Citation
Network approaches to wicked problems

Material Information

Title:
Network approaches to wicked problems
Creator:
Miller, Annie Blythe ( author )
Language:
English
Physical Description:
1 electronic file (119 pages) : ;

Thesis/Dissertation Information

Degree:
Doctorate ( Doctor of philosophy)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
School of Public Affairs, CU Denver
Degree Disciplines:
Public affairs
Committee Chair:
Teske, Paul
Committee Members:
Dodge, Mary
Alejano-Steele, AJ
Ronquillo, John

Subjects

Subjects / Keywords:
Human trafficking ( lcsh )
Political planning ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Wicked problems are policy and social issues where policy makers and citizens agree that action must be taken but the solutions are contested. Wicked problems are often cross jurisdictional, connected to moral and ethical public policy debates, and appear to require collaborative efforts to solve. This dissertation is a community-based project that supports efforts by the Laboratory to Combat Human Trafficking (LCHT) to reach their mission by analyzing how networks (equivalent to partnerships conceptually) act to achieve effectiveness together. This research explores comparisons across 186 networks attempting to solve the same wicked problem, which may give researchers leverage to effectively understand many of the complexities of public management organizational arrangements. The first chapter provides historical analysis of the network and wicked problem literatures and the following chapters address different dimensions of networks and network effectiveness. The second chapter examines network membership and describes which organizations may need networks to accomplish their mission. Chapter 3 focuses on network effectiveness by testing a model proposed by Turrini et al. (2010), which suggests that structural, contextual, and network determinants lead to network effectiveness. The final chapter utilizes qualitative techniques to summarize and contextualize the network effectiveness findings.
Thesis:
Thesis (Ph.D.)--University of Colorado Denver
Bibliography:
Includes bibliographical references.
System Details:
System requirements: Adobe Reader.
Restriction:
Embargo ended 05/15/2019
Statement of Responsibility:
by Annie Blyther Miller.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
on10030 ( NOTIS )
1003045755 ( OCLC )
on1003045755

Downloads

This item has the following downloads:


Full Text
NETWORK APPROACHES TO WICKED PROBLEMS
by
ANNIE BLYTHE MILLER B.S., Colorado State University, 2004 M.S., Miami University, 2007 M.A., University of Colorado Boulder, 2010
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Public Affairs Program
2017


©2017
ANNIE BLYTHE MILLER
ALL RIGHTS RESERVED


This thesis for the Doctor of Philosophy degree by Annie Blythe Miller has been approved for the Public Affairs Program by
Paul Teske, Chair Mary Dodge, Advisor AJ Alejano-Steele John Ronquillo
Date: May 13, 2017
m


Miller, Annie Blythe (Ph.D, Public Affairs Program)
Network Approaches to Wicked Problems Thesis directed by Professor Mary Dodge
ABSTRACT
Wicked problems are policy and social issues where policy makers and citizens agree that action must be taken but the solutions are contested. Wicked problems are often cross jurisdictional, connected to moral and ethical public policy debates, and appear to require collaborative efforts to solve. This dissertation is a community-based project that supports efforts by the Laboratory to Combat Human Trafficking (LCHT) to reach their mission by analyzing how networks (equivalent to partnerships conceptually) act to achieve effectiveness together. This research explores comparisons across 186 networks attempting to solve the same wicked problem, which may give researchers leverage to effectively understand many of the complexities of public management organizational arrangements.
The first chapter provides historical analysis of the network and wicked problem literatures and the following chapters address different dimensions of networks and network effectiveness. The second chapter examines network membership and describes which organizations may need networks to accomplish their mission. Chapter 3 focuses on network effectiveness by testing a model proposed by Turrini et al. (2010), which suggests that structural, contextual, and network determinants lead to network effectiveness. The final chapter utilizes qualitative techniques to summarize and contextualize the network effectiveness findings.
The form and content of this abstract are approved. I recommend its publication.
Approved: Mary Dodge
IV


TABLE OF CONTENTS
INTRODUCTION..................................................................
Network Approaches to Wicked Problems......................................
Human Trafficking: International Attention and U.S. Intervention...........
Partnerships and Networks..................................................
Networks and Effectiveness.................................................
Perceptions of Networks....................................................
Research Orientation.......................................................
Partnership in Practice....................................................
Data Collection and Research Design........................................
Limitations and Considerations.............................................
Organization and Structure.................................................
Conclusions and Findings...................................................
JOINERS: A QCA ANALYSIS OF NECESSARY AND SUFFICIENT CONDITIONS
LEADING TO ORGANIZATIONS JOINING NETWORKS.....................................
Abstract...................................................................
Introduction...............................................................
Theory.....................................................................
Data and Measures..........................................................
Methodology................................................................
Calibration..............................................................
Publicness...............................................................
Funding..................................................................
.. 1
.. 3
.. 6
10
12
15
17
18
20
23
24
26
27
27
27
30
33
35
36
37
39
v


Coproduction..............................................................40
Theoretical Model.........................................................42
Analysis.....................................................................43
Necessary and Sufficient Conditions in Combatting Human Trafficking Networks.43
Sufficient Conditions.....................................................45
Limited Diversity.........................................................47
Missingness...............................................................48
Conclusion...................................................................50
NETWORK EFFECTIVENESS: COLLABORATING TO COMBAT HUMAN
TRAFFICKING....................................................................52
Abstract.....................................................................52
Introduction.................................................................52
Wicked Problems and Human Trafficking.....................................54
Whole Networks............................................................57
Theory.......................................................................59
Methodology..................................................................60
Factor Analysis...........................................................61
Results......................................................................66
Network Effectiveness.....................................................69
Conclusion...................................................................70
PERCEPTIONS OF WHOLE NETWORKS: ATTEMPTS TO END WICKED
PROBLEMS.......................................................................73
Abstract.....................................................................73
vi


Practitioner Points
73
Introduction.....................................................................74
Partnerships and Networks........................................................76
Theory...........................................................................78
4Ps: Prevention, Protection, Prosecution, and Partnership.....................78
Data.............................................................................79
Methodology......................................................................80
Results..........................................................................80
State and National Differences................................................81
Prevention, Protection, Prosecution, and Partnership (the 4Ps)................85
Network Formation and Sustainability..........................................85
Trust and Conflict............................................................87
Tensions and Negotiations.....................................................89
Discussion.......................................................................91
Conclusion.......................................................................92
REFERENCES.......................................................................95
APPENDIX A: Terms...............................................................106
APPENDIX B: LCHT Survey Description.............................................108
APPENDIX C: Factor Analysis Robustness Details..................................109
Appendix C.l: Principle Factor Graphs........................................109
vii


LIST OF TABLES
TABLE
1.2: Determinants of Whole Network Effectiveness....................................14
1.3: Data Collection and Analysis Phases............................................21
2.1: Organizational Conditions Influencing Network Membership.......................42
2.2: Theory, Coproduction, Publicness and Funding Truth Table.......................46
2.3: Theory, Coproduction, and Publicness Truth Table...............................49
3.1: Hypothesized Factor Loadings...................................................63
3.2: Iterated Principle Factors Analysis............................................67
3.3: Iterated Principle Factor Items................................................68
4.1: Conflict in Network Arrangements...............................................88
viii


LIST OF FIGURES
FIGURE
3.1: Turrini et al. (2010) Network Effectiveness Framework Overview


CHAPTER 1
INTRODUCTION
Problem, Purpose, and Key Questions
Networks often are implemented as a way to leverage multiple agencies or organizations to achieve greater success than a single entity. Networks, coalitions, and partnerships are not a particularly new phenomenon in the social sector. While the field of networks continues to receive attention, Provan and Milward (2001) called for greater attention on the study of network level outcomes. At its broadest level, this research is framed by the following research question: What factors contribute to network level effectiveness?
First, out of concern for selection effects when studying networks, the research examines organizations that opt into networks and compares them to organizations that do not. Selection effects may guide positive or associative gains made by networks and this truncation may be leading researchers and practitioners to make inappropriate policy or resource allocation decisions in public service goods provision. The research focus in Chapter 2 is: What are the necessary and sufficient conditions leading to network membership? Additionally, do organizations that fail to join networks have structural advantages that rarely require network membership?
After determining the factors that contribute to network membership, Chapter 3 explores determinants of network effectiveness. Conceptually network effectiveness is the ability of networks to achieve goals and make an impact on a social, political or cultural issue. Network effectiveness, however, has yet to be operationalized and while theory exists to suggest which factors may contribute to effectiveness, no specific construct or measure for
1


network effectiveness exists. Some of the hypothesized indicators of network effectiveness include trust, cooperation, goal achievement, stability, effective communication, conflict resolution, and network duration. Applying the unified framework for network effectiveness proposed by Turrini et al. (2010), the chapter establishes two factors that appear to capture network effectiveness and network dysfunction.
In order to reach a nuanced understanding of network effectiveness, Chapter 4 concludes this dissertation with a qualitative exploration of the survey data. Nearly every question in the network battery utilized in this research included comment boxes, many with rich and deeply informative narratives about the state of the network. One respondent, for example, had this to say when asked about the sustainability of the partnership: “Goals set by members’ individual agencies often determine/derail their goals in the partnership.” Another respondent reported that the network allows them “to harmonize state and federal law enforcement, nonprofits, city services, social services, and the community in a holistic antitrafficking and victim assistance effort.” The final chapter is a thematic and content analysis that provides rich description supporting the empirical findings of previous chapters.
Exploring network function and dysfunction, operationalized as effectiveness, contributes to the empirical study in the public management and network literatures. This dissertation represents more than scholarly contributions. The findings here have an immediate and direct impact on the work done by the Laboratory to Combat Human Trafficking. Additionally, the contribution goes beyond the human trafficking awareness movement; trafficking is one of many wicked problems that plague communities across the globe. Resolving conflict actively in partnerships, seeking out incentives for membership in network for organizations that may not perform activities related to prevention or protection
2


of survivors, and establishing rules and norms within networks that contribute to the identified indictors of effectiveness will help improve the functioning of many types of partnerships. This chapter provides an introduction to wicked problems, network determinants, network effectiveness as well as the specific issue of human trafficking.
Network Approaches to Wicked Problems
Wicked problems are generally considered policy and social issues where policy makers and citizens agree that action must be taken but the solutions are contested. Citizens may disagree about the root causes or identification, and the problem persists after numerous attempts to combat the issue. Human trafficking is one example of a contemporary wicked problem with global implications. As borders become more porous, information about the quality of life in other places is easily shared, and the pace of economic globalization grows, so does the demand for labor, personal safety, or freedom. Trafficking in humans increases in conflict areas and vulnerable populations are especially at risk. Trafficking spans various types of crimes and impacts adult men as well as women and children. Too often, bureaucracies may oversimplify the problem as a sex crime only affecting youth and children. Head (2008) suggested this simplification is a direct result of the growth in needing to establish targets and goals that can be met by bureaucrats and managers in the public sector. Simplification allows for quantification of some aspects of this crime, but may lead to negative externalities for other potential victims and survivors of trafficking. Wicked problems like human trafficking appear intractable because they are complex, require participation from multiple sectors, and solutions vary among communities.
Human trafficking is particularly problematic because combating it requires solutions, collaborations, and partnership across many levels of governance and jurisdictions. A truly
3


wicked problem, according to Roberts (2000), is one where “nothing really bounds the problem solving process - it is experienced as ambiguous, fluid, complex, political, and frustrating as hell” (p. 2). Current efforts to end trafficking, particularly those funded by the United States government and the United Nations, have political will to end child sex trafficking. Labor trafficking, border smuggling, drug running, and other forms of coercion, however, while less flashy, also require time, attention, funding, and political will to achieve reduction and elimination of the problems. Trafficking prevalence shifts geographically as conflict situations, black markets, and labor needs change globally. When a single municipality, county, state, or country decreases trafficking, the crime may easily shift into jurisdictions that are less vigilant just as a markets shift with changes in supply and demand. This situation further complicates problem identification, jurisdictional authority, and potential solutions. Ultimately, human trafficking is both wicked and frustrating. Many communities recognize the need to combat this problem, to provide attention and energy to the root causes by finding ways to prevent the crime, protect survivors, and prosecute the offense. Networks are arrangements that allow organizations and individuals from different sectors, problem orientations, and service provision areas to come together and work to end wicked problems from multiple angles.
Wicked problems are plagued with two types of uncertainty, cognitive and strategic, that lead agencies to take different approaches to potential solutions (van Bueren, Klijn, & Koppenjan, 2003). Cognitive uncertainties include a lack of scientific understanding about the root causes of a problem or an inability to identify the causal relationships leading to the problem. Strategic uncertainty occurs when there is disagreement on the solution of a wicked problem. Strategic uncertainty, particularly in networks, occurs because different actors may
4


opt to take different approaches to problem resolution, have high transaction costs for information sharing, and disagree about the root causes as a result of cognitive uncertainty. In the trafficking movement, for example, many agencies utilize a harm reduction model when confronting the issue. If an agency can reduce an individual’s vulnerabilities by providing housing, workforce training, health care, food, or other recourses, the supply of available humans vulnerable to trafficking will be reduced. This supply-side solution contrasts with demand-side attempts to end trafficking. Agencies working on the demand-side seek to increase penalties for johns, create educational programs aimed at reshaping the thought patterns that lead pimps and johns to abuse people, and to shame individuals who use trafficked labor. It may appear unproblematic that agencies take each of these perspectives and work from both directions; however, one of the fundamental reasons that organizations adopt a network approach to solving wicked problems is to share scarce resources. The gains from the network can be lost as a result of the cognitive and strategic uncertainty.
Networks often are identified as managerial and social arrangements that may help end some forms of wicked problems (van Bueren, Klijn, & Koppenjan. 2003). Trafficking, which is considered a form of slavery, occurs locally but ending this servitude requires collaborative, multi-level, multi-jurisdictional efforts. Networks, beyond markets and hierarchies, are an organizational form that can allow agencies to share scarce resources.
Also, organizational networks can support employees and individuals to collaborate more effectively, may foster learning and knowledge, and provide efficiencies through interdependencies in supply chains (Podolny & Page, 1998; Powell, 2003). Network forms of organization have grown in prominence because they more effectively reflect empirical observations of the world. Borgatti and Foster (2003) described the prominence of networks
5


