survey, you will
be asked to reflect on the impact participating in the council has had on your
organization/agency. If only a part of the work of your organization/agency
concerns domestic violence, please respond on behalf of the particular unit,
office, or work group concerned with domestic violence.

 

150

 

I SECTION A: COUNCIL INVOLVEMENT

Regarding YOU:

 

1) Are you a member of a domestic violence coordinating council? (circle one) Yes
No

1a) If No - How many months ago did you stop being a member?
PLEASE CONTINUE EVEN IF YOU ANSWERED NO
2) When did you become involved with this council? Month Year

3) Approximately what % of council meetings have you personally attended over the
past 12 months? %

4) Do you hold a leadership position within your council or task force (e.g.,
chairperson)?
(circle one) Yes No

43) If Yes: Describe your position

 

5) At what level is your position within your organization or agency? (circle one)

a. Directorl head administrator
b. Middle-level administratorlsupervisor/coordinator
c. Project/program staff

Regarding YOUR ORGANIZAw (it AGENCY
6) When did your organization/agency become involved with this council?
__ Month ___Year
7) What % of council meetings has your organization or agency been represented at
(by you or 0zome one else from your organization/agency) over the past 12 months?
8) Yourself included, how much influence do the council representatives from your

organization/agency have to change practices or policies within your organization?
(circle one)

a) none b) very little c) some d) quite a bit e) a great deal

9) To what extent do you feel that historically there has been a culture of cooperation
among organizations and agencies in your community? (circle one)

a) not at all b) very little c) some d) quite a bit e) a great deal

151

 

l SECTION 3: COUNCIL RELATIONSHIPS I

The next set of questions asks about the nature of the relationships you and your
organization/agency have with the other members of your council. Remember,
your answers are completely confidential.

l-T #J¥#*T

 

10. Below is a list of the different stakeholder groups represented on your council. Please I
circle the number to the right that indicates how often over the past year your
organization/agency has communicated with any of the council members belonging to that
. roup outside of council meetings.

Never Once Once Once Two Every Don't Know

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

or every 3 a or week ( NI )
twice months month three
over times
the a

past month

year
a. Batterer's Intervention Service 1 2 3 4 5 6

Providers
b. Local Business 1 2 3 4 5 6
c. CivicNolunteer Organizations 1 2 3 4 5 6
d. Courts (i.e. judges) 1 2 3 4 ' 5 6
e. Cultural/Ethnic Groups 1 2 3 4 5 6
f. Domestic Violence Shelters/Service 1 2 3 4 5 6
Providers

9. Domestic Violence Survivors 1 2 3 4 5 6
h. Health Service Organizations 1 2 3 4 5 6
i. Higher Education 1 2 3 4 5 6
j. Legal Aids 1 2 3 4 5 6
k. Local Government 1 2 3 4 5 6
l. Media 1 2 3 4 5 6
m. Mental Health Agencies/Organizations 1 2 3 4 5 6
n. Police/Law Enforcement Agencies 1 2 3 4 5 6
o. Pre-SchoolIK-12 Education 1 2 3 4 5 6
p. Probation 1 2 3 4 5 6
q. Prosecuting Attomey’s Office 1 2 3 4 5 6
r. Religious Organizations 1 2 3 4 5 6
3. Social Services (i.e., HA) 1 2 3 4 5 6

 

 

152

III which you feel the members on your council representing that group share a common
I philosophy with your organization/agency regarding what domestic violence Is and how

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I11. Now, for each Stakeholder group below, please circle the number that Indicates the extent to
it should be addressed.
I I Not at A To Quite A Entirely I m not
I all Little some a bit great familiar
I 1 extent deal enough with
I' I them to know
I I (4)
‘ a. BattereI’s Intervention Service 1 2 3 4 5 6
I [ Providers
; b. Local Business 1 2 3 4 5 6
c. CivicNolunteer Organizations 1 2 3 4 5 6
I d. Courts (i.e. judges) 1 2 3 4 5 6
I e. Cultural/Ethnic Groups 1 2 3 4 5 6
f. Domestic Violence Shelters/Service 1 2 3 4 5 6
I I Providers
I 9. Domestic Vlolence Survivors 1 2 3 4 5 6
h. Health Services 1 2 3 4 5 6
i. Higher Education 1 2 3 4 5 6
j. Legal Aids 1 2 3 4 5 6
k. Local Government 1 2 3 4 5 6
i I. Media 1 2 3 4 5 6
l m. Mental Health 1 2 3 4 5 6
n. Police/Law Enforcement Agencies 1 2 3 4 5 6
o. Pre-SchooVK-12 Education 1 2 3 4 5 6
p. Probation 1 2 3 4 5 6
. q. Prosecuting Attorney's Office 1 2 3 4 5 6
l r. Religious Organizations 1 2 3 4 5 6
3. Social Services (i.e., FIA) 1 2 3 4 5 6
l

 

153

 

 

14712 Nw, for each stakhleodrgroup below, pleas cirle teh uer that iitst
: to which you feel the members on your council representing that group would be responsive
‘ if you or someone from your organization/agency brought a concern or issue to their

ll

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

attention.
Not at A little To Quite A Entirely I'm not
all some a bit great familiar
extent deal enough with
them to know
( 4 l
a. Batterer's Intervention Service 1 2 3 4 5 6
Providers
b. Local Business 1 2 3 4 5 6
c. Civic/Volunteer Organizations 1 2 3 4 5 6
d. Courts (i.e. judges) 1 2 3 4 5 6
e. Cultural/Ethnic Groups 1 2 3 4 5 6
f. Domestic VIolence Shelters/Service 1 2 3 4 5 6
Providers
9. Domestic VIolence Survivors 1 2 3 4 5 6
h. Health Services 1 2 3 4 5 6
i. Higher Education 1 2 3 4 5 6
j. Legal Aids 1 2 3 4 5 6
k. Local Government 1 2 3 4 5 6
I. Media 1 2 3 4 5 6
m. Mental Health 1 2 3 4 5 6
l n. Police/Law Enforcement Agencies 1 2 3 4 5 6
o. Pre—SchoollK-12 Education 1 2 3 4 5 6
p. Probation 1 2 3 4 5 6
q. Prosecuting Attorney's Office 1 2 3 4 5 6
r. Religious Organizations 1 2 3 4 5 6
l
’ s. Social Services (i.e., HA) 1 2 3 4 5 6
l