as “part of a general shift, beginning in the second half of the 20th century, away from individualist, essentialist and atomistic explanations toward more relational, contextual and systemic understandings” (p. 991). Organizations across the globe must learn how to work together, share resources and knowledge, and realize their connectedness across jurisdictional boundaries. The next section provides context and history to the wicked problem of human trafficking. The remainder of the chapter discusses the dependent variable under consideration in this study, network effectiveness. Additionally, the chapter examines perceptions of networks and how they are utilized; focuses on the research questions attended to throughout the dissertation; describes the research and methodological approach; and introduces the remaining chapters.
Human Trafficking: International Attention and U.S. Intervention
Fighting human trafficking, as a modern social movement, began 15 years ago, but as awareness grows, it is clear this is not a new phenomenon. The United Nations recognized the need for the protection of individuals from force, fraud, or coercion many decades earlier. The Declaration of Human Rights (UN General Assembly, 1948), in articles three, four, and five, state that individuals should retain the rights to personal liberties, the right to life, freedom from slavery, security of their own person, and the ability to live free from inhumane, cruel, or tortuous treatment. Trafficking is broadly considered an act where individuals are forced into some action, fraudulently led to behave or act in a particularly way, or coerced by another individual (Alejano-Steele, 2013; Trafficking Victims Protection Act of 2000). Hart (2009) suggested that human trafficking is the modern term for any form of slavery. Many people think that human trafficking is primarily sex trafficking or coercion into sex acts; however, trafficking is not limited to prostitution or sex crimes and the victims
6


of this crime are not limited to children or women. According to Alejano-Steele (2013), “although greater attention has been given to human trafficking as a global human rights violation that exploits women as commodities, there continues to be a lag in applying these lenses domestically to our sisters and daughters in the United States” (p. 148). Many of the challenges associated with focusing on the issue domestically connect directly to the complex web of national legislation and jurisdictional authorities engaged in combating trafficking. The complexity of human trafficking as a crime warrants the involvement of multiple agencies, and highlights why network effectiveness and the ability of agencies to collaborate is the means to ending wicked social problems like trafficking.
The Office to Monitor and Combat Trafficking in Persons, part of the U.S. Department of State, identifies several forms of servitude as trafficking including: debt bondage, forced labor, sex trafficking, bonded labor, involuntary domestic servitude, child soldiers, forced child labor, and child sex trafficking. Early discourses related to trafficking of persons tended to focus primarily on trafficking children and women for sex (Gallagher, 2010). The United States Congress, through passing the Trafficking Victims Protection Act of 2000, expanded the definition of trafficking to include labor, migrant trafficking, and noted that while trafficking often crosses boarders, the problem must be addressed both internationally and domestically.
The laws aimed at combating human trafficking in the United States are a complex mix of Acts and Reauthorizations folded into several different pieces of legislation. Table 1.1 provides an overview of each piece of legislation and the major contribution made through adoption. In the United States, trafficking protections, in their most current form are a legal component of the Violence Against Women Reauthorization Act of 2013 (VAWA,
7


Public Law 113-4) and The Trafficking Victims Protection Reauthorization Act, which passed as an attachment to the 2013 VAWA reauthorization. This legislation established protections such as temporary U-Visas for survivors of trafficking in the United States. This legislation as the title suggests may not ultimately protect male labor and sex trafficking victims in the same capacity as protections offered for women and children.
______________Table 1.1: United States Legislation to Combat Human Trafficking________________
Legislation____________Year
Action
The Mann Act 1910, Reauthorized 1978, 1986 Criminalizes sex trafficking especially as it relates to crossing state or international lines.

The Tariff Act 1930 Goods cannot be imported if indentured laborers, forced laborers or slaves, made them.

Racketeer Influenced and Corrupt Organizations Act (RICO) 1961 RICO gave law enforcement agencies a way to prosecute organized crimes especially groups organized to force individuals to participate in felonious activities or fraudulent practices.

Trafficking Victims Protection Act 2000 The first comprehensive legislation on human trafficking that designated human trafficking as a federal offense. Established the first 3 Ps -prevention, protection and prosecution and the Office to Monitor and Combat Trafficking in Persons (State Department). Created the T-visa, a legal status that allows trafficking survivors to remain in the U.S. legally.

Prosecutorial Remedies and Other Tools To End the Exploitation of Children Today (PROTECT) Act 2003 Created enhanced penalties for individuals engaged in tourism practices to participate in sexual acts with children. Also established the Amber Alert System to find lost, runaway, and kidnapped children.

8


Table 1.1: United States Legislation to Combat Human Trafficking (Cont.)
U.S. Leadership on HIV/AIDS, Tuberculosis, and Malaria Act 2003 Restricts aid and international funding for countries and organizations that do not have specific antitrafficking policies.

Trafficking Victims Protection Reauthorization Act 2005 Authorizes additional grant aid to combat trafficking and establishes a pilot program to shelter minors. Grant aid provided both domestically and internationally.

Trafficking Victims Protection Reauthorization Act 2008 Expands the federal jurisdiction to prosecute trafficking and expands the definition of the crime of sex trafficking. Authorizes DOJ to create a model for state legislation.

Customs and Facilitations and Trade Enforcement Act 2009 Prohibits the import of goods made by trafficked or indentured individuals.

Trafficking Victims Protection Reauthorization Act 2013 Funds anti-trafficking efforts until 2017 as an amendment to the reauthorized VAWA. Strengthens collaboration with state and local law enforcement. Seeks to prevent child marriage.

National Defense Authorization Act 2013 Prevent and limit government contracts with agencies in foreign countries that may engage in trafficking.

*Compiled on August 22nd, 2015 based on analysis from Polaris Project and the Office to Combat and Monitor Trafficking in Persons.
Women and children are most vulnerable to sex and labor trafficking and represent approximately 54% of victims internationally (International Labour Organization, 2015); however, 46% of victims are male and may not receive appropriate interventions, support, or resources. Additionally, the Reauthorization Act specifically addresses international trafficking apart from trafficking in the United States. International trafficking efforts are
9


centralized through the U.S. Department of State while the Department of Justice, Department of Health and Human Services, and Department of Labor distribute grants and oversee activities related to trafficking domestically. Many agencies and many agendas intersect to combat trafficking in the United States.
Partnership formation is specifically encouraged as part of the 2013 VAWA
reauthorization. The legislation specifically states:
In coordination and cooperation with other officials at the Department of State, officials at the Department of Labor, and other relevant officials of the United States Government, shall promote, build, and sustain partnerships between the United States Government and private entities, including foundations, universities, corporations, community-based organizations, and other nongovernmental organizations, to ensure that: (1) United States citizens do not use any item, product, or material produced or extracted with the use and labor from victims of severe forms of trafficking; and (2) such entities do not contribute to trafficking in persons involving sexual exploitation.
The Trafficking Victims Act of 2000 was amended in the VAWA-R 2013 Act to
reflect the need that partnerships, especially across the regional bureaus within the State
Department set goals for decreasing trafficking in partnership with international government
partners and that the bureaus create metrics for assessing the ability of organizations to meet
those goals. This dissertation addresses the effectiveness of partnerships across the United
States and the State of Colorado to determine which characteristics, contexts, and functions
lead to positive outcomes toward ending trafficking.
Partnerships and Networks
The language of partnerships and networks is interchangeable in this study. Many terms in the public administration and management literature appears to describe the network concept under consideration here. One indictment of the broadly and quickly expanding literature on networks is that the term ‘networks’ is increasingly nebulous, represents a
10


different concept to different people, and is inconsistently applied (Borgatti & Foster, 2003). This section identifies the author’s conceptual definition of networks and connects that concept to the operationalization of the concept in the survey from which data for this project were drawn.
The notion that several organizations or agencies work together toward achieving a common goal is the heart of understanding these network arrangements. O’Toole and Meier (2004) argued that networks are an interdependent group of multiple organizations or parts of organizations whereby the connections among the actors are not arranged as superior to each other. This definition suggests that members of the network are not subordinate or report to one another. Provan and Kenis (2008) specifically defined networks as “groups of three or more legally autonomous organizations that work together to achieve not only their own goals but also a collective goal” (p 231). Provan and Kenis (2008) also argued that network configurations often require lead organizations or network administrators to provide an effective governance mechanism for positive outcomes. The structure of networks, leadership of the network, and goal-directed behavior are all considered here. The Laboratory to Combat Human Trafficking, the community partner connected to this project, defines ‘partnership’ research as:
Partnership measures acknowledge that combating human trafficking requires a comprehensive response through the cooperation of multiple sectors.
Partnerships bring together diverse experiences, amplify messages, and leverage resources. For the purpose of this survey, an anti-human trafficking partnership refers to a ... [e.g., a cooperative relationship between two or more organizations established for the purpose of jointly combating human trafficking in some way.]
This language, while utilized specifically for the survey tool, suggests ways in which a group of organizations may join together to achieve their stated goals.
11


More specifically, this study examines various “whole” public management networks. Public management networks can be identified as “agencies involved in public policy making and/or administrative structure through which public goods and services may be planned, designed, produced, and delivered (and any or all of the activities)” (McGuire & Agranoff, 2007, p. 1). Agranoff and McGuire (2001) first called for focused and continued efforts to study public management networks. This need for research was heralded by many scholars and the field has arguably developed through three traditions: that of social network analysis in sociology, policy networks in political science, and service provision networks in public management (Berry et al., 2004). Berry et al. (2004) find contributions across all three traditions and suggest that overall structure matters when it comes to understanding networks, an individuals’ place and the centrality of an organization also matters within the network, and these structures and positions can influence the effectiveness of the network itself (see also, Agranoff & McGuire, 2001).
This dissertation examines which organizations may opt into and out of networks combating human trafficking, identifies structural, contextual, and network determinants of effectiveness and ineffectiveness. Additionally, the research explores comparisons across multiple networks attempting to solve the same wicked problem, which may give researchers leverage to more effectively understand many of the complexities of public management arrangements. This study focuses on the public administration and management network tradition and contributes to the past and current literature by employing an empirical study of whole networks that provide public goods to prevent human trafficking, protect survivors, and prosecute perpetrators.
Networks and Effectiveness
12


Networks form when an organization can gain from sharing resources (e.g., information and staff). Sharing resources means reaching new populations, providing resources to clients more effectively or more efficiently, partnering to expand services, or acquiring additional funds through mandated collaboration. Scholars agree that complex arrangements among individuals, organizations, and institutions complicate governing (Ansell & Gash, 2008; Provan & Kenis, 2008; Provan & Milward, 2001). Contributions that build network theory may provide greater insight into clarifying, understanding, and managing the complexity that arises because of networked collaborations and governance. Klijn (2005) specifically suggested that the network perspective “tries to develop concepts (networks, games, perceptions, etc.) and methods of analyzing (such as formal network analysis, explicating network rules and frames of reference) in order to clarify the complexity of the interaction processes and their outcomes (p. 266).” Berry et al. (2004) argued “the conceptual frameworks and key terms employed across the literature have created a complex and often confusing picture (p. 539).” Following Berry and colleagues’ suggestion that the field of study comes from advances in social network analysis methodology in sociology, studies in political science of policy networks, and examinations of public service provision in public management, this study deepens the literature in the third vein. This dissertation examines network concepts like structure, function, and context to understand some of the complexity inherent to managing effective public service networks.
Public service provision networks continue to expand. Question on the utility, purpose, and necessity of effectively understanding networks are fully recognized as essential. While, scholars continue to agree that while studying networks matters, there is still uncertainty about the functioning and effectiveness of whole networks (McGuire &
13


Agranoff, 2007; Provan & Lemaire, 2012; Raab, Mannak, & Cambre, 2015; Turrini et al. 2010). Scholars attribute this conceptual gap to a dearth of studies examining “whole” networks (Kilduff & Tsai, 2003; O’Toole, 2015; Provan, Fish, & Sydow, 2007; Provan & Lemaire, 2012).
Table 1.2: Determinants of Whole Network Effectiveness
Contextual
Structural Characteristics________Characteristics________________Network Functioning
Composition: Prevent, protect and prosecute External to Network Managerial work

Sector: Government, Nonprofit, Private Grant required partnership Training Members

Size Resource Munificence Comprehensive Service Provision

Formalized Mission or Vision Statement

Voice: Includes survivor/ rep from vulnerable pop Goals

Local: Includes local community member Perception of network leader

Centralization

Broker / leader
Provan and Lemaire (2012) argued, “whole network research shifts the focus from the ties that an actor has (an egocentric micro approach) and focuses instead on all of the ties among
14


a set of actors, which is a macro approach” (p. 639). The macro approach applied here looks at both national and state-level networks that seek to end human trafficking.
Of the limited studies examining whole networks, the definition of network effectiveness varies (Turrini et al., 2010). Three types of effectiveness: client level, community level, and network level were identified by Provan and Milward (2001) as indicators of network success (Turrini et al., 2010). Network level effectiveness refers to the sustainability, legitimacy, and maintenance of the network. Ferlie and Pettigrew (1996) examined a single network for duration and effectiveness and several other studies have examined the ability of networks to achieve stated goals, innovate, and remain sustainable, but empirical work has failed to address all of these types of effectiveness along with indicators of ineffectiveness in whole networks. This dissertation builds the network effectiveness literature by applying network concepts like structure, context, and function drawn from Turrini et al.’s (2010) scholarship to determine the factors that lead to effectiveness and ineffectiveness in whole networks (Provan & Milward, 1995; Turrini et al., 2010). Examining the composition and capacity of these networks may lead to understanding necessary and sufficient conditions for network-level effectiveness as well as enhance understanding of how to create more effective network structures to overcome wicked social problems.
Perceptions of Networks
The increasing attention on the network organizing structure assumes that networks provide gains in service provision through collaboration (Provan & Kenis, 2008) and that networks form through affinity, reputation, or similar goal orientation/attainment (Agranoff & McGuire, 2001; McPherson, Smith-Lovin, & Cook, 2001). Although, networks may
15


actually work against the altruistic aims of the organizations involved (O’Toole & Meier, 2004). Additionally, networks may be ineffective tools to provide services for trafficking and violence survivors as some evidence suggests that additional trauma and secondary victimization may occur as a result of reporting incidents to multiple providers (Campbell et al., 1999; Orth, 2002). Another indictment against networks is that overcoming the transaction costs related to knowledge and resource sharing is immensely difficult (Weber & Khademian, 2008). Network managers and organizational leaders that act as collaborative capacity builders may be able to overcome this particular challenge. Chapter 4 explores how determinants like leadership and sustainability of the network may contribute to ineffectiveness.
The networks under consideration in this study all seek to end human trafficking.
Each organization that joins the network has its own mission, funding structure, and action targets. These organizations may find that joining a network fails to contribute to their mission, may extend the organization beyond its capacity, or may increase administrative burdens on individuals who represent the organization. This dissertation takes two different approaches to understand what O’Toole and Meier (2004) refer to as “the dark side” of networks. First, the research compares organizations in the movement who participate in networks with those that do not. Perhaps, organizations that have substantial resources and funding, jurisdictional advantages like prosecutorial authority, and those already successful in providing victim advocacy in the local community may not need to leverage collaborations to effectively participate in the movement to end human trafficking.
Work prior to Grannovetter’s (1973) study on the power of weak ties in networks considered the negative implications of networks, however since the late 1970s, most
16


attention focuses on the neutral ties or positive gains made through networks (Labianca, 2014). While this study does not directly address the negative ties between organizations in a network, it does consider which determinants lead to conflict, competition, or derision at the whole network level. This approach is an especially important consideration as organizations are incentivized through federal and state funds to collaborate, share resources, and jointly produce public goods.
Research Orientation
Different conceptions of mixed methods approaches abound (Creswell & Clark,
2008; Tashakkori & Teddlie, 2010). Tashakkori and Teddlie (2010) provided perspective on an integrated framework and “meta-inferences.” This epistemology of mixed methods is the most coherent and relevant approach for understanding, describing, explaining, and predicting the success and failure of networks combating human trafficking. Truly mixed method research must be a sum greater than its qualitative and quantitative parts; more specifically, “such research can simultaneously address both exploratory and confirmatory questions, thereby gathering information that can result in ‘meta-inferences’ .. .that neither the quantitative nor qualitative perspectives could do alone” (Tashakkori & Teddlie, 2008, p. 101). The research design for this dissertation is an attempt to combine methodological approaches from various research traditions and to pay special attention to the consideration that through the partnership with LCHT this work also is intended to be action research (Lingard, Albert, & Levinson, 2008). The epistemological perspective of Guba and Lincoln (1989), who argued that starting first with research questions and then selecting appropriate methods, guides this thesis. The following section describes the methodology and research design used in this work. The mixed method approach utilized here combines qualitative
17


comparative analysis utilizing Boolean algebra, factor analysis, and content analysis to examine organizations’ choices for network membership and effectiveness and ineffectiveness of networks working to solve a wicked problem.
Partnership in Practice
The Laboratory to Combat Human Trafficking (LCHT), a lead organization of a network (Provan & Kenis, 2008) based in Colorado, graciously agreed to provide data from a national and statewide survey effort as well as other valuable support. Comparing network effectiveness was possible only through collaboration with their organization as they have the credibility necessary for collecting data on trafficking partnerships nationally and created the survey utilized here. As such, this work is a collaborative and community-based project. Without the participation, contributions, and support of LCHT, this study would not be possible. Many of the contributions from this paper are realized through engaged scholarship (Boyer, 1996; Van de Ven, 2007; Hollander & Saltmarsh, 2000). Engaged scholars seek to work jointly with community partners to achieve better understanding of the problems, assist in clarifying complexities, and create opportunities for social change. LCHT approached the researcher to assist in testing their theoretical model for combating trafficking.
The theoretical model developed by LCHT identifies four components of the movement necessary to end human trafficking: prevention, protection, prosecution, and partnership (4Ps). The U.S. Department of State’s Office to Monitor and Combat Trafficking in Person’s adopted LCHT’s 4Ps model, yet the model itself remains untested. The theory driving the hypotheses tested here is that the network composition of actors devoted to protection, prevention, and prosecution influence the ability of partnerships to achieve positive network - level outcomes (i.e., effectiveness). By testing the 4Ps model, this study
18


determines if organizing networks around these functions actually changes the effectiveness
of networks combating trafficking across the United States.
The data provided by LCHT for this project developed from a multi-year planning
and development process. The survey was designed to determine how many efforts are aimed
at combating human trafficking, the extent of the success of those efforts, and to identify
promising practices conducted in communities across Colorado and the United States.
According the institutional review board submission completed by Alejano-Steele (2012):
Subjects are law enforcement, prosecutors, government officials, service providers, and community members involved in the anti-trafficking movement throughout the United States. These subjects will be asked to take a survey to answer questions about the services available for people who have experienced human trafficking, the ways in which anti- human trafficking efforts are approached by the criminal justice system, the various prevention efforts for human trafficking and the various partnerships that exist in the anti-human trafficking movement. All subjects will be sent a survey electronically through their email in which they will receive an explanation of the study and an invitation to participate in the study. In the email, they will be directed to a Zoomerang link for subjects to click on to initiate the survey. Service providers from Colorado will additionally complete the Community Needs Assessment Survey, which will also be distributed electronically. All subjects will be directed to a link to the survey in the program Zoomerang that will contain an informed consent. The sample size for this study is 1500.
Subjects will be identified through literature reviews, website reviews of anti-human trafficking agencies, suggestions made by pre-eminent colleagues in the field and through recommendations from the Laboratory to Combat Human Trafficking’s Advisory Board for the Colorado Project. Subjects will be sent an email explaining the study and inviting them to participate two weeks prior to the survey being distributed. They will be directed to a link to the survey in Zoomerang that will contain an informed consent.
One hundred and eighty-six respondents completed the national survey and 132 respondents completed the Colorado state survey.
The survey included 247 items and several respondents reported feeling survey fatigue as a result of the length. The partnership questions were the final section of the survey, which may mean that respondents who felt fatigued, but may have failed to report in
19


the final comment box of the partnership section, were systematically less likely to respond to that section. How this systematic bias may impact findings is unclear, but 71% of respondents completed item 191: Is your organization involved in a partnership effort to combat human trafficking (formal or informal)? Of those respondents 59% indicated they did not belonging to a partnership. In the national survey, 66% of respondents made it to question 191 and of those respondents, 22% reported that they do not belong to a partnership. These descriptive results suggest that while many respondents failed to fully complete the survey, there does not appear to be evidence to suggest the respondents who did complete the network batter reflect a predisposition or selection bias to participate in networks or not participate.
Data Collection and Research Design
At times, research design and methodology is its own form of a wicked problem. The solutions are diverse, contested, and often depend on the epistemological perspective of the researcher (Brady & Collier, 2010; Gerring, 2011; King, Keohane, & Verba, 1994). The research design proposed here is iterative and occurs in several different phases in order to draw as much leverage as possible on making inferences. King, Keohane, and Verba (1995) argued that the distinctions between qualitative and quantitative methods are nearly unimportant; instead, they suggest that the logic of inference ought to apply to all social science studies. Instead of applying a single logic, this research design employs the logic of inference (King, Keohane, and Verba 1994) and the logic of inquiry (Gerring, 2011), to lead to sources of leverage on social science phenomena (Brady & Collier, 2010). In order to achieve the most robust findings, several phases of research are necessary. Table 1.3 identifies the multiple phases, describes the purpose of each, and identifies the method
20


utilized for that phase of the design. To achieve a mixed method research design, several phases of research were exploratory while others were confirmatory.
Table 1.3: Data Collection and Analysis Phases
Research Stage Goal Method

Phase 1: Specifying variables; Determining variance; Seeking additional data Create all independent and dependent variables; exploratory Exploratory factor Analysis (determine if the network quality battery in LCHT’s survey matches hypothesized constructs); Collect all necessary data files from LCHT; Assess the size,
sources if necessary quality, limitations, and strengths of data sources
Determine
Phase 2: Test exploratory factor analysis from national data on state-level data validity and reliability of Factor Analysis of Dependent Variable; confirmatory Confirmatory factor analysis of national data
Phase 3: Identify causal mechanisms Theory refinement; Exploratory Data Analysis (EDA) Qualitative Comparative Analysis (QCA) - specifically, fuzzy-set QCA for two different condition sets
Check for
Phase 4: Review potential specification problems, validity, accuracy, robustness and misspecification; confirmatory Kennedy’s (2010) Diagnostics and Ten Commandments of Applied Econometrics
Provide Content analysis of items
Phase 5: Compile all comments for qualitative analysis evidentiary support for findings; exploratory and confirmatory included as comment boxes in the partnership section of LCHT’s survey; Connect themes drawn from quantitative chapters to improve robustness and validity
Phase one of the proposed design is a factor analysis of the national survey. The researcher posits that by conducting testing and exploratory data analysis on the national survey data, the Colorado survey can serve as a confirmatory data set. King, Keohane, and
21


Verba (1994) suggested that theory development and testing should not occur on the same data so the methodology proposed here seeks to avoid this logic inference problem. In line with the dearth of literature on network ineffectiveness and the expansiveness of the literature on the organizational form of networks, a hypothesis was drawn for an exploratory factor analysis of one of the batteries included in the partnership section of the survey. Table 1.4 details the proposed factor loadings tested in phase 1.
_________________________Table 1.4: Hypothesized Factor Loadings_________________________
_____________Effectiveness______________________________________Dysfunction______________
There is a great deal of trust among members Conflict can arise among members because of the different agency/organizational missions
Member sometimes socialize together Conflict arises among members because of competing definitions of human trafficking
There is good communication among members There is competition among members that work within the same communities
The partnership will remain strong if the current leader leaves There are difficulties sharing information among members about victims of trafficking due to confidentiality policies
There is low turnover of members in the partnership
The members of the partnership are able to resolve conflict effectively
The partnership has achieved its annual goals over the past year
According to Dillon and Goldstein (1984), exploratory factor analysis is a way of determining if items load on similar factors. The intent of phase one factor analysis is to determine if two underlying, or latent, concepts are present. Explanatory factor analysis is
22


most commonly utilized when a researcher has no strong theoretical priors about how the data may perform around a particular number of factors (DeVellis, 2012; Dillion &
Goldstein, 1984). As such, exploratory analysis provides a helpful mechanism to move forward in determining if these items, in fact, represent the two hypothesized concepts of network efficiency and inefficiency.
After conducting the exploratory factor analysis, a confirmatory analysis on the Colorado survey was conducted. The factors did indeed load similarly across the two different sets of data. The final chapter, a robustness check on the thematic and theoretic chapters preceding, presents a content and thematic analysis of comments provided by respondents in the partnership section of the survey.
Limitations and Considerations
The study of networks, especially the proposed confirmatory analysis proposed here, has two important limitations. First, the Turrini et al. (2010) model of network effectiveness represents a large amalgamation of various types of indicators. The survey sent out to representatives of organizations in and out of networks does not contain measures or indicators of all the proposed determinants of network effectiveness in their suggested model. As such, several of the determinants are unexplored in this study. Future research in this area is necessary in order to fully test the Turrini et al. (2010) model.
An additional consideration is missing data. There is “missingness” present across both surveys and throughout the independent and dependent variables. Allison (2001) argued there are several ways to manage missing data including listwise and pairwise deletion, imputation, and dummy variable adjustment. In each of the following chapters, the missingness is handled in different ways that best address the more specific research question
23


under consideration and the selected methodology. In chapter 2, missingness is most appropriately handled by listwise deletion of cases. Since the Qualitative Comparative Analysis (QCA) requires the observation of combinations of conditions leading to network membership, imputation or other means of predicting values would be inappropriate. Logical remainders, the cases not empirically observed, are an important part of the QCA, which also makes imputation problematic. In chapter 3, missingness is also handled through casewise deletion. However, the missingness in this chapter is likely random as opposed to other chapters where datasets and variables are drawn from all four sections of the survey (where some consistent type of missing bias may be observed as respondents opt into and out of the additional sections of the survey). Chapter 3 utilizes a single battery in the fourth (partnership) section of the survey. In this chapter, 71 cases had full information and were utilized for the factor analysis. Chapter 4 is the qualitative chapter and the researchers opted to examine all comment boxes completed by respondents. If a respondent felt strongly enough to comment, the researchers were compelled to analyze those responses regardless of completion in other sections or on other items.
Organization and Structure
The format of the dissertation proposed here is slightly different from a traditionally structured dissertation. The “three-article” format is gaining popularity, especially as the expectation for publications rise in the tenure-track hiring market. This research follows that format by structuring each chapter as a stand-alone article ready for submission for peer review. This first chapter provides historical and contextual analysis of network literature, and the following chapters address different dimensions of partnerships. The second chapter examines network membership and describes which organizations may need networks to
24


accomplish their mission and goals and why other organizations may not. This chapter is dedicated to assisting LCHT in their efforts to test the 4P model. Chapter 3 focuses on network effectiveness by testing a model proposed by Turrini et al. (2010), which suggests that structural, contextual, and network determinants lead to network effectiveness. The intention of the concluding chapter is to enhance conceptual understanding of the networks through the respondents’ own words in a qualitative manner; this chapter effectively rounds out the preceding quantitative chapters into a mixed-method dissertation. All of the findings from each chapter support a broad interpretation of network approaches to wicked problems.
These network approaches include the consideration of coproduction, the idea that organizations must create outputs directly related to the goals the network seeks to achieve if they are to join a network in the first place. Previous research had not fully illuminated this condition and it may enhance practical work of agencies and network administrative organizations seeking to create effective networks. Second, dysfunction and effectiveness are unique and identifiable constructs at the network level. Effectiveness, comprised most significantly of trust, achieving shared goals, good communication, and resolving conflict, appears to drive the ability of network level outcomes. Additional qualitative support for the need for conflict resolution, as well as the need for funding and genuine commitment, in networks comes to the front in Chapter 4. Respondents supported the claims across the network literature suggesting that networks increase education or awareness of a social issue, improve service provision, allow for communication throughout communities, and can produce policy/legislation. These findings expand the network literature and highlight how examinations at the network level will contribute to more effectively understanding client, organization, and community level outcomes.
25