 

154

l l

. l 13. Some groups may have relatively more or less to offer i
violence.

helping counrlscto addrsrs domestic

For each stakeholder group below, please circle the number that indicates the extent to which the
members on your council representing that group offer a unique perspective or form of

expertise that has been valuable in understanding and/or addressing domestic violence.

fl

' l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A little To Quite A Entirely I’m not familiar
some a bit great enough with
extent deal them to know

( 4 )
a. Batterer‘s Intervention Service 2 3 4 5 6
Providers
b. Local Business 2 3 4 5 6
c. CivicNolunteer Organizations 2 3 4 5 6
d. Courts (i.e. judges) 2 3 4 5 6
‘ e. Cultural/Ethnic Groups 2 3 4 5 6
i f. Domestic Violence Shelters/Service 2 3 4 5 6
1 Providers
9. Domestic Violence Survivors 2 3 4 5 6
h. Health Services 2 3 4 5 6
i. Higher Education 2 3 4 5 6
j. Legal Aids 2 3 4 5 6
_ ' k Local Government 2 3 4 5 6
} f l. Media 2 3 4 5 6
I m. Mental Health 2 3 4 5 6 I
n. Police/Law Enforcement Agencies 2 3 4 5 6 I
o. Pre-SchoollK-12 Education 2 3 4 5 6
p. Probation 2 3 4 5 6 I
! q. Prosecuting Attomey’s Office 2 3 4 5 6 I
r. Religious Organizations 2 3 4 5 6
Social Services (i.e.. FIA) 2 3 4 5 6 J

 

155

E 14 Now,foraech stkeholder Mrupbcw, plesea Circel teta i een E
l which the members on your council representing that group can be trusted to follow through on E
l

what they say they will do.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Not at A little To Quite A Entirely I'm not familiar
all some a bit great enough with
extent deal them to know
( 1’ l
. Batterer‘s Intervention Service Providers 1 2 3 4 5 6
. Local Business 1 2 3 4 5 6
. Civic/Volunteer Organizations 1 2 3 4 5 6
. Courts (i.e. judges) 1 2 3 4 5 6
. Cultural/Ethnic Groups 1 2 3 4 5 6
. Domestic Violence Shelters/Service 1 2 3 4 5 6
Providers
. Domestic Violence Survivors 1 2 3 4 5 6
. Health Services 1 2 3 4 5 6 .
'. Higher Education 1 2 3 4 5 6
'. Legal Aids 1 2 3 4 5 6
. Local Government 1 2 3 4 5 6
. Media 1 2 3 4 5 6 l
. Mental Health 1 2 3 4 5 6
. Police/Law Enforcement Agencies 1 2 3 4 5 6 E
. Pre-SchoolIK-12 Education 1 2 3 4 5 6
. Probation 1 2 3 4 5 6 E
. Prosecuting Attomey’s Office 1 2 3 4 5 6
. Religious Organizations 1 2 3 4 5 6
. Social Services (i.e., FIA) 1 2 3 4 5 6

 

156

15)Sometimes other organizations and agencies can be a resource in trying to
accomplish certain goals. For example, they may provide you with valuable
information, provide you access to additional resources (e.g., staff time, office

space), or write letters in support of your organization/agency to help in obtaining
grants or contracts.

In thinking about the other organizations/agencies who are members of your
council, please list below the names of organizations or agencies that have

helped or supported your organization/agency in the past three years (up to
10):

 

 

 

 

 

 

 

 

 

 

16) In thinking about the other organizations/agencies who are members of your
council, please list below the names of those organizations/agencies with

whom your organization or agency currently coordinates services and
actions (up to 10):

 

 

 

 

 

 

 

 

 

 

 

157

COMMUNITY ASSESSMENT

Some communities may differ in the types of work that are most needed for

improving their response to domestic violence.

the following areas:

 

.Developing protocols to better coordinate the
practices of different organizations! agencies
concerning their response to domestic violence

None

1

Alittle

2

Some

3

Quite a
bit
4

A great
deal

5

 

Developing new or improving existing practices,
programs, and/or services relevant to domestic
violence perpetrators or survivors

 

Working to change public as well as institutional
policies related to domestic violence

 

Engaging in public education and/or prevention-
focused activities concerning domestic violence

158

 

 

1

“dines—.1
“mm.

1?"?

 

 

E SECTION C: COUNCIL EFFECTIVENESS E
Below are statements describing the effects a council may have on a community.
Please circle the number that indicates the degree to which each statement

[reoresents the effect our council has __d on our communi .

 

 

 

E I 8. Towhextent d el the fforts 0 or council have:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E Not A Som Most Quit A Don‘t E E
at all little e ly e a grea Know , .
what bit i (4) E I
deal E E
a. Increased the ability of organizations/agencies to 1 2 3 4 5 6 '
coordinate their efforts?
b. Influenced changes in the practices of 1 2 3 4 5 6
organizations/agencies that have increased batterer
accountability?
c. Influenced the policy of agencies regarding their 1 2 3 4 5 6
response to domestic violence?
.. ' E
E l
d. Got people talking about domestic violence? 1 2 3 4 5 6 E E
e. Been effective at facilitating needed changes in your 1 2 3 4 5 6 E E
community regarding how it responds to domestic E E
violence? E E
f. Has been effective at helping to reduce instances of 1 2 3‘ 4 5 6
domestic violence in your community? . E
E
9. Has been productive in accomplishing what it set out to 1 2 3 4 5 6 E
do? .
h. Resulted in members seeing their organization/agency 1 2 3 4 5 6 E
as part of a broader system for responding to domestic . .
violence? E E
i. Increased members’ knowledge of the strengths as well 1 2 3 4 5 6 E E
E as limitations of each others organizations/agencies? E E
.. E
j. Resulted in organizations and agencies working 1 2 3 4 5 6
together more efficiently?
l
k. Expanded or improved existing programs or services 1 2 3 4 5 6
for women and children affected by domestic violence?
E I. Influenced changes in the practices of 1 2 3 4 5 6 E
organizations/agencies that have increased women’s E
E safety? E
m. Played a role in developing new programs or services 1 2 3 4 5 6 E
for battered women and their children? E
E n. Stimulated policy changes within your organization 1 2 3 4 5 6
E regarding its response to domestic violence?
E
E 0. Made the public more aware of what they can do if they 1 2 3 4 5 6
E I or someone they know is being affected by domestic
E 1 violence?
E E
E I p. Made good headway at fixing the biggest problems with 1 2 3 4 5 6
E E your community’s response to domestic violence?