Conclusions and Findings
The following three chapters cumulatively support the development of the network literature in several key ways. First, chapters 2 and 3 contribute to theory building. Chapter 2, the exploration of why organizations join networks, findings indicate that organizations who need to coproduce outputs tend to join networks. The condition of coproduction, working together to protect survivors or preventing the crime of human trafficking, is a sufficient condition for network membership. In chapter 3, a handful of key items like trust, achieving shared goals, good communication and resolving conflict comprise a factor termed network effectiveness. The items - difficulty resolving conflict, competing for funding, difficulties sharing information, and competition among service providers in the same communities -comprise a factor of network dysfunction. These findings suggest that additional empirical work on network effectiveness and dysfunction is now possible. The final chapter in this thesis details the role of conflict in network arrangements. This qualitative chapter includes findings that suggest conflict is particularly detrimental to the effectiveness of networks, and likely related to lack of information sharing, genuine trust, competition and poor communication. Network leaders ought to carefully consider the structures and tools available to partners for conflict resolution as well as techniques for engaging all partners in meaningful and goal oriented discussions. This thesis has policy implications for the formation and sustainability of networks, the management of partnership and network arrangements, and provides a methodological advance of a much larger sample of networks.
26


CHAPTER 2
JOINERS: A QCA ANALYSIS OF NECESSARY AND SUFFICIENT CONDITIONS LEADING TO ORGANIZATIONS JOINING NETWORKS
Abstract
Networks achieve gains in service provision primarily through collaboration, strategic planning, and shared resources. This chapter compares multiple service provision networks, aimed at combating human trafficking, to determine how publicness, coproduction, theoretical models and funding influence network membership. Various agencies were surveyed about their efforts to end human trafficking. Respondents reported how the protection of survivors, prevention of trafficking, and prosecution of perpetrators occurs as well as their relationship to a partnership. By utilizing fuzzy set logic in Qualitative Comparative Analysis (QCA), aspects of an organization’s responsiveness to public demands, theoretical orientation to ending trafficking, funding, and service provision lead to network membership. Sufficiency conditions suggest that when an organization must coproduce service outcomes, they will join the network. An additional combination of conditions: not having a theoretical orientation to the work, having funding from government grants, and not being constrained by responsiveness to public demands also leads to network membership. While this second combination is a logically feasible outcome, the coproduction pathway explains the majority of cases following a sufficient condition analysis. Coproduction is a necessary condition for organizations joining networks.
Introduction
Networks, an increasingly utilized organizational form, allow independent organizations to work jointly to achieve shared goals. These partnerships permit
27


organizations to specialize, share information or resources, expand services to fill provisioning gaps, and provide a supportive space to develop new practices aimed at solving wicked problems. This introduction describes how network participants view their roles, the challenges associated with networks, and the potential networks have for combating human trafficking. An underlying assumption in the network literature is that organizations that join networks to produce goods and services achieve gains they could not by working independently. These shared gains produce positive externalities for clients, consumers, or customers. Why then, do some organizations opt out of networks? Why do organizations, especially those with high levels of specialization, join networks? Do the advantages of network membership typically outweigh all the additional managerial, communication, complexity, and coordination challenges? Do the resources, knowledge, and time required for effective networks come at a cost too high for smaller, sparsely funded organizations? This chapter explores these tensions and tradeoffs to determine some characteristics of organizations most likely to join networks. The networks under consideration here are all working to end human trafficking. The compositions of the networks vary as does the size, funding, leadership, and formality of the network.
Many organizations within the networks describe the positive externalities attained by collaboration. Network members argue that the partnership allows them “to harmonize state and federal law enforcement, nonprofits, city services, social services, and the community in a holistic anti-trafficking and victim assistance effort.” The partnership is also viewed as “collaborative efforts to enhance flow of information, identification of victims, rescue, service provision, investigations, prosecutions and follow up.” These comments - drawn
28


directly from the survey - suggest networks and partnerships can create space for collaborative advances in the community as well as lead to enhanced services for clients.
Partnership members, however, also report challenges related to network membership. A respondent to the survey had this to say when asked about the sustainability of the partnership, “goals set by members’ individual agencies often determine/derail their goals in the partnership.” Respondents also indicate that many of the networks form for reasons other than human trafficking. They cite domestic violence coalitions, housing first programs, or prosecution networks as impetus for recognizing and combating human trafficking. This respondent, for example, reports that, “participants do more than just human trafficking. Trafficking, like poverty, substance abuse, mental illness, is just one more cause of homelessness that we must understand better in order to better serve our constituents.” Based solely on this comment, the respondent’s primary consideration is ending homelessness and he/she recognizes that trafficking contributes to the problem of homelessness.
Many respondents note that their partnership either openly and honestly addressed conflict or ignored it to the determent of the partnership. Several respondents report that conflict is ignored, avoided or not discussed; each of these respondents commented that this lack of confrontation appears problematic. Building trust also appeared to be an area where respondents reported some challenge. A respondent suggests, for example, that “I have not seen this addressed other than ‘we have to trust one another’ but it sounds like a directive, not followed up with substance.” There are often tradeoffs and drawbacks to consider with any organizational form or method for service provision. This work attempts to understand how
29


and why organizations may opt into collaborative efforts while other organizations do not. Clearly, the opportunities and challenges of network collaboration are complex.
Theory
There are several ways to conceptualize the formation, structure, and outcomes of networks. Early conceptions and developments in the network literature focus on shifting the theoretic understandings of organizational arrangements. Many researchers asserted that actors beyond the firm and the government are responsible for providing goods and services in modern economies (Coase, 1937; Ostrom 1972; Williamson, 1979). Third sector entities that manage public and common goods exist and often have important social functions.
These functions, such as protecting fisheries, forests, and access to nature; social service provision; and housing supports, are goods that may be required to end wicked problems. Networks provide a mechanism for the intersection of third sector, government, and private actors. Network participation, formation, and membership are most often discussed through the notion that network structure and governance mechanisms create positive externalities over markets, hierarchies, or the bureaucratic forms of organization (Coase, 1937; Jones, Hesterly & Borgatti, 1997; Ostrom & Ostrom, 1971; Williamson, 1967, 1979). The gains typically associated with networks, like lowering transaction costs for collaboration, are those that allow members of the network to improve service, delivery, production, opportunity, or innovation.
Solving wicked problems, especially those like human trafficking, where traffickers can quickly move to different jurisdictions or transition from sex trafficking to labor trafficking, may require networks comprised of many agencies all working to end this crime. Networks, because they have the capacity to overcome collective action problems, are one
30


possible governance and organizational solutions that may increase the capacity of agencies seeking to combat trafficking (Kickert, Klijn, & Koppenjan, 1997; Powell, 2003). Additionally, networks can support information sharing, technical knowledge production, and service delivery. In the context of this study, networks, formed by two or more individual organizations, are a collaborative effort to achieve a shared goal or goals (Provan & Kenis, 2008).
Berry, Brower, Choi, Goa, Jang, Kwon, & Word (2004) discussed three different views of networks. These three views, drawn primarily from sociology, political science, and public management disciplines, promote networks as ways to understand relationships between and among organizations through social network analysis techniques, as groups of policy actors in various subsystems that act to change or amend public policy, and finally as groups coming together, primarily after the New Public Management movement, to directly provide public services to citizens. Lecy, Mergel, & Schmitz (2014) similarly viewed the network literature in public administration in three different veins. They argued that networks are conceptualized and studied as policy formation, governance, or implementation networks. While it is immensely helpful to distinguish how scholars are using the language of networks; organizations, especially those in the non-profit sector, likely participate as actors in advocacy, lobbying, service provision, and contractual partnerships with government agencies. The distinctions across network traditions are helpful for scholars, but practitioners may not view these concepts as separate types of networks. The public management vein, seemingly a combination of the public management and implementation theoretical constructs, most closely connects to the research question guiding this research, because it describes the ability of organizations to work together to achieve shared goals. This research
31


focuses on the following question: how do structural and contextual conditions influence network membership?
Human trafficking, a contemporary term for slavery of any kind, requires multisector participation and collaboration if efforts to combat it are to be successful. However, it is unclear why some organizations may opt out of network membership. Organizations that can capture the full supply chain of the public good production process may not join networks since they do not require other organizations to effectively provide services. Additionally, networks where the success of the consumptive good does not depend upon the production or consumption of other goods may lead organizations to opt of networks (Katz & Shapiro, 1985). Additionally, Suarez (2011) argued that organizations that are mandated by grant requirements to form networks, often view these requirements as onerous, counterproductive, or as a dilution of funds that could make a much larger impact if spent on one specific program or target. Non-profit agencies therefore may remain conflicted about network membership. The need for funding may strongly outweigh any of the challenges associated with mandated partnership.
An expansive literature exists that discusses the organizational form of networks, the rise of networks, and calls to further study networks (Agranoff & McGuire, 2004; Ferlie & Pettigrew, 1996; Kickert, Klijn & Koppenjan, 1997; McGuire & Agranoff, 2007; Podolny & Page, 1998; Provan & Kenis, 2008; Provan & Lemaire, 2012). There is a much smaller literature that raises concerns about the impacts of network membership for individual organizations (O’Toole, 1997, 2015; O’Toole & Meier, 2004; Raab & Milward, 2003). Not all organizations may share in the belief that network membership will allow them to more effectively achieve their goals. Additionally, the conventional assumption that New Public
32


Management (Osborne, 1993) strategies lead government organizations to partner and contract may not always ring true. In the realm of human trafficking, many of the key actors are prosecutorial or law enforcement. There is little reason to believe that law enforcement agencies will be incentivized to contract out their jurisdictional authority and the courts will not grant prosecutorial power to other entities. Without incentives to partner, will judicial or prosecutorial entities join networks? These organizations may experience pressure from local advocacy groups or crime victim service providers, but that influence may or may not be enough to drive police departments and district attorneys into joining. The following sections detail the data and measures as well as the QCA methodology utilized to explore the necessary and sufficient conditions that lead to network membership.
Data and Measures
This research is a collaborative effort with the Colorado-based Laboratory to Combat Human Trafficking (LCHT). Community-engaged research is both philosophical and participatory as the research seeks to produce knowledge and insight in partnership with community-based organizations and individuals (Anderson et al., 2012; Mikesell, Bromley,
& Khodyakov, 2013). The goal of this research is to support LCHT’s efforts to end human trafficking by identifying some of the conditions that may lead organizations to join networks. As noted above, the network literature often argues that when organizations can join together they can reap greater collective benefits, improve service provision efforts to vulnerable community members, and enhance the quality of services provided through reduction in redundancies. With this research in mind, LCHT hopes to support projects that may make organizations more able to join networks. Through this research collaboration, this
33


chapter explores the relationship between characteristics of an organization and the outcome
of that organization opting to join a network (i.e., network membership).
LCHT created and administered a survey to determine the effectiveness and intention
of work done nationally and locally to combat human trafficking. The survey was developed
as a tool for communities to assess their own efforts to end trafficking and as a project that
would systematically capture efforts across the country. Alejano-Steele (2013), co-Founder
and co-Director of LCHT, suggests that:
as the anti-trafficking movement reflects upon 10 years since the passage of the federal Trafficking Victims Protection Act, we see a landscape of scattered efforts, frustrated communities that must cobble resources to combat a vast and complex problem, victims falling through the cracks, and traffickers escaping punishment. Currently, there is no comprehensive process to gauge the number of anti-trafficking efforts underway in the U.S., let alone how effective or successful those efforts are in preventing people from being trafficked.
The survey was developed as a tool to support other organizations and networks seeking to end human trafficking. The 4P model was utilized and the survey is structured around prevention, protection, prosecution, and partnerships. A survey description with further details is included in Appendix B.
Network Membership
One hundred and eighty-six respondents completed the national survey. Each organization respondent was asked to answer the following question, “Is your organization involved in a partnership effort to combat human trafficking (formal or informal)?” This question elicited responses of either yes or no and the creation of the dependent variable is a dichotomous value of the coded response. Survey fatigue is a concern as several respondents noted that the length and detail of the survey proved challenging. The partnership question, described above is question 191 in the survey and is in the last section of the survey. Most
34


respondents did not work through all 191 questions as there were significant skip patterns, but this is something to consider. Of the 186 respondents in the national survey, no information could be collected on nine cases. The respondents agreed to take the survey, but failed to complete any responses beyond consent. One hundred and twenty-two respondents completed question 191 out of a possible 177. Just over 30% of the respondents opted out of the partnership section in the survey. There does not appear to be a consistent selection bias on the types or characteristics of organizations that opt not to respond. In this data, this missingness will be characterized as missing at random and those cases will be considered list-wise deletions in the case configurations (Allison, 2001).
Methodology
A qualitative comparative analysis (QCA) was selected for this exploratory data analysis (EDA) because it allows a researcher to view cases jointly and to further determine necessary and sufficient conditions for network membership (Ragin, 2008). QCA is a useful tool for exploring the research question posed here because QCA techniques can summarize data succinctly in the form of a truth table, check the coherence of data, check hypotheses and other existing theories through necessary and sufficient condition testing, test conjectures made based on existing theories, and create opportunities for new theory development based on the results of the case configurations (Rihoux & Ragin, 2009). An example of a particular case utilized here is a non-profit organization that subscribes to a theoretical framework for combating trafficking, participates in protection activities but does not engage in advocacy, training, or prevention activities. There are many organizations that share these characteristics that responded to the survey. According to Miethe, Hart, and Regoeczi (2008), “through applications of EDA, researchers are often able to uncover underlying structures in
35


the data, extract trends among variables, detect outliers and other anomalies, and develop more parsimonious models” (p. 231). This study uncovers trends in network membership that allows the LCHT to better serve survivors, vulnerable individuals, and organizations working to end human trafficking.
The intended goal of this methodology is to determine what conditions may be sufficient and/or necessary for an organization to opt into joining a human trafficking network. The QCA method functions much like a circuit board in a simple electronic device. For example, an organization that adheres to a theoretical model for their work has that particular switch in the on position. The most likely path to network membership may or may not be through that particular switch. Network membership, the outcome of interest in this paper, happened or did not happen in that particular context. Different switch configurations from the conditions of interest may all contribute to membership being flipped on or off. In QCA, a type of set theory, the notion of an independent variable is replaced with conditions that may be necessary or sufficient to lead to an outcome (Schneider & Wagemann, 2012). The process for determining which conditions lead to the network membership outcome is called calibration. In order to calibrate the set, the conditions must be assigned values from 0 to 1. In fuzzy set QCA those values determine the level of membership that case has in that condition. Each condition are discussed here to outline the level of membership possible. Calibration
Several different characteristics of organizations may play a role in determining if that organization will opt to join a network. Both contextual and organizational characteristics are important considerations for network membership. Just as network effectiveness includes structural and contextual factors (Turrini et al., 2010), both of these
36


factors are important considerations for individual organizations when making the calculation if joining the network will be worthwhile. The sector, focus area of the organization (prosecute, prevent, or protect), grant funded organization and a theoretic model are all included as possible attributes.
The process of calibrating this fuzzy set follows the technique described by Schneider and Wageman (2012). Calibration requires detailed theoretical and content knowledge and the researcher must determine and fully outline the membership of each case in the condition. The section below highlights each condition and frames the theoretical expectation guiding the selection of that condition. The end of the calibration section describes other considerations related to the conditions and calibration technique.
Publicness
The New Public Management movement in public administration and management heralded the idea that contractual and delegational arrangements between non-profits, private sector entities, and governments would be a cost-effective and efficient way to deliver public services (Ferlie, 1996; Hood, 1995). These arrangements also increase complexity, require more intentional communication efforts, and are often structured without effective oversight (Hood, 1991; Kaboolian, 1998; Lane, 2005). Networks providing services to combat human trafficking require shared goal setting, communication, collaboration, and structures that promote information sharing.
Organizations residing within the different sectors may have differing incentives to join networks. As previously mentioned, a governmental agency may have full jurisdiction to prosecute or to provide funding for a non-profit organization to do the work of prevention; if that is the case, the incentive may not be strong enough to incur the costs of network
37


membership and will induce non-participation. The survey respondents do not report why they enter into the network and they likely may not know themselves all the reasons for those choices. In order to calibrate the sector condition, there must be a clear dimension on whether each case is either more in or more out in terms of membership. There are a couple dimensions of sectors that may best illuminate the level of membership.
The calibration considered here follows in a vein of public management and organizational leadership that views organizations having more and less responsibility to the public broadly. Bozeman (1987) argued that all organizations are public and have some level of responsibility to respond to, or at least attend to, political authority and the will of the people. In order to operationalize how the sector of the organization may matter for network membership, the degree of publicness was assigned for each organization. Some organization types will likely be required to be more directly responsive to political authority or may be less responsive to political authority.
Each of the organization types explored here, governmental, business, and non-profit, all experience varied levels of publicness. The for-profit organizations that contribute to this work in dynamic ways are less responsive to the public directly. Their governance structures require no broad public oversight, and the CEO primarily determines the growth of the company. Businesses are considered entirely out of the public space and rarely respond to political will. Typically, government agencies ought to be viewed as the most responsive to public demands are they have the broadest permissiveness for action in the public sector and are asked to represent the community beyond single issue areas. As such, government agencies working on human trafficking are considered the most public organizations. However, there are some types of government agencies that tend to have little responsiveness
38


to public demand as they are rather removed from the electoral space and while they are providing a public service, they have strict legal mandates for their work - the prosecutorial and judicial arenas of government. These prosecutorial and judicial organizations are calibrated as being mostly out of the publicness condition as they have little need to respond to the public broadly or to specific types if political will. Non-profit organizations, governed by a board of directors and incorporated with a mission to do some kind of philanthropic or public service, are viewed as both public and somewhat responsive to political authority. Since non-profits tend to narrow focus to specific types of work of action, they may have a clearly identified or deep commitment to a specific population within the community. This choice, or ability to determine their own notion of ‘public,’ is viewed here as making them slightly less fully in the condition of publicness. Therefore the full condition of publicness is calibrated as government =1, non-profit = .75, judicial = .25, and business = 0.
Funding
Some agencies combating human trafficking in the United States received federal grant money through various reauthorizations of the Victims of Trafficking Act of 2000. Additionally, several portions of the 2013 Reauthorization of the Violence Against Women Act (VAWA 2013) provide states, law enforcement agencies and non-profit organizations financial grants to incentivize combating human trafficking. Some of this grant funding encouraged organizations to form or create networks. The funding provided to combat sex trafficking of minors, for example, authorized block grants to four communities where the population was greater than 5,000,000 people and included certain activities such as providing mental health counseling, shelter, and case management permissible under the grant.
39


While this additional burden may place constraints on the agency and focus their work in an unintended direction, the funding is essential, according to the respondents. As a result, funding for the partnership increases the likelihood that members join partnerships. This result is derived from the emphasis agencies place on funding that either carries some sort of encouraging language around network formation or the burdens of reporting and collecting data are best shared across providers doing similar community interventions.
While the notion that an agency that receives funding will join the network is straightforward, it is worthwhile to explore the combination of conditions that lead to joining a network when an agency is unfunded for partnership membership. Including this condition allows for combinations of conditions, which may be leading to the outcome of interest.
The calibration of this condition is also relatively straightforward. The organizations that receive funding for partnership receive a 1 as they are fully members of that condition. Organizations who do not receive funding retain the 0 label as they are fully excluded from that condition.
Coproduction
LCHT and the United States Department of State (State Department) utilize a framework for understanding the types of activities conducted by organizations seeking to combat human trafficking. The 4Ps include protection, prosecution, prevention, and partnership as a comprehensive way to view the work they and other agencies conduct. The individual agencies in the sample agreed to complete four different survey sections; as noted in the introduction, agencies who are able to achieve their goals without coproduction, that is without requiring a consumer, client, or another organizations to assist them in task, project or outcome completion, are likely more burdened by network membership. Coproduction can
40


be viewed as “an integrating mechanism and an incentive for resource mobilization,” as well as place additional administrative burdens on the organization due to the commitments required for network membership (Boviard, 2007, p. 846). Organizations engaged in coproduction or services may actually need other agencies to produce public goods, or, and perhaps equally meaningful, may feel pressured to join the network if it is a social norm in the community. The requirement that other agencies or individuals are required to produce the intended outputs of the organization can increase the expectation that an organization will need the network membership in order to achieve those goals. The result is that agencies may determine that the costs (administrative burdens) of network membership are outweighed by the need to build relationships that will enhance the possibility that other actors will coproduce the desired outcomes.
The 4Ps are a helpful tool for calibration when considering the case membership of individual agencies in regard to coproduction. Agencies who do not participate in prevention activities or provide services or support to survivors may have full capture of their intended outcome. These organizations are calibrated as 0 for coproduction. Organizations that participate in both prevention and protection are calibrated at 1. Organizations that participate in either prevention or protection are calibrated at .75 because they are more fully in the membership than out of it. This is an important distinction in QCA as set membership does not allow for cases to remain at the midpoint; the midpoint is a theoretically useless distinction when determining necessary and sufficient conditions (Schneider & Wagemann, 2012). Protection and prevention activities tend to be those that requires some form of coproduction, therefore, those cases are closer to full membership in the set than out of it.
41


Theoretical Model
Additionally, LCHT was interested in understanding if harm reduction, demand reduction, behavioral interventions or other types of social science models are articulated in the fight to end sex and labor trafficking. As such, this analysis includes the condition that the agency with a conceptual or theoretical model is worth further exploration. Each organization was offered three different types of theoretical models: behavioral, harm reduction, and demand reduction as options. They were also able to select that they do not utilize a theoretical model to guide their work. While there is not necessarily a strong theoretical prediction that would stem from an organization’s interest in operating from a model and their willingness to join a network, it is a condition worth testing as it may operate in conjunction (as a necessary or sufficient condition) with other conditions more theoretically driven. Cases where respondents selected a theoretical model for combating human trafficking are listed as fully in the set members. The cases where organizations do not use a model are calibrated as zeros.
Table 2.1: Organizational Conditions Influencing Network Membership
Independent Variable Concept Indicator

Publicness Responsiveness to public demands 1 - government, .75 - non-profit, .25 -judicial, and 0 - business
Funding Organizations have the capacity through federal funding to coordinate or provide services 1 - full set membership, 0 - no membership
42


Table 2.1: Organizational Conditions Influencing Network Membership (Cont.)
Coproduction Organizations increase awareness, advocacy and education towards addressing a community's systemic vulnerability 1 - full set membership, .75 - partial set membership, 0 - no set membership
Theory Agency utilizes a theoretical model to drive their work. 1 - full set membership, 0 - no membership
Analysis
This chapter offers better understanding of which organizations may opt into networks. Set-theory (Miethe et al., 2008), a type of qualitative case analysis (Ragin, 2000), allows for determining which conditions are likely necessary or sufficient for joining networks. The notations of set-theory follow from Boolean and fuzzy algebra and utilize logical AND/OR/NOT operations (Schneider & Wageman, 2012). One hundred and eighty-six cases were included in the analysis. The researcher utilized the QCA and QCAGUI packages in R (Dusa, 2007; Dusa & Thiem, 2015) to determine the logical minimization or the analysis of sufficiency (Schneider & Wageman, 2012). The empirical output of logical minimization is a truth table. Table 2.2 details each of the set relations that may lead to network membership.
Necessary and Sufficient Conditions in Combatting Human Trafficking Networks
Ragin (2008) suggested that a Standard Analysis, one in which three different types
of solutions to QCA are explored, consists of:
(a) the complex solution, or what we prefer to call the conservative solution, where no assumption about any logical remainder is made; (b) the most parsimonious solution, which is based on simplifying assumptions, that is, those remainders are included into
43


the logical minimization that contribute to parsimony; and (c) the intermediate solution, which relies on easy counterfactuals, that is, only those simplifying assumptions are included that are in line with theory-driven directional expectations. (Schneider & Wagemann, 2013, p. 211)
Schneider and Wagemann (2012, p. 91 - 92) recommended conducting two analysis procedures in order to determine both necessary and sufficient conditions. This statistical procedure includes a top-down analysis of sufficiency that produces the causal combinations of conditions leading to the outcome with a logical minimization process resulting in only the empirically verifiable combinations and a bottom-up analysis of necessity that reveals the assertion that “a logical AND conjunction of two or more conditions can only be necessary for Y if, and only if, all single conditions involved in the conjunction are necessary on their own” (p. 92). However, this assumption of necessary conditions independently leading to the outcome is the focus of a new methodological debate.
This debate in the analysis and description of findings in the QCA technique emerged between Thiem (2016) and Schneider and Wagemann (2012)’s Theory-Guided Enhanced Standard Analysis (T/ESA) procedure. Thiem (2016) argued that when applying T/ESA, utilizing a minimization procedure that distinguishes necessary conditions from sufficient in a two step process (Schneider & Wagemann, 2003; ~ 2012) is erroneous due to faulty assumptions related to the mathematical propositions of Boolean algebra and the combination of a researcher’s application of causality to a phenomenon under study. Essentially, Thiem (2016) suggested that the assertion from Schnieder and Wagemann (2012) claiming that the bottom-up approach requires each “necessary condition also be an isolated cause of given effects” (p. 480) leads to a conservative solution not intermediate or parsimonious solutions. This result is problematic in the reporting of QCA analyses as it leaves the processes related to both conducting an analysis and reporting results in a shared
44


language for peers problematic. As such, the following sections begin with an analysis of sufficient conditions, logical remainders, missingness and an additional robustness test on findings. A bottom-up process for finding necessary conditions was conducted, but no conditions in this dataset appeared to support any necessary condition1. As noted below, the presence of coproduction is a sufficient pathway to network membership but cases (very limited numbers) are present that support additional pathways to network membership. The empirical observations presented here suggest that the parsimonious and conservative solutions do not differ and presenting an intermediate solution is not possible based on Thiem’s (2016) arguments against T/ESA.
Sufficient Conditions
The truth table below describes four conditions leading to network membership. Following the results of Table 2.2, the primitive expression for all condition sets on the outcome of network membership results in the following:
TCFP + TCF-P + TC-F-P + TC-FP + -TCFP + -TCF-P + -TC-FP + -TCF-P + ~T~CF~P Y
The process of logical minimization requires reducing these primitive conditions down to the subset relations between the condition and the outcome. In eight of the nine possible sufficient subset relations, Coproduction values greater than .05 produce the network membership outcome. Thus the subset relations can be logically minimized to:
C + tFp -> Y
Had the empirical observation of COPRODUCTION as the only pathway consistent with the outcome network membership, COPRODUCTION could be considered a necessary condition. However, the additional pathway suggests that coproduction is a sufficient pathway with high coverage.
45