159

 

 

Influenced city, county, or state legislation concerning
domestic violence?

 

Made your community less tolerant of domestic
violence?

 

Consistently moved the council closer to achieving its

 

 

ORGANIZATIONAL IMPACT

Next are statements describing the impact belonging to a council may have on
member organizations and agencies. Please circle the number that shows the

degree to which each statement represents the impact belonging to the council
has had on your organization or agency

 

19. For my organzaitionlalmncy, particpation in the coucil has led to:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Not A To Quite A great Not ' .
at all little some a bit deal applicable
extent to my
organizatio.
n E .
(ill .
a. The generation of new ideas for improving our practices 1 2 3 4 5 : ;
and/or services E
E
b. The acquisition of useful knowledge about services, 1 2 3 4 5 E
programs, or people in the community E E
E
c. A decrease in the number or severity of barriers we face 1 2 3 4 5 E E
in accomplishing our mission E .'
' E
d. A greater ability to identify the source of the problems we 1 2 3 4 5 E E
encounter in order to come up with more effective E E
solutions E E
e. A heightened public profile for my organization/agency 1 2 3 4 5 E E
Increased our ability to affect public policy 1 2 3 4 5 E E
E
9. Increased access to tools, best practices, and/or other 1 2 3 4 5 E
information that has informed the work of my E E
organization E E
h. Greater knowledge about how the system works and 1 2 3 4 5 E E
how organizations and agendas affect one another E
E
An increase in our ability to find the answers to 1 2 3 4 5
questions or problems that arise
An improvement in our ability to compete for grants 1 2 3 4 5
and/or other funding opportunities
k. Increased utilization of my organization/agency's 1 2 3 4 5
expertise or services

160

organization could have on its own

 

An enhanced ability to meet the needs of my
organization's constituency or clients

 

Increased understanding of the dynamics of domestic
violence

 

Increased knowledge about how to best interact with
other organizations in order to accomplish our objectives

 

Increased access to more or different types of resources
(e.g., funding, staff, equipment, office space).

 

Increased knowledge of the limitations and constraints
faced by other organizations

 

 

 

 

 

E ‘ 20. For m or anizationlagency, participation in the council has led to:
E Not A To Quite A great Not

 

 

 

 

 

E E

E

E E at all little some a bit deal applicable

l E extent to my

E E rganizatio .

I E ’ (1 I

E E a. An increase in the level of respect and credibility we 1 2 3 4 5

E E have with other agencies and organizations

E

. E

E E b. An increase in other members' trust that we can be 1 2 3 4 5

E E relied upon to do what we say we're going to do

E .
E ,

E E c. An increase in how responsive other organizations and 1 2 3 4 5 E ‘

E E agencies are to our questions or concerns

E E

E . d. An increase in how supportive other organizations and 1 2 3 4 5

E E agencies are of my organization/agency

I I e. A reduction in the differences in philosophy between my 1 2 3 4 5

I E organization and the other organizations on the council

 

161

 

SECTION D: COUNCIL LEADERSHIP E

 

 

Councils often have several people who take on leadership roles at different
times within the council. Across this group of people, please circle the number
below that best indicates the degree to which the statement to the left reflects the
leadership in your council.

3. Theweders in ycoil: ’5 _

disagree What What
Disa ree A ree

Talk about the values behind our work 1 3 4

 

Are well respected and/or admired within our community 1 3 4

 

Help members see the interconnections between their work 3 4

 

Inspire people to do more than what is expected of them

 

Cause me to rethink things which I have never questioned
before

 

Make me less critical of creative ideas

 

Take time to find out what the concerns and issues of my
organization/agency are

 

Would contact me if there was something on the next
meeting's agenda that concerned me or my
on anization/a - ency

Let me know that my organization/agency and l are valued
members of the council

 

Are adept at fostering respect, trust, inclusiveness, and
openness in the council

 

Are competent at integrating diverse viewpoints and
managing conflicts among members

 

Help us all stay focused on what we’re trying to achieve

 

Plan meetings effectively and efficiently

 

Are adept at obtaining resources in the community

 

The leadership in our council is shared among the different
stakeholders

 

162

 

 

SECTION E: PARTICIPANT DEMOGRAPHIC AND CHARACTERISTICS J

 

26) What year were you born?
30) Is the group/organizationlagency you
represent on the council a:
27) Are you (circle one): Male Female
3. Private for profit

 

28) What best describes your racial/ethnic business! organization
background? b. Not for profit organization
a. African American/Black c. Public agency/institution
b. Asian/Pacific Islander d- Community-based group
C. Hispanic/Latino
d. Native American 31a) If answered a or b: What is the
e. White/Caucasian size of your organization? (# of
f. Other. staff)?___
29) What is the highest degree you have 30) How many years have you worked for
received (circle one) the organization or agency you currently
a. Did not graduate from high work for? _
school

GED or high school diploma 32) How many years have you been in your

Associate's degree current posmon?

Master’s degree 33) How many years have you worked in the

b

c

d. Bachelor’s degree

e

f Ph.D., MD, or J0 area of domestic violence?
9

Other?

 

163

THANK YOU”

You have completed the CCR Project survey! Please return this survey directly
back to the CCR Project office at MSU in the included postage paid envelop or to:
Brandy Nowell, Department of Psychology, Michigan State University, East
Lansing, MI 48824. If you have any questions or concerns, please contact me at
nowellbr@msu.edu or by phone at (517) 353-9965. The findings generated from
this project will be ready to distribute and will be mailed to your council directly
in early 2006.

164

Appendix G: Follow up reminder

Dear #NAME#,

This is a just reminder email concerning the CCR Project Survey. Your knowledge
and experiences are valuable to helping us understand CCR5 and how to make them
more effective. While participation is completely voluntary, we would like to
encourage you to take the time to contribute. The survey should take no longer than
40 minutes to complete.