Table 2.2: Theory, Coproduction, Publicness and Funding Truth Table
Network Conditions Membership
Cases Frequency T C F P N
15,27,32,41,46,47,52,70,151,176,183 11 1 1 1 1 1
121,148 2 1 1 1 0 1
2,7,12,50,53,64,65,120 8 1 1 0 0 1
0 1 0 0 0 1
0 1 0 1 1 1
0 1 0 0 1 1
10,40,54,72,83,108,109,123,128,157,174,181 12 1 1 0 1 1
0 1 0 1 0 1
17,21,24,25,26,31,49,101,106,177 10 0 1 1 1 1
43,44 2 0 1 1 0 1
141,149,152,165 4 0 1 0 0 1
0 0 0 0 0 1
0 0 0 1 1 1
0 0 0 0 1 1
16,20,38,56,57,63,68,71,90,94,100,124,153,163, 179,182 16 0 1 0 1 1
133 1 0 0 1 0 1
Total Cases 66
The condition of the existence of Coproduction, that is, the organization is conducting prevention and/or protection activities, leads to network membership in 65 of the 66 cases. The single case of tFp comes as a result of a grant writing firm that joined a network (possibly paid to write grants) and responded that they do not coproduce any services related to preventing trafficking or protecting survivors, they are not responsive to public demands as they are a business, and yet they do receive funds and seek funds for their work to combat human trafficking. While this is an interesting logical outcome, it is a rare condition leading to network membership.
There are two parameters of fit in QCA: consistency and coverage. Consistency is the the way in which a researcher handles combinations of conditions that lead to both the
46


outcome and the absence of the outcome. In this particular empirical setting, after missingness is handled across all four conditions, there are not cases observed where the outcome, network membership, is not present. This finding should not be interpreted to suggest that all cases with this combination of conditions always produces network membership; instead, the consistency measure is a value of 1, meaning that all cases observed that lead to network membership fall into the two sufficient paths detailed above. The Coproduction path, has both high raw and unique coverage. Coverage, a measure of fit that explains the amount of overlap between the subset condition and the outcome of interest, can be explored through both raw coverage and unique coverage measures (Dusa & Theim, 2012; Schneider & Wagemann, 2012). Coproduction, has a raw coverage of .883 and unique coverage of .818. Coverage values range from 0 to 1. Clearly, a large amount of the outcome network membership occurs through the Coproduction pathway. The absence of theory and publicness with the presence of Funding (tFp), has a raw coverage of .080 and a unique coverage of .015. As previously discussed, this path accounts for a very small subset of the outcome. Two individual cases cover both solution terms (cases 43 and 44) and are therefore calculated as contributing to raw coverage for each pathway. This reasserts the finding that tFp is likely a very rare sufficient pathway to network membership.
Limited Diversity
Limited diversity occurs when logical remainders are present during the analysis (Schneider & Wagemann, 2012). Logical remainders, possible subset relations that are not empirically observed in the data, are nonetheless important considerations for researchers. In this analysis, for example, the logical remainder tcFP would be an organization that fails to subscribe to a particularly theory for the work, does not do protection of survivors or
47


prevention of the crime, but does receive funding and is responsible to the public. As noted in the introduction, many partners in the human trafficking prevention and reduction space are focused on other types of crime or vulnerabilities in the community. A non-profit day shelter, for example, may be funded through Office of Victim’s Assistance and Bureau of Justice Affairs funds, may be required to join a network, and yet may not provide any direct services or efforts against trafficking. While this scenario was not empirically observed, the pathway to network membership may still be present.
An argument could be made that tFp is a logical remainder itself as there is only a singe case observed. The definition of logical remainders from Schneider and Wagemann (2012) is “truth table rows that lack enough empirical evidence to be subjected to a test of sufficiency” (p. 152). Since 66 cases were examined, the concept of arithmetic remainders (too few cases for a single case in all possible combinations) would neither apply nor would the scenario of this being an impossible remainder as it was empirically observed. Clearly, there is no clear guidance on how much empirical evidence is enough to support the existence of the sufficient pathway. Typically, one single case is enough and as such, tcFP remains included in this analysis as a sufficient pathway to network membership.
Missingness
Missingness is an additional consideration necessary for a complete analysis of the conditions leading to network membership. Missingness does not appear to be missingness at random (Allison, 2001) on the funding variable. Of the 120 cases dropped from the analysis, over half (63), were missing on both the funding condition and the network membership outcome. The funding variable is the only condition drawn from the partnership section of the survey. The partnership section was the final section (of four total) and it appears that
48


respondents may have suffered from survey fatigue by the time they reached question number 191 (the first question of the fourth section). As such, the researcher opted to additionally conduct a second QCA analysis, which dropped this funding variable to compare the results and to create a robustness check on the set relations found in the prior section. While the presence of funding and the non-presence of theory, coproduction, and publicness did result in a case empirically, that particular case was a unique condition in the data. Table 2.3 details 93 cases and provides support for two findings in the full four conditions of Table 2.2 (above).
Table 2.3: Theory, Coproduction, and Publicness Truth Table
Conditions Network
___________ Membership
Cases Frequency T c p N
4,10,15,27,32,37,40,41,46,47,52,54,58,7 0,71,72,83,108,109,123,126,128,151,15 7,161,167,174,176,179,181,183,186 32 1 1 1 1
2,7,12,50,53,64,65,120,121,148,178 9 1 1 0 1
0 1 0 1 1
0 1 0 0 1
8,14,16,17,19,20,21,24,25,26,31,35,363 8,49,56,57,63,68,75,90,94,97,100,101,1 06,115,118,124,125,135,140,143,153,16 3,169,173,175,177,182 40 0 1 1 1
43,44,119,141,149,152,165, 7 0 1 0 1
42,60,87 3 0 0 1 1
111,133 2 0 0 0 1
Total Cases 93
The logical minimization of the Table 2.3: Three Conditions results in the equation: C + tc -> Y
Coproduction is again found to be a sufficient condition to network membership with a much larger number of cases. Of 93 cases, 88 cases were observed with the presence of
49


coproduction leading to network membership. Another sufficient pathway, no coproduction and no theory (regardless of the presence of absence of publicness), represents five cases; tc has much smaller coverage but again meets the empirical threshold for consideration as a sufficient pathway. No theory coupled with no coproduction is interesting for interpretation. These organizations may seemingly have no direct motivation for direct membership but each of the five cases here all work to end trafficking through other avenues of support. The case observed and mentioned above is a grant writer, another makes art to raise social justice awareness, yet another views themselves as a funding pass through for special services and the last two work on policy and advocacy for women broadly and intersect with trafficking through gender-based violence reduction. These five cases may suggest motivation for networks to consider partners not directly connected to the explicit mission but those that can provide additional viewpoints, contribute to effectiveness through other avenues, or simply because the organization is interested in supporting the collaborative efforts.
Conclusion
Coproduction appears to be an important condition for network membership.
Network membership is substantially driven by organizations that work toward efforts protecting survivors and/or preventing trafficking. These organizations may come to the work with diverse theoretical orientations, from public or private organizations and may or may not lack funding to combat trafficking. In many ways, this is an incredibly refreshing and powerful finding. Perhaps this is truly a necessary condition - organizations that want to work toward multiple different avenues of ending trafficking tend to believe that they cannot do that work alone. This finding breathes more life into the interpretations that while networks may have constraints, limitations, and costs, organizations that are engaged in
50


service provision recognize that those tradeoffs are worth the time, resources, and positive externalities gained from joining networks.
Future studies should further explore and understand the relationships among sufficient pathways to not joining networks. The full analysis of all four conditions did not reveal any pathways to non-membership; however, the robustness check of dropping funding, suggests that there are five sufficient pathways to nonmembership. These are worthy of additional consideration in future papers. Prosecution agencies and judicial government agencies are not clearly implicated as joining or not joining networks as discussed previously as a possible pathway to membership. The notion of responsiveness to the public does not directly create a subset relationship with network membership but it may still be implicated in sufficient pathways to non-membership. Further examination may provide additional insights into network arrangements that support and do not support agencies with jurisdictional authority or issue capture.
Network effectiveness, the ability of networks to sustain collective performance, trust each other, work together, share information, and achieve goals, may be impacted by the different pathways to membership. The networks that contain organizations from pathways achieved not through Coproduction may enhance the network effectiveness through additional and diverse perspectives or may inhibit and slow networks through tangential missions, lack of attention to issues at the core of trafficking, or perceptions that preventing other forms of violence are more important. Understanding the more complex arrangements of networks and the formation and sustainability of those networks is key to enhancing effectiveness, partnerships seeking to end wicked problems, and the complexities of working together at the organizational level.
51


CHAPTER 3
NETWORK EFFECTIVENESS: COLLABORATING TO COMBAT HUMAN
TRAFFICKING
Abstract
Networks, collaborations of more than three agencies working to achieve shared goals, hope to attain collective gains through partnerships, sharing resources, and/or providing more comprehensive services. The goal of many of these networks is to tackle wicked problems, those too complex, systemic, or challenging to end without a broader effort. There is still little empirical evidence that these efficiencies are achieved as often as touted or that network governance is effectively structured to provide public goods and services. Network effectiveness, the ability of multiple organizations to trust each other, work together, share information, socialize, achieve goals, and resolve conflict is examined across 186 networks that are all working to combat human trafficking. Ending human trafficking, a form of slavery that preys on vulnerable people living within institutions of inequality, will require effective collaborative governance. Sampling networks seeking to achieve similar goals allows for comparison where similar contextual and structural challenges likely exist. Through factor analysis, two distinct concepts emerged from the data: effectiveness and dysfunction. The specific items that comprise these factors are useful for inclusion in future research, for more deeply understanding the differences among client, community, and network level effectiveness, and encouraging network formation and maintenance that attends to both increasing effectiveness and decreasing dysfunction.
Introduction
52


Foundations, governments, private entities and non-profits are increasingly developing networks that improve service delivery, make a collective impact greater than they could do alone, and share resources (e.g., scarce funding). As agencies increase their interdependencies, they also increase the layers of bureaucracy, administration, and authority. Due to these increasing cognitive challenges related to additional layers of personnel as well as competing strategic goals for ending wicked problems, the ability of the networks to achieve goals, resolve conflict, or develop trust may suffer. Some networks seem to overcome challenges more effectively than others or manage to find clarity where uncertainty previously ruled. By developing and testing a theory of network effectiveness, this study enhances understanding of the management and administration of networks.
The vast literature on networks can be understood as developing in three research traditions and represent conceptually different ways to understand the study of networks. The three traditions of network research: sociology, political science, and public management produced innovations in social network analysis, policy innovation, change, and agenda setting, and development of managerial structures and behaviors as well as network performance and outcomes (Berry et al., 2004). These traditions built on each other, intertwined, and provided powerful insights like the concept of “structural holes” (Burt,
1992), density (Marsden, 1990), capacity to end wicked problems through partnerships (van Bueren, Klijn, & Koppenjan. 2003; Head, 2008), and the capacity for policy innovation and knowledge transfer across a network (Rhodes, 1997; Robinson, 2006; Weber & Khademian, 2008). Through a process of tracing citation relationships, Lecy, Mergel, and Schmitz (2014) argued that scholars do a poor job of defining the way in which they conceptualize the network under consideration, the research tradition influencing the work, or the research
53


program. The literature tracing process led these scholars to structure the public administration literature on networks in three conceptual areas: policy formation, governance, and implementation (Lecy, Mergel, & Schmitz, 2014). The major development provided in this research is in the implementation of networks vein and focuses specifically on comparing the capacity of whole networks to solve wicked problems.
Wicked Problems and Human Trafficking
Poverty, human rights abuses, unemployment, homelessness, and violence prevention are all forms of wicked problems. These problems share the attributes of being sustained in communities over long periods of time, having multiple and competing definitions, needing public and administrative attention, and requiring solutions that may be contested by different segments of the population (i.e., the solutions are often politicized). These issues plague society because of the immense cognitive and strategic uncertainty that surround both the problem and the solutions (van Bueren, Klijn, & Koppenjan. 2003).
Roberts (2000) provided a conceptual distinction around various types of problems. The first type, simple problems, may initially present a challenge, but when expertise is applied there is a clear definition of the problem and systematic ways to solve the issue. Complex problems present themselves as ones where there is consensus on what constitutes the problem. Citizens and policy makers can agree that the status quo is the problem and that something needs to be done. Complex problems, however, have multiple solutions. The final type of public problems is considered wicked because both the identification and possible solutions are contested. Citizens and government officials disagree about the root causes or identification, and the problem persists after numerous attempts to combat the issue.
54