If you would like to participate, just click on the link below. If you do not want to
participate and would like to request that no further reminders be sent to you, please
reply to this email. If you have any questions or difficulty accessing the survey,
please feel free to contact me by this email or by phone at (517) 353-9965.

TO GO TO THE SURVEY, COPY AND PASTE THIS URL INTO YOUR WEB
BROWSER OR CLICK HERE: #LOGIN_URL#

Thank you for your time.
Sincerely,
Brandy Nowell

Michigan State University
Department of Psychology

165

Appendix H: DVCC Effectiveness at promoting community change

DVCC EEFFECTIVENESS AT PROMOTING COMMUNITY CHANGE
SUB-SCALE SUMMARY

 

SUB-SCALE BREAKDOWN

 

Dimension

Items (To what extent do you feel the efforts of your
council have...)

 

a. Coordinating existing services

0

O

0

Increased the ability of organizations/agencies to
coordinate their efforts.

Resulted in organizations and agencies working
together more efficiently.

Resulted in members seeing their
organization/agency as part of a broader system
for responding to domestic violence

Increased members’ knowledge of the strengths
as well as limitations of each other’s
organizations/agencies.

 

b. Improvement of existing
services/practices and/or
development of new

Influenced changes in the practices of
organizations/agencies that have increased
women’s safety

Influenced changes in the practices of
organizations/agencies that have increased
batterer accountability.

Expanded or improved existing programs or
services for women and children affected by
domestic violence

Played a role in developing new programs or
services for battered women and their children

 

c. Policy change

Influenced the policy of agencies regarding their
response to domestic violence.

Stimulated policy changes within my

organization regarding our response to domestic

violence.

Organized to influence city, county, or state

legflation concemigg domestic violence

 

(1. Public Education (short term type of
programming)

Got people talking about domestic violence
Made our community less tolerant of domestic
violence.

Made the public more aware about what they
can do if they or someone they know is being
affected by domestic violence

 

e. Goal accomplishment

 

 

Been effective at facilitating needed changes in
our community regarding how we respond to
domestic violence

Made good headway at fixing the biggest
problems with our community’s response to
domestic violence

Has been effective at helping to reduce instances
of domestic violence in our community
Consistently moved the council closer to
achieving its goals.

Has been productive in accomplishing what it
set out to do.

 

166

 

Appendix I: Qualitative Analysis for Strengthened Organizational Capacity Scale

MEASUREMENT DEVELOPMENT:
STRENGTHENED ORGANIZATIONAL CAPACITY SCALE
CROSSWALK OF QUALITATIVE INTERVIEWS WITH EXISTING LITERATURE

 

 

 

Dimension Key informant themes Literature Items
Participation in the
, * council has led to: '
Information V Increased access to Bailey & Koney 1. Increased
attainment/ information [data] on (2000) access to tools,
Organizational community statistics 0 Alliances best practice,
learning V Gain access to benefit and/or other
information about organizations information that
available resources through has informed
and best practices increase the work of my
V Increased access to information organization
useful tools resources sharing 2. Acquisition of
(i.e., information Foster-Fishman et useful
books) al 2001 knowledge
V Gain knowledge of o Qualitative about services,
what other study of best programs, or
orgs/agencies are practice people in the
doing utilization in community
V Increased non-profits 3. Greater
understanding of the suggested that knowledge
constraints and inter- about how the
limitations of the organizational system works
other agencies/orgs relationships and how
V Increased knowledge were key to organizations
of how the system organizational and agencies
works and what learning and affect one
resources are best practice another
available, how they utilization 4. Increased
work Ahuja 2000 understanding
V Keep up to date about 0 Proposes that of the dynamics
what’s going on in inter-org of domestic
the community and linkages can be violence
with other programs an conduit 5. Increased
V Greater through with knowledge of
understanding of how information is the limitations
changes impact the shared and constraints
entire system (Freeman, face by other
V Increased 1997) access to organizations.
_ understandiwf the best practices

 

 

 

 

 

167

 

issues involved in
DV cases

Improved knowledge
about how to handle
cases more
effectively

(Rogers &
Larson, 1984).

Burt 1992

o Inter-
organization
a1
relationships
assists orgs
to keep up to
date on
developing
opportunities

Reitan, 1998

o organizations
form
relationships
with other
organizations
to increase
op for
organization
a1 learning

 

 

Increased
capacity for
solving
problems that
arise

 

Increased access to
information for how
to deal with problems
Improved
communication with
other agencies — can
ask questions without
defensiveness
Increased ability to
solve problems facing
organization

Better able to
navigate the system
and get problems
solved because
personally familiar
with the people and
you know how to
work with/influence
them

Increased knowledge
about how to best
work with other

 

Alter & Hage 1993

0 Increased
ability to
manage
uncertainty,
solve
problems
(Trist, 1983;
Aldrick,
1979; &
Hage, 1988)

Abuja 2000

O

Proposes that
inter-org
linkages
produce
increased ability
for problem
solving
(Freeman,

1982)

 

. A decrease in

the number or
severity of
barriers we face
in
accomplishing
our mission

. Increased

ability to find
the answers to
questions or
problems that
arise

. *Increased

knowledge
about how to
best interact
with other
organizations in
order to
accomplish our
objectives

. Increased

 

168

 

 

organizations and
agencies to get the
best outcome

More
knowledgable/sawy
about how to impact
the system — more

ability to
correctly
identify the root
of problems we
encounter in
order to come
up with more

 

 

empowered to act effective
[we don’t waste time solutions
saying ‘can you
belive that? We know
what to do or who to
call]
Better able to
diagnose the root of
problems and come
up with more
effective solutions
(increased the #
possible solutions
that can be applied to
any given problem)

Sustainability Increased utility of Austin (2000) An
one’s organization 0 More client improvement in
‘they call us’ exchanges our ability to
More referrals into compete for
services Foster-Fishman et grants and/or
Made resources a1 (2001) other funding
available to carry out o participation in opportunities
activities council was Increased
organization would related to access to more
have otherwise not greater or different
been able to get centrality in the types of
involved in referral resources (e. g.,
More competitive network funding, staff,
getting grants (due to equipment,
letters and MOUs) Austin (2000) office space).
Increased access to 0 Improved Increased
resources for services to utilization of
expanding clients my
programming/service organization/ag
3 Access to resources ency’s expertise
(equipment, space) as or services
a result of affiliation Enhanced