Human trafficking is one example of a contemporary wicked problem with global implications. As borders become more porous, information about the quality of life in other places is easily shared, and as the pace of economic globalization grows so does the demand for labor, personal safety, or freedom. Trafficking in humans increases in conflict areas and vulnerable populations are especially at risk. Trafficking spans various types of crimes and impacts adult men as well as women and children. Too often, bureaucracies may oversimplify the problem as a sex crime only affecting youth and children. Head (2008) suggested this simplification is a direct result of the growth in needing to establish targets and goals that can be met by bureaucrats and managers in the public sector. Simplification allows for quantification of some aspects of this crime, but may lead to negative externalities for other potential victims and survivors of trafficking. Wicked problems like human trafficking appear intractable because they are complex, require participation from multiple sectors, and solutions vary among communities.
Human trafficking represents both a criminal offense (federal statute created by the Trafficking Victims Protection Act of 2000) and a violation of an individual’s human rights. Trafficking is “defined by the fact that victims” agency and other human rights are violated through force, fraud, or coercion” (Foot, 2016, p. 2). This violation primarily comes as a result of some form of vulnerability. Women and children, the most frequent victims of trafficking, are typically more marginalized in societies across the globe or more likely to be perpetrated against (Roberts, 2008). Men, especially those who feel they must provide for themselves and their families are more frequently trafficked to provide labor. While vulnerable populations are at greater risk to be forced, coerced, or fraudulently deceived, trafficking is also viewed today as any modem form of slavery. The root causes of
55


contemporary slavery are debatable and may vary depending upon the state, region, or country. This complexity leads to the cognitive and strategic uncertainty that defines wicked problems.
Human trafficking is particularly problematic because combating it requires solutions, collaborations, and partnerships across many levels of governance and jurisdictions. A truly wicked problem, according to Roberts (2000), is one where “nothing really bounds the problem solving process - it is experienced as ambiguous, fluid, complex, political, and frustrating as hell” (p. 2). Current efforts to end trafficking, especially those funded by the United States government and the United Nations, have political will to end child sex trafficking. Labor trafficking, border smuggling, drug running, and other forms of coercion, while less flashy also require time, attention, funding, and political will if they are to end. Trafficking prevalence shifts geographically as conflict situations, black markets, and labor needs change globally. When a single municipality, county, state, or country decreases trafficking, the crime may easily shift into jurisdictions that are more permissive just as a markets shift with changes in supply and demand. This situation further complicates problem identification, jurisdictional authority, and potential solutions. Ultimately, human trafficking is both wicked and frustrating as hell. Many communities recognize the need to combat this problem, to provide attention and energy to the root causes by finding ways to prevent the crime, protect survivors, and prosecute the offense. Combating the multiple causes and outcomes related to trafficking also demands effective communication across agencies and governments, timely information sharing, and prevention, protection, and prosecution efforts that build on collaboration among all parties. Networks are arrangements that allow
56


organizations and individuals from different sectors, problem orientations, and service provision areas to come together and work to end wicked problems from multiple angles.
Human trafficking represents an example of what Roberts (2000) described as a truly wicked problem. She identifies three possible strategies for coping with wicked problems: authoritative, competitive, and collaborative. The focus of this research is to examine how whole networks function in collaboration to produce outcomes aimed at ending human trafficking.
Whole Networks
Since networks have the collaborative capacity to combat wicked problems, scholars typically view them through a positive lens (Ansell & Gash, 2008; Klijn & Koppenjan, 2000; Koliba, Meek, & Zia, 2010; Meier & O'Toole, 2003; Padgett & Ansell, 1993; Provan & Milward, 1995). One problematic aspect of the broadly and quickly expanding literature on networks is that the term “networks” is increasingly nebulous, represents a different concept to different people, and is inconsistently applied (Borgatti & Foster, 2003). This section identifies a conceptual definition of networks and connects that concept to the operationalization of networks in the survey from which the data for this project were drawn.
The notion that several organizations or agencies work together toward achieving a common goal is the heart of understanding these network arrangements. O’Toole and Meier (2004) argued that networks are an interdependent group of multiple organizations or parts of organizations whereby the connections among the actors are not arranged as superior to each other. This definition suggests that members of the network are neither subordinate nor report to one another. Provan and Kenis (2008) specifically defined networks as “groups of three or more legally autonomous organizations that work together to achieve not only their own
57


goals but also a collective goal” (p. 231). Provan and Kenis (2008) also argued that network
configurations often require lead organizations or network administrators to provide an
effective governance mechanism for positive outcomes. The structure of networks, leadership
of the network, and goal-directed behavior are all considered here. Specifically, this article
addresses the research questions from the “network as a form of governance’ approach as
opposed to a “network analytical” approach (Provan & Kenis, 2008). The network as
governance focus pays special attention to meta-level characteristics of the network. The
network leadership, funding, structure, stability, and size are all important considerations in
this perspective. The Laboratory to Combat Human Trafficking, the community partner
connected to this project, defines partnership research as:
Partnership measures acknowledge that combating human trafficking requires a comprehensive response through the cooperation of multiple sectors.
Partnerships bring together diverse experiences, amplify messages, and leverage resources. For the purpose of this survey, an anti-human trafficking partnership refers to a ... [e.g., a cooperative relationship between two or more organizations established for the purpose of jointly combating human trafficking in some way.]
This language, while utilized specifically for the survey tool, suggests ways in which a group of organizations may collaborate to achieve their stated goals related to human trafficking.
More specifically, this study examines various “whole” public management networks. Public management networks include “agencies involved in public policy making and/or administrative structure through which public goods and services may be planned, designed, produced, and delivered (and any or all of the activities)” (McGuire & Agranoff, 2007). O’Toole (1997) and Agranoff and McGuire (2001) called for focused and continued efforts to study public management networks. While other networks are informal, lack funding mechanisms for combating trafficking, and have little regulatory authority to prosecute or
58


criminalize trafficking in their local communities. This article addresses the following research question: How can we measure and understand network level effectiveness?
Theory
A comprehensive theory on network effectiveness is currently nonexistent (Turrini et al., 2010; O’Toole, 2015). However, based on previous work related to network effectiveness (Provan & Milward, 1995; Provan & Sebastian, 1998; Turrini et al., 2010) the structural, functional, managerial and contextual descriptions of networks relate to the performance of networks (Klijn & Koppenjan, 2000; Provan & Milward, 2000; Klijn, 2005; Provan & Kenis, 2008). The goal of this research is to explore the relationships among indicators of network performance. The most comprehensive attempt to identify a framework for understanding network effectiveness comes from Turrini et al. (2010). They posit the following:
CCNTEXTUAl
CHARACTCMSTKS
STWUCHJWAI ana
OTMEKVAISVE ... **».,
CHAfiACTERlSr/CS "**-.... _________________________________
MnKWK
-----------------1 ____FERFORIAANZB
FUNCTIONING I________________)
CHARACTERISTICS f
' s
MA’JAC,fR\Al RGiFS [•'*
Figure 3.1: Turrini et al. (2010) Network Effectiveness Framework Overview (© Public Administration, Wiley Online Library)
The Turrini et al. (2010) framework is a way to conceptualize the determinants of network effectiveness.
fctmxxK
FERrQFtUiMZB
59


Of the limited studies examining whole networks, the definition of network effectiveness varies (Turrini et al., 2010). Provan and Milward (2001) identified three types of effectiveness: client level, community level, and network level as indicators of network success. Network level effectiveness refers to the sustainability, legitimacy, and maintenance of the network. Ferlie and Pettigrew (1996) examined a single network for duration and effectiveness and a handful of other studies have examined the ability of networks to achieve stated goals, innovate, and remain sustainable, but empirical work has failed to address all of these types of effectiveness. The conceptual definition of network effectiveness utilized for this study is the network level capacity of the actors to work together to achieve goals related to combating human trafficking. Achieving goals is a key tenant in the definition, though working together may not be as easily operationalized. Working together represents the network’s ability to communicate, build trust, and resolve conflict. Clearly, effectiveness differs from efficiency and utility. Simon (1965), with interpretation from Baldwin (1985), clarified the notion that efficiency and utility imply a relationship between costs and outputs whereas effectiveness is solely a conception of the benefits, outputs, or outcomes of the work. The current research analysis attends to the observable outcomes of working together and achieving goals, not the costs imposed on the actors to achieve those outcomes.
Future research will begin to address the relationships among the hypothesized determinants and network performance, but first, an empirical operationalization of network effectiveness is necessary. The following section discusses the results of a factor analysis of items collected from a national survey of networks seeking to combat human trafficking.
Methodology
60


In 2013, organizations involved in combating human trafficking completed a survey conducted by a self-identified, organization that seeks to end this wicked social problem. The Laboratory to Combat Human Trafficking (LCHT) supports research, response, and reporting efforts in Colorado and works closely with organizations across the state developed to prevent the crime, protect survivors, and prosecute perpetrators. Two surveys were administered: one was a national sample and the other was distributed to organizations within a single state.2 Since the surveys were administered simultaneously, this study utilizes only national survey data as a sample of the population of networks. The national survey is larger, likely more representative of unique, whole networks, and covers much broader geographic areas. The national respondents were selected through purposive and convenience sampling, which lead to a sample of 453 organizations that received the survey. One hundred and eighty-six organizations returned surveys and the executive or program director in the agency completed the survey the majority of the time (55%). Further details about the methodology of the survey or the types of organizations included in the sample is available at http://coloradoproiect.combathumantrafficking.org/resultsandfmdings/nationalreport. The following section details the use of factor analysis to determine how the items in the survey represent network effectiveness.
Factor Analysis
Factor analysis, a technique in which a researcher must determine how to rotate and group items tapping a single latent construct (DeVellis, 2012), was utilized to explore survey items believed to influence the functioning, performance, and overall effectiveness of the networks. The survey contained a question battery asking the respondent to consider several
2
Complete access to survey questions, methodology, and sampling can be found at http://coloradoproject.combathumantrafficking.org/resultsandfindings/nationalreport
61


statements related to the partnership they opted to describe. Respondents who belong to multiple partnerships were asked to think of the most diverse partnership in terms of sector membership. Also, each agency was allowed to complete the battery multiple times for each partnership; however, only the most diverse partnership is included in this data set. The battery includes items on a five point Likert scale related to statements that refer to the partnership. One item states, for example, “there is a great deal of trust among members” and the respondent selected strongly disagree, disagree, neutral, agree, or strongly agree.3
A total of 11 statements were included in the battery (see Table 3.1 for specific survey items). A factor analysis was conducted to determine if these items might effectively represent fewer, underlying constructs (Jolliffe, 2002; Brown, 2015). Based on the statements, generally positively or negatively related to the functioning of the network, the researcher hypothesized that there would be two distinct factors and that exploratory factor analysis (EFA) through principle axis factoring and rotation is most appropriate. As there is little empirical evidence or measures currently in use for network effectiveness, this approach is an exploratory factor analysis rather than confirmatory. Brown (2015) suggested that confirmatory factor analysis (CFA) is only appropriate when there are specific theoretical priors for the communality estimate, the structure of the variables, matrices and variance, and how the indicators are related to the factor or construct. Exploratory factor analysis is the most appropriate technique based on the state of theory on network effectiveness and the oft-unreliable nature of confirmatory factor analysis (Borsboom, 2006).
3
A quick note on the missingness in this battery. There are 186 respondents for the larger survey and 122 report on membership in a network. Of those 122, 95 respondents report that they are, indeed, part of a network combatting human trafficking. A factor analysis program will generally use casewise deletion leading to a total N value of 71 for the factor analysis.
62


Specification of the exploratory factor analysis is arguably one of the most important considerations when attempting to identify the factors that represent suspected latent concepts (Fabrigar et al., 1999; Preacher & MacCallum, 2003). These considerations include selecting the matrix of coefficients to be analyzed, the number of factors to be extracted, the extraction method, if the factors should be rotated, and the interpretation of the resulting factor scores (Thompson, 2004). The following section outlines each of those choice points.
Since the statements are generally written in positive terms or negative terms, a theoretical assumption is that the positive items will likely be correlated and the more negative statements will likely also have a correlated relationship. Additionally, trust, communication, and achieving goals together are implicated as ways in which networks can be viewed as effective (Provan & Milward, 2000; Provan & Kenis, 2008; Turrini et al., 2010). Table 3.1 shows the expected factor loadings from the survey items: one that is the latent representation of effectiveness and representation of network dysfunction:
______________Table 3,1: Hypothesized Factor Loadings_____________
________Effectiveness________________________Dysfunction__________
There is a great deal of trust among members Conflict can arise among members because of the different agency/organizati onal missions
Member sometimes socialize together Conflict arises among members because of competing definitions of human trafficking
There is good communication among members There is competition among members that work within the same communities
The partnership will remain strong if the current leader leaves There are difficulties sharing information among members about victims of trafficking due to confidentiality policies
63


Table 3.1: Hypothesized Factor Loadings (Cont.)
Effectiveness Dysfunction
There is low turnover of members in the partnership
The members of the partnership are able to resolve conflict effectively
The partnership has achieved it annual goals over the past year
Several different techniques for extracting the factors are available to researchers looking to explore the relationship between multiple variables, which may have a relationship to an underlying construct. Generally, researchers would consider principle component analysis and common factor analysis (Galbraith, Moustaki, Bartholomew, & Steele, 2002; Jolliffe, 2002; Warner, 2008) Principle component analysis (PCA) is a sum of the variance of the variables - the factors are created directly from the original data and represent essentially a sum of variance from the variables included in the factor under consideration (Galbraith, Moustaki, Bartholomew, & Steele, 2002; Warner, 2002). Fabrigar et al. (1999) cautioned that researchers may fail to realize that PCA is a data reduction technique and fails to account for the individual variance and error terms of each variable in the factor analysis. Principle factor analysis or common factors, however, is a multiple regression technique that utilizes the shared variance of a set of variables to determine a communality estimate. This estimate is the variance that is predicted as a result of the correlation among the variables that comprise the factor (Warner, 2002). This communality value is replaced in the R matrix and several statistical packages can run this estimate over numerous iterations to improve the accuracy and reliability of the estimate (Fabrigar & Wegener, 2012). Another distinction between PCA and the common factoring reported here
64


is that PCA does not distinguish between common and unique variance. Common factoring is preferable here because it is likely that trust, communication, socializing, and other items that may compose one of the factors lack similar error terms or distributions (Fabrigar &
Wegener, 2012). Therefore, the common factoring is utilized here.
Both maximum likelihood extraction and the principle factors extraction techniques estimate parameters of unique variance and factor loadings (common factoring), but maximum likelihood (ML) techniques allow for statistical testing of correlations and factor loadings (Fabrigar, Wegener, MacCallum, & Strahan, 1999). Therefore, ML estimations must assume that the included variables are normally distributed and do not violate other normality assumptions. Principle factors allow for non-normality, but are often a challenge to report and compare the reliability or validity of results (Preacher & MacCallum, 2003). Principle factoring can be iterative; the process utilizes the sample correlation matrix initially then estimates new communality estimates in the matrix until the process reaches convergence (Preacher & MacCallum, 2003). Based on findings for exploratory factor analysis, only PAF is utilized here as it appears to outperform ML and inferential statistics are unnecessary (or reliable) when conducting exploratory analysis (Costello & Osborne, 2009; De Winter & Dodou, 2012).
A decision of rotating the axis or not must also be considered (Fabrigar & Wegener, 2012). Rotating the factor analysis allows for the axis of each factor to shift in order to describe a more elegant and simple analysis for interpretation (Fabrigar & Wegener, 2012; StataCorp, 2012). Orthogonal rotation ensures the assumption that the factors that are uncorrelated remain in place, whereas oblique rotation allows for factors to violate that assumption and be correlated. Since this is exploratory factor analysis, oblique rotation was
65