 

with the council

 

 

ability to meet

 

169

 

 

 

 

 

 

V Improve existing
services — expand
services

 

 

the needs of my
organization’s
constituency or
clients

 

170

 

 

 

 

 

 

Influence/

visibility

 

Gain influence with
other organizations in
community — better
able to advocate for
clients

Increased attention
and visibility to the
other agencies — ‘they
are more likely to
throw us a bone’
‘they know we are
looking’

Have a forum for
influencing the
thinking/strategies of
other
agencies/organization
3

Greater access to
important people in
the community
Opportunity to have
influence into
decisions that would
otherwise not be
involved in

Increased the
visibility of
organization within
the community as a
result to being
affiliated with the
council

Increased community
support for and
confidence in the
program = increases
the longevity of
programs

Increased credibility
in the community and
with the other
stakeholders

 

Alter & Hage 1993
o Gain
influence of
domain
(Alinsky,
1971

Bailey &

Koney,2001

0 Provide a
forum 0 affect
the distribution
of power
among
organizations
and give single
orgs more
power in a
particular issue
domain than it
could have
amassed alOne
(Black, 1983;
Rosenthal &
Mizrahi, 1994)

Proven & Milward

(2001)

o Legitimacy:
acquiring
status and
acceptabilit
y in the
community

Bailey & Koney:

2000

0 Increase
organization’s
input into the
management of
broader issues
that may be
beyond their
individual
scope or
capabilities

 

. Increased level

of respect and
credibility with
other agencies
and
organizations

. Enhanced

ability to affect
public policy

. Ability to have

greater impact
than my
organization
could have on
its own

. Heightened

public profile of
my
organization/ag
ency

 

171

 

 

 

 

 

Mizrahi &
Rosenthal (2001)
o Developme

nt of
broader
community
support for
individual
member
org’s cause

 

172

 

 

 

 

REFERENCES

173

Adler, P. S., & Kwon, S. (2002). Social capital: Prospects for a new concept. Academy of
Management Review, 27(1), 17-40.

Ahuja, G. (2000). Collaboration networks, structural holes, and innovation: A
longitudinal study. Administration Science Quarterly, 45(3), 425-455.

Aiken, L., & West, S. (1991). Multiple Regression: Testing and Interpreting Interactions.
Thousand Oaks: Sage.

Aldrich, H., & Zimmer, C. (1986). Entrepreneurship through social networks. In R. W.
Smilor (Ed.), The Art and Science of Entrepreneurship (pp. 3-24). Cambridge:
Ballinger.

Allen, N., Watt, K., & Hess, J. Z. (2003, July). Collaboration and social change in the
domestic violence context: Exploring potentiality and pitfalls. Paper presented at

the 9th Biennial Conference of the Society for Community Research and Action,
Las Vegas, NM.

Allen, N. E. (2001). Coordinating councils as vehicles for achieving a coordinated
community response to domestic violence: An examination of the correlates of
council eflectiveness. Allen,-Nicole-Elizabeth: Michigan State U., US.

Allport, G. (195 8). The Nature of Prejudice. Garden City, NY: Doubleday.
Alter, C., & Hage, J. (1993). Organizations Working Together. Newbury Park: Sage.

Ancona, D., & Caldwell, D. (1992). Demography and design: Predictors of new product
team performance. Organization Science, 3, 321-341.

Austin, J. (2000). The Collaboration Challenge: How Nonprofits and Business Succeed
Through Strategic Alliances. San Francisco: J ossey-Bass.

Bailey, D., & Koney, K. M. (2000). Strategic Alliances Among Health and Human
Service Organizations: From Afiiliations to Consolidations. Thousand Oaks:
Sage.

Baker, W. (2000). Achieving Success Through Social Capital. San Francisco: J ossey-
Bass.

Balkundi, P., & Harrison, D. (2004). Networks, leaders, and time in teams: Connections
to viability and performance. Manuscript submitted for publication.

174

Bantel, K., & Jackson, S. (1989). Top management and innovations in banking: Does the
composition of the team make a difference? Strategic Management Journal, 10,
1 07-1 24.

Bartunek, J. M., Foster Fishman, P. G., & Keys, C. B. (1996). Using collaborative
advocacy to foster intergroup cooperation: A joint insider—outsider investigation.
Human Relations, 49(6), 701-733.

Bond, M. A., & Keys, C. B. (2000). Strengthening parent-community member relations
on agency boards: Comparative case study. Mental Retardation, 38(5), 422-435.

Borgatti, S. P., & Foster, P. C. (2003). The network paradigm in organizational research:
A review and typology. Journal of Management, 29(6), 991 -101 3.

Brass, D. J ., & Burkhardt, M. (1992). Centrality and power in organizations. In R. Eccles
(Ed.), Networks and Organizations (pp. 191-215). Cambridge: Harvard Business
School Press.

Bryson, J ., & Crosby, B. (1992). Leadership for the Common Good. San Francisco:
Jossey-Bass.

Building a coordinated community response seminar. (2000). Traverse City: Prosecuting
Attorney's Association of Michigan Violence Against Women Project.

Burt, R. S. (1992). Structural Holes: The Social Structure of Competition. Cambridge:
Harvard University Press.

Burt, R. S. (2000). The network structure of social capital. In B. M. Staw (Ed.), Research
in organizational behavior (V 01. 22): Elsevier Science.

Butterfoss, F. D., Goodman, R. M., & Wandersman, A. (1993). Community coalitions for
prevention and health promotion. Health Education Research, 8(3), 315-330.

Campbell, J. C., Dienemann, J ., Kub, J ., Wurmser, T., & Loy, E. (1999). Collaboration as
a partnership. Violence Against Women, 5(10), 1 140-1 157.

Chaiken, S., & Stangor, C. (1987). Attitudes and attitude change. Annual Review of
Psychology, 38, 575-630.

Clark, N. M., Baker, E. A., Chawla, A., & Maru, M. (1993). Sustaining collaborative

problem solving: Strategies from a study in six Asian countries. Health Education
Research, 8(3), 385-402.

175

Clark, S. J ., Burt, M. R., Schulte, M. M., & Maguire, K. (1996). Coordinated community
response to domestic violence in six communities: Beyond the Justice System
(Final report to the US. Department of Health and Human Services). Washington,
D.C.: The Urban Institute.

Coe, B. A. (1988). Open focus: Implementing projects in multi-organizational settings.
International Journal of Public Administration, 1 1(4), 503-526.

Cohen, D., & Prusak, L. (2001). In Good Company: How Social Capital Makes
Organizations Work. Boston: Harvard Business School Press.

Cohen, J. (1988). Statistical Power Analysis for the behavioral sciences (2nd Edition
ed.). Hillsdale, NJ: Lawrence Erlbaum.

Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal
of Sociology, 94, 895-81 11.

Cook, T. D., & Shadish, W. R. (1986). Program evaluation: The wordly science. Annual
Review of Psychology, 3 7, 193-232.

Cooke, P., & Wills, D. (1999). Small firms, social capital and the enhancement of
business performance through innovation programmes. Small Business
Economics, 13(3), 219-234.

de J ong, A. (1996). Inter-organizational collaboration in the policy preparation process,
Huxham, Chris (Ed). (1996). Creating collaborative advantage. (pp. 165 175).
Thousand Oaks, CA, US: Sage Publications.

Dirks, K. T., & Ferrin, D. L. (2001). The role of trust in organizational settings.
Organization Science, 12(4), 450-467.

F elner, R., Felner, T., & Silverman, M. (2001). Prevention in mental health and social
intervention: Conceptual and methodological issues in the evolution of the science
and practice of prevention. In E. Seidman (Ed.), Handbook of Community
Psychology (pp. 9-42). New York: Kluwer.

Floge, L., & Merrill, D. (1985). Tokenism reconsidered: Male nurses and female
physicians in a hospital setting. Social Forces, 64, 925-947.

Foster Fishman, P. G., Salem, D. A., Allen, N. A., & Fahrbach, K. (2001). Facilitating

interorganizational collaboration: The contributions of interorganizational
alliances. American Journal of Community Psychology, 29(6), 875-905.

176

Foster-Fishman, P., Berkowitz, S., Lounsbury, D., Jacobson, S., & Allen, N. (2001).
Building collaborative capacity in community coalitions: A review and

integrative framework. American Journal of Community Psychology, 29(2), 241-
261.

Fowler, S. A., Donegan, M., Lueke, B., Hadden, D. S., & Phillips, B. (2000). Evaluating
community collaboration in writing interagency agreements on the age 3
transition. Exceptional Children, 67(1), 35-50.

Frank, K. A., & Zhao, Y. (2004). Subgroups as meso-level entity in social organization in
schools. In B. Schneider (Ed.), Social Organization of Schools (pp. 279-318).
New York: Sage.

Gajda, R. (2003). Utilizing collaboration theory to evaluate strategic alliances. Paper
presented at the American Evaluation Association, Reno, NV.

Gamache, D., & Asmus, M. (1999). Enhancing networking among service providers:
Elements of successful coordination strategies. In E. L. Pence (Ed.), Coordinating

Community Responses to Domestic Violence: Lessons from Duluth and Beyond.
Thousand Oaks: Sage.

Glisson, C., & James, L. (1992). The interorganizational coordination of services to
children in state custody. In H. Schmid (Ed.), Organizational change and
development in human service organizations (pp. 65-80). New York: The
Hawthorn Press.

Gray, B. (1989). Collaborating: Finding Common Ground for Multiparty Problems. San
Francisco: Jossey-Bass.

Gray, B. (1996). Cross-sectoral partners: Collaborative alliances among business,
government and communities, Huxham, Chris (Ed). (1996). Creating
collaborative advantage. (pp. 57 79). Thousand Oaks, CA, US: Sage Publications.

Gray, B. (2004). Strong opposition: Frame-based resistance to collaboration. Journal of
Community and Applied Social Psychology, 14(3), 166-176.

Halfors, D., Cho, H., Livert, D., & Kadushin, C. (2002). Fighting Back Against
Substance Abuse: Are Community Coalitions Winning? American Journal of
Prevention Medicine, 23(4), 237-245.

Hardy, C., & Phillips, N. (1998). Strategies of engagement: Lessons from the critical
examination of collaboration and conflict in an interorganizational domain.
Organization Science, 9(2), 217-230.

Hargadon, A., & Sutton, R. (1997). Technology brokering and innovation in a product
development firm. Administrative Science Quarterly, 42(4), 716-749.

177

Hart, B. J. (1995). Coordinated community approaches to domestic violence. Paper
presented at the Strategic Planning Workshop on Violence Against Women,
National Institute of Justice, Washington DC.

Himmelman, A. T. (I 996). On the theory and practice of transformational collaboration:
From social service to social justice, Huxham, Chris (Ed). (1996). Creating
collaborative advantage. (pp. 19 43). Thousand Oaks, CA, US: Sage Publications,
Inc.

Hofrnann, D. A., Griffin, M., & Gavin, M. (2000). The application of hierarchical linear
modeling to organizational research. In S. Kozlowski (Ed.), Multilevel theory,
research, and methods in organizations. San Francisco: J ossey Bass.

Hoge, M. A., & Howenstine, R. A. (1997). Organizational development strategies for
integrating mental health services. Community Mental Health Journal, 33(3), 175-
187.

Hogue, T. (1993). Community based collaboration: Community wellness multiplied.
Oregon State University: Oregon Center for Community Leadership. Retrieved
January, 2004, from the World Wide Web:
http://crs.uvm.edu/nnco/collab/wellness.htm1

Huxham, C. (1996). Collaboration and collaborative advantage. In C. Huxham (Ed.),
Creating Collaborative Advantage (pp. 1 - 18). Thousand Oaks, CA, US: Sage
Publications, Inc.

Ibarra, H. (1993). Personal networks of women and minorities in management: A
conceptual framework. The Academy of Management Review, 18(1), 56-87.

Jackson, S. (1992). Team composition in organizational settings: Issue in managing a
diverse work force. In J. Simpson (Ed.), Group process and productivity. Beverly
Hills: Sage.

Jackson, S. E., & Joshi, A. (2004). Diversity in social context: A multi-attribute,

multilevel analysis of team diversity and sales performance. Journal of
Organizational Behavior, 25 , 675-702.

Jenkins, P., & Davidson, B. (1999). Consensus and contradictions in understanding
domestic violence: Implications for policy and model programs. In S. J.
McElhaney (Ed.), Violence in homes and communities: Prevention, intervention,
and treatment. Thousand Oaks: Sage.

Jones, R. B. (2003). Capacity building in human service organizations. New Directions
for Philanthropic F undraising, 40(summer), 15 - 27.

178

Kameda, T., Ohtsubo, Y., & Takezawa, M. (1997). Centrality in sociocognitive networks
and social influence: An illustration in a group decision-making context. Journal
of Personality and Social Psychology, 73(2), 296-309.

Kegler, M. C., Steckler, A., Malek, S. H., & McLeroy, K. (1998). A multiple case study
of implementation in 10 local Project ASSIST coalitions in North Carolina.
Health Education Research, 13(2), 225-238.

Kelly, J. G., Ryan, A. M., Altman, B. E., & Stelzner, S. P. (2001). Understanding and
changing social systems: An ecological view. In E. Seidman (Ed.), Handbook of
Community Psychology (pp. 133-160). New York: Kluwer.

Kieman, N., Kieman, M., Oyler, M. A., & Gilles, C. (2005). Is a web survey as effective
as a mail survey?: A field experiment among computer users. American Journal
of Evaluation, 26(2), 245-252.

Knoke, D. (1999). Organizational networks and corporate social capital. In S. M. Gabbay
(Ed.), Corporate Social Capital and Liability (pp. 17 - 42). Norwell, MA: Kluwer
Academic.

Kohm, A. (1998). Cooperating to survive and thrive: Innovative enterprises among non-
profit organizations. Nonprofit World, 16(3), 36-44.

Kraatz, M. (1998). Learning by association?: lnterorganizational networks and adaptation
to environmental change. Academy of Management Journal, 41(6), 621-643.

Krackhardt, D. (1992). The strength of strong ties: The importance of philos in
organizations. In R. Eccles (Ed.), Networks and Organizations: Structure, Form,
and Action (pp. 216-239). Boston, MA: Harvard Business School Press.

Kretzmann, J. P. (1995, September/October 1995). Building Communities From the
Inside Out. Shelterforce Magazine.

Lasker, R. D., Weiss, E. S., & Miller, R. (2001). Partnership synergy: A practical
framework for studying and strengthening the collaborative advantage. The
Milbank Quarterly, 79(2), 179-204.

Leenders, R. T., & Gabbay, S. M. (1999). Corporate Social Capital and Liability. Boson:
Kluwer Academic.

Letts, C., Ryan, W., & Grossman, A. (1999). High performance non-profit organizations:
Managing upstream for greater impact. New York: Wiley.

Lin, N. (2001). Social Capital: A Theory of Social Structure and Action. Cambridge:
Cambridge University Press.

179

 

Lincoln, 1. R., & Miller, J. (1979). Work and fi'iendship ties in organizations: A
comparative analysis of relational networks. Administrative Science Quarterly,
24, 181 - 199.

Linney, J. A. (2000). Assessing Ecological Constructs and Community Contexts. In E.
Seidman (Ed.), Handbook of Community Psychology (pp. 647-668). New York:
Kluwer Academic/Plenum.

Lonsway, K., Carrington, S., Aguirre, R, Wood, M., Moore, M., Harrington, P., Smeal,
E., & Spillar, K. (2002). Equality Denied: The Status of Women in Policing:
National Center for Women and Policing.

Markoff, L., Finkelstein, N., Kreiner, P., & Prost, C. (2005). Relational systems change:
Implementing a model of change in integrating services for women with
substance abuse and mental health disorders and histories of trauma. Journal of
Behavioral Health Services and Research, 32(2), 227-241.

Martin, J. (1992). Cultures in Organizations: Three Perspectives. New York: Oxford
University Press.

Martin, S. E. (1980). Breaking and Entering: Police Women on Patrol. Berkeley:
University of California Press.

Mattessich, P., Murray-Close, M., & Monsey, B. (2001). Collaboration: What makes it
work (2nd ed.). Saint Paul: Amherst Wilder Foundation.

McDonald, M. B. (2003). Social capital in a community collaborative network
Mcdonald,-Mary-B: Michigan State U., US.

McLaughlin, J. A., & Covert, R. A. (1984). Evaluating interagency collaborations.
Chapel Hill: Technical Assistance Development System.

Mizrahi, T., & Rosenthal, B. B. (1992). Managing dynamic tensions in social change
coalitions. In J. Morrison (Ed.), Community Organization and Social

Administration: Advances, Trends, and Emerging Principles. New York:
Haworth.

Morgan, J. L., Guetzloe, E. C., & Swan, W. W. (1991). Leadership for Local Interagency
Coordinating Councils. Journal of Early Intervention, 15(3), 255-267.

Mulroy, E. (1997). Building a neighborhood network: lnterorganizational collaboration to
prevent child abuse and neglect. Social Work, 42(3), 255-264.

Nahapiet, J ., & Ghoshal, S. (1998). Social capital, intellectual capital, and the
organizational advantage. The Academy of Management Review, 23(2), 242-266.

180

Neilsen, R. (1988). Cooperative strategy. Strategic Management Journal, 9, 475-492.

Nissen, L., Merrigan, D., & Kraft, M. (2005). Moving mountains together: Strategic
community leadership and systems change. Child Welfare, 84(2), 123-141.

O'Looney, J. (1994). Modeling collaboration and social services integration: A single
states experience with developmental and non-developmental models.
Administration in Social Work, 18(1), 61-86.

Parks, C. D., & Hulbert, L. G. (1995). High and low trusters' responses to fear in payoff
matrix. Journal of Conflict Resolution, 39, 718-730.

Pence, E. (1999). Some thoughts on philosophy. In E. L. Pence (Ed.), Coordinating
community responses to domestic violence: Lessons from Duluth and beyond.
Thousand Oaks: Sage.

Pence, E., & Gallaway, S. (1985). Criminal justice response to domestic assault cases: A
guide for policy development: Report prepared by the Domestic Abuse
Intervention Project.

Pence, E., & Shepard, M. (1999). Developing a coordinated community response. In E.
Pence (Ed.), Coordinated Community Responses to Domestic Violence: Lessons
fiom Duluth and Beyond (pp. 3-24). Thousand Oaks: Sage.

Portes, A. (1998). Social capital: Its origins and applications in modern sociology. Annual
Review of Sociology, 24, 1-24.

Post, L. A., Klevens, J., Maxwell, C. D., Shelley, G., & Ingram, E. (2004). The impact of
coordinated community responses on communities' attitudes and rates of intimate
partner violence. Manuscript submitted for publication.

Powell, W. W., Koput, K. W., & Smith-Doer, L. (1996). lnterorganizational collaboration
and the locus of innovation: Networks of learning in biotechnology.
Administrative Science Quarterly, 41 (1 16-145).

Provan, K., & Milward, H. (2001). Do networks really work? A framework for
evaluating public-sector organizational networks. Public Administration Review,
61(4), 414-423.

Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community.
New York, NY, US: Touchstone Books/Simon and Schuster, Inc.

Raudenbush, S. W., & Liu, X. (2000). Statistical power and optimal design for multisite
randomized trials. Psychological Methods, 5(2), 199-213.

181

Reagans, R., & Zuckerman, E. W. (2001). Networks, diversity, and productivity: The
social capital of corporate R&D teams. Organization Science, 12(4), 502-517.

Reilly, T. (2001). Collaboration in action: An uncertain process. Administration in Social
Work, 25(1), 53-74.

Riessman, F. (1990). Restructuring help: A human service paradigm for the 19905.
American Journal of Community Psychology, 18(2), 221-230.

Rivard, J. C., Johnsen, M. C., Morrissey, J. P., & Starrett, B. E. (1999). The dynamics of
interagency collaboration: How linkages develop for child welfare and juvenile

justice sectors in a system of care demonstration. Journal of Social Service
Research, 25(3), 61-82.

Rosenkoetter, S. E., Shotts, C. K., Streufert, C. A., Rosenkoetter, L. 1., Campbell, M., &
Torrez, J. (1995). Local interagecy coordinating councils as infrastructure for
early intervention: One state's implementation. Topics in Early Childhood Special
Education, 15(3), 264-280.

Roussos, S. T., & F awcett, S. B. (2000). A review of collaborative partnerships as a

strategy for improving community health. Annual Review of Public Health, 21,
369-402.

Schuller, T., Baron, S., & Field, J. (2001). Social capital: A review and critique. In T.
Schuller (Ed.), Social Capital: Critical Perspectives. Oxford: Oxford University
Press.

Shan, W., Walker, G., & Kogut, B. (1994). Interfirm cooperation and startup innovation
in the biotechnology industry. Strategic Management Journal, 15(5), 387-3 94.

Shepard, M. F ., Falk, D. R., & Elliott, B. A. (2002). Enhancing Coordinated Community
Responses to Reduce Recidivism in Cases of Domestic Violence. Journal of
Interpersonal Violence, 17(5), 551-569.

Stevenson, J. F., & Mitchell, R. E. (2003). Community-Level Collaboration for
Substance Abuse Prevention. Journal of Primary Prevention, 23(3), 371-404.

Sullivan, C., & Allen, N. (2001). Evaluating coordinated community responses for
abused women and their children. In J. L. Edleson (Ed.), Domestic Violence in the
Lives of Children: The Future of Research, Intervention, and Social Policy (pp.
269-282). Washington, D.C.: American Psychological Association.

Tabachnick, B., & Fidell, L. (2001). Using Multivariate Statistics (4th Edition ed.).
Boston: Allyn & Bacon.

182

Talmud, I. (1999). Corporate social capital and liability: A conditional approach to three
consequences of corporate social structure. In R. Eccles (Ed.), Corporate Social
Capital and Liability. Boston: Kluwer Academic Publishers.

Tapper, D., Klienman, P., & Nakashian, M. (1997). An interagency collaboration strategy
for linking schools with social and criminal justice services. Social Work in
Education, 19(3), 176-188.

Tjaden, P., & Thoennes, N. (2000). Extent, Nature, and Consequences of Intimate
Partner Violence (Research report NCJ 181867). Washington, D.C.: National
Institute for Justice Center for Disease Control.

Trickett, E. J ., Barone, C., & Watts, R. (2001). Contextual influences in mental health
consultation: Toward an ecological perspective on radiating change. In E.
Seidman (Ed.), Handbook of Community Psychology (pp. 303-330). New York:
Kluwer.

Tsai, W. (2000). Social capital, strategic relatedness and the formation of
intraorganizational linkages. Strategic Management Journal, 21(9), 925-940.

Tsai, W., & Ghoshal, S. (1998). Social capital and value creation: The role of intrafirm
networks. The Academy of Management Journal, 41(4), 464-476.

Tsui, A. S., Egan, T. D., & O'Reilly, C. (1992). Being different: Relational demography
and organizational attachment. Administrative Science Quarterly, 3 7, 549-579.

Wasserman, S., & Faust, K. (1994). Social network analysis: Methods and applications.
New York, NY, US: Cambridge University Press.

Webber, S. S., & Donahue, L. M. (2001). Impact of highly and less job-related diversity
on work group cohesion and performance: A meta-analysis. Journal of
Management, 27, 141-162.

Williams, K. Y., & O'Reilly, C. (1998). Demography and diversity in organizations: A
review of 40 years of research. Research in Organizational Behavior, 20, 77-140.

Wittenbaum, G. M. (2000). The bias toward discussing shared information: Why are
high-status group members immune? Communication Research, 27(3), 379-401.

Wittenbaum, G. M., Hubbell, A. P., & Zuckerman, C. (1999). Mutual enhancement:
Toward an understanding of the collective preference for shared information.
Journal of Personality and Social Psychology, 77(5), 967-978.

Wittenbaum, G. M., & Park, E. S. (2001). The collective preference for shared
information. Current Directions in Psychological Science, 10(2), 70-73.

183

Wolff, T. (2001). The future of community coalition building. American Journal of
Community Psychology, 29(2), 263-268.

Yoder, J. D. (1991). Rethinking tokenism: Looking beyond numbers. Gender and
Society, 5(2), 178-192.

184

   
  

\lijlljjlgujjllgjjm

9

      

will;