utilized. There is no evidence to suggest that each of the items in the battery are uncorrelated, and if fact, it is likely that the ability to resolve conflict effectively is somewhat correlated to the relevance of conflicts that arise out of mission differences or from varying definitions of human trafficking. Additionally, not only could the items be correlated, the dysfunction of networks, one of the hypothesized factors may or not may not be correlated with the effectiveness of networks. Theoretically, a network may remain effective and at the same time have dysfunction. For example, it is not unrealistic to expect that network members may disagree on the definitions of the wicked problem, to have conflicts as a result of differing missions or to have challenges related to sharing information and yet, with dysfunction present, the network will still have some level of effectiveness that can be measured and evaluated. The dysfunctions in the network may or may not have a relationship with the effectiveness of the network and further studies will need to explore the relationship between network determinants, dysfunction, and effectiveness. As such, oblique rotation is the most appropriate selection.
Results
In order to utilize the iterative option of principle factors, the first step was to examine non-iterative models to determine that two factors were likely present. Current theory supports this hypothesis, but it was also necessary to test. In Kaiser criterion, screeplot and parallel analysis, two factors were the best fit. The eigenvalues reported below support both a scree test and parallel analyses of the iterative principle factor extraction.4 Typically eigenvalues greater than one would suggest that a distinct factor is present and this is often referred to as the Kaiser criterion (Fabrigar & Wegener, 2012; DeVillis, 2012; Dillon & Goldstein, 1984). The eigenvalues for factor 1 and factor 2 are greater than one and the scree
4 Images of Scree Test and Parallel Analysis provided in Appendix C.
66


plot appears to drop off considerably after the second factor. Parallel analysis results also suggest that two factors are present.
_______________Table 3.2: Iterated Principle Factors Analysis_____________
Variable
Factorl Factor2 Uniqueness
Trust 0.891 -0.095 0.176
Socialize 0.453 0.200 0.777
Compete for Funding -0.068 0.587 0.641
Good Communication 0.851 -0.116 0.239
Mission Differences lead to Conflict -0.159 0.401 0.798
Definition of HT leads to Conflict -0.397 0.425 0.621
Competition in Same Community 0.014 0.738 0.457
New Leadership 0.300 -0.026 0.908
Low Turnover 0.529 0.267 0.683
Difficulty Sharing Information 0.177 0.565 0.673
Resolve Conflict Effectively 0.859 -0.151 0.209
Achieved Past Goals 0.694 -0.203 0.443
The graph for the screeplot and the parallel analysis are located in Appendix C. Based on all three of these results, the hypothesis is generally supported that the rotated, iterative principle factor model of two distinct factors suggests that while two factors are present, several items may not be a good fit for the resulting factors. Table 3.2 details the parameters of fit for the factor analysis.
The loading plot, the item communalities and an exploration of the Chronbach’s alpha leads to the conclusion that several of the included items from the battery fail to strongly load onto either factor 1 or factor 2. The item communalities reported above suggest that socializing, mission differences lead to conflict, and a strong network with a leadership change fail to meet general social science guidelines; although a hard and fast cut-point is not identified the current research standards suggest that values between .32 and .7 may be
67


considered when the item theoretically fits the factor and items above .7 should generally be included (Costello & Osborne, 2009; Velicer & Fava, 1998).
____________________________Table 3,3: Iterated Principle Factor Items_________________________
Factor 1
Effectiveness
(Eigenvalue = 3.82)
Factor 2
________Dysfunction
(Eigenvalue = 1.55)
Does Not Load
Proportion of variance attributed to this factor: .68 Proportion of variance attributed to this factor: .35

There is a great deal of trust among members Conflict arises among members because of competing definitions of human trafficking The partnership will remain strong if the current leader leaves

There is good communication among members There are difficulties sharing information among members about victims of trafficking due to confidentiality policies Conflict can arise among members because of the different agency/ organizational missions

There is low turnover of members in the partnership Partners compete for funding Member sometimes socialize together

The members of the partnership are able to resolve conflict effectively There is competition among members that work within the same communities

The partnership has achieved it annual goals over the past year
Uniqueness is generally considered to represent some level of measurement error as well as the communality of a variable in relation to a factor; uniqueness values above .6 (considered
68


high) are variables that are not closely correlated with the factors (StataCorp, 2012). Based on these criteria, factor 1 was further explored with five items and factor 2 with four items. Table 3.3 details the items that appear to load on factor 1, factor 2, or do not load on either.
Factor 1, titled Network Effectiveness as it appears to represent positive, whole network outcomes, explains approximately two-thirds or 68% of the variance in the matrix. Network dysfunction, factor 2, describes 35% of the variance in the test battery and the items that appear to load on this factor have lower communalities and higher uniqueness values. The less robust findings for factor 2 can be explored in future scholarship and for the remainder of this particular research, network effectiveness, factor 1, is the primary factor under consideration. Effectiveness more strongly demonstrates items that meet appropriate standards for inclusion based on theory, interitem correlations, and uniqueness values. Network Effectiveness
Network effectiveness, the central concept under consideration here, is worthy of further, detailed attention. Since prior research on network effectiveness primarily considers client or community level effectiveness (Provan & Milward, 1995, 2001; O’Toole, 2015; Turrini et al., 2010), the following results demonstrate how items load on factor 1, network effectiveness.
Table 3.4 highlights the results of a Chronbach alpha correlation test on the items that appear to load on network effectiveness. Based on the results included in Table 3.4, most of the item-rest correlations among the variables are close to each other meaning that the item correlates with the other items comprising the factor; however, the low turnover variable appears to be much lower than the other included items (Velicer & Fava, 1998). When this item is dropped from the factor, the average interitem correlation increases to .73 and the
69


overall alpha increases to .92.5.Generally values greater than .6 on the average interitem correlation are preferable and suggest a reliable indicator, but alphas above .8 are often acceptable (Statacorp, 2012). Therefore, dropping the turnover item may best be considered a theoretical choice for future research.
Table 3.4: Factor 1 (Network Effectiveness) Chronbach Alpha of Items
Item Obs Sign item-test correlation item-rest correlation average interitem correlation alpha
Trust 73 + 0.903 0.837 0.528 0.817
Good Communication 73 + 0.889 0.816 0.537 0.823
Low Turnover 73 + 0.602 0.408 0.733 0.917
Resolve Conflict 73 + 0.890 0.817 0.537 0.822
Achieved Past Goals 72 + 0.805 0.687 0.594 0.854
Test scale 0.586 0.876
Conclusion
Establishing the construct of network effectiveness and the construct of dysfunction are a useful advancement in the network effectiveness literature for several reasons. First, the two constructs may now be utilized to develop empirical variables. Now that these two distinct latent concepts have been identified through exploratory factor analysis, future studies can further develop the measurement of each factor through confirmatory factor analysis, and ultimately, structural equation modeling techniques can be utilized to explore relationships between contextual, managerial, functional, and structural network determinants that may lead to effectiveness.
Conceptually, the two factors reflect the effectiveness of the network as a combination of trust development, achieved shared goals, good communication, low
See Table 4: Chronbach’s Alpha Without Low Turnover in the Appendix for complete details.
70


membership turnover, and conflict resolution. Somewhat surprisingly, a change of leadership did not correlate with effectiveness. Leadership, and the turnover of leadership, may more effectively represent a determinant of effectiveness as hypothesized earlier rather than correlate as a function of effectiveness. Additionally, differences in organizational missions and the willingness of participants to socialize together do not appear to correlate strongly enough to comprise the network effectiveness factor either. Socializing, much like leadership change, may instead be predictors of effectiveness rather than a component. Mission differences that create conflict across organizations may fail to represent effectiveness or dysfunction for several reasons. A third (33%) of respondents reported that they were neutral about the statement that conflict can arise as a result of differing missions - it is not unrealistic to expect that respondents are neutral as they might be able to identify agencies within the network that make the network less effective as a result of differing missions as well as agencies who make the network more effective regardless of mission differences. Since it is currently unclear why so many respondent report neutral feelings, this item is not clearly an aspect of network effectiveness.
Future work may need to more closely connect network effectiveness to indicators that represent the outcomes of human trafficking networks actions. For example, only achieving past goals represents the ability of these networks to achieve community or client level outcomes. In order to determine how network effectiveness will actually decrease human trafficking, or directly influence the root causes of wicked problems, requires further operationalization of network effectiveness concepts and measures. Additional work is needed to support the construct validity of each factor as this will enhance and guide the development of empirical analysis of network-level effectiveness.
71


Provan and Kenis (2008) argued that trust and goal consensus appear to be important aspects related to network effectiveness. Additionally, Provan and Milward (2001) suggested that network level effectiveness must address the ability of organizations to work together, to fund and maintain information sharing as wells as member commitments to network goals. The results provided here suggest that trust, low turnover, achieving goals, and communicating effectively are all important dimensions of effectiveness. The addition of the concept of conflict resolution is new to the theoretical discussion of effectiveness. Little attention in current literature attends to conflict resolution developments across networks suggesting this is an area fruitful for further application and consideration. O’Toole (2015) asserted that developing opportunities to study networks as dependent and independent variables is a much needed development for the advancement of network theory. This study directly develops in that vein. Practically, the development of this network effectiveness variable will benefit grantors, grantees, community agencies, and may drive network administrative organizations to pay increasing attention to the items that comprise network effectiveness.
72


CHAPTER 4
PERCEPTIONS OF WHOLE NETWORKS: ATTEMPTS TO END WICKED
PROBLEMS
Abstract
Networks require collaboration from multiple partners; some partners may have goals that fail to effectively align with the mission of the network. Other agency members, however, may argue that a network is the only way to accomplish goals related to service provision, equity, or solving wicked problems. Therefore, how do participants define, describe and contextualize the networks to which they belong? Additionally, how do these perceptions shape the effectiveness and sustainability of the network? A qualitative analysis of 134 whole networks combating human trafficking revealed that networks are typically able to improve service provision, allow for communication throughout communities, increase education or awareness of a social issue, and produce policy/legislation. Network concerns included conflict, genuine commitment and the lack of funding for partnerships -while these coalitions may demonstrate that they can provide enhanced services, the network’s ability to achieve shared goals is often limited.
Practitioner Points
• Respondents suggest that conflict arises in networks and there appears to be two modes for conflict management - handling it openly and honestly with all partners or avoiding it altogether.
• Conflict in networks decreases the attainment of network goals and the sustainability of the partnership.
• Trust, a key to any organizational setting where teamwork is required, is built in three key ways: repeated interactions like regular meetings and communications, handling conflict, turnover, and leadership selection in an open and reflective way, and by working together to achieve shared goals.
• Granting agencies and umbrella non-profit organizations may opt to consider incentivizing network partnerships that support opportunities for open lines of
73


communication, developing conflict resolution skills in individual participants in the network, and promoting opportunities for creating shared goals across a community.
• Networks that form to achieve goals beyond information sharing and dissemination tend to have lower levels of perceived conflict and build trust more effectively.
Introduction
Networks, interdependent and structurally equal organizational arrangements comprised of two or more agencies seeking to achieve shared goals (Kilduff & Tsai, 2003; Meier, 2004; Provan & Kenis, 2008), are prevalent in communities across the globe. Agencies often recognize that a single entity is unable to meet the needs of clients in a community (Brummel, 2010; Siddiki, 2015). Therefore, these agencies seek out other resource providers, partners, or groups to help reach organizational metrics. These agencies may or may not need to rely upon the other organizations to reach their own goals and this distinction may have significant impact on their commitment, attention, effort, and communication with the network to which they belong. Community-level needs may fail to align with the focus of individual organizations. Or, the provision of services may vary based on the political will around a particular issue in the community to the detriment of other services. Exploring the ways in which organizational participants view the effectiveness, leadership, and sustainability of their networks may lend itself to more effectively understanding the nature of network-level leadership, outcomes from collaborative networks and possible community-level impacts.
Agencies also may enter a network due to a policy, which mandates membership or network composition (Saz-Carranza, Salvador Iborra, & Albareda, 2015). Lepore (2016), for example, argued that services for poor children like child protective services come as a result of a lack of service provision for poor women. Public attention and political will is invested
74


in a sprawling juvenile justice system and departments of child welfare while money spent preventing teen pregnancy, training and supporting young women from vulnerable backgrounds, or providing job placement opportunities may effectively support the wellbeing of America’s children. Community agencies attempting to decrease the prevalence of family violence or at-risk children may recognize this political reality and partner together to support women in vulnerable situations. The child welfare agency’s goals may be to evaluate and assess cases and complete risk analyses, for example, which can be empirically evaluated but fail to address the root cause of family violence. In order to systematically address the the problem, case workers and supervisors may recognize a need to work jointly with prevention partners for long-term, community-level results. These networks, seeking to end interpersonal violence, may have goals that diverge from those of the specific agency, but partnerships may be worthwhile nonetheless.
Preventing violence is one of many wicked problems facing communities across the globe. ‘Wicked problems’ are those that are contested, systemic, frustrating, and may require multiple, contextual solutions to end them (Head & Alford, 2015). Human trafficking is one example of a contemporary wicked problem with global implications. As borders become less defined, information about the quality of life in other places is more easily shared, and the pace of economic globalization grows, so does the demand for labor, personal safety, or freedom. Trafficking in humans increases in conflict areas and vulnerable populations are especially at risk. Trafficking spans various types of crimes and impacts adult men as well as women and children. Too often, bureaucracies may oversimplify the problem as a sex crime only affecting youth and children. Head (2008) suggested this simplification is a direct result of the growth in needing to establish targets and goals that can be met by bureaucrats and
75


managers in the public sector. Simplification allows for quantification of some aspects of this crime, but may lead to negative externalities for other potential victims and survivors of trafficking because the root cause may not be directly addressed. Wicked problems like human trafficking appear intractable because they are complex, require participation from multiple sectors, and the solutions may vary among communities.
Partnerships and Networks
The language of partnerships and networks is interchangeable in this study. Many terms in the public administration and management literatures appear to describe the network concept under consideration here. One indictment of the broadly and quickly expanding literature on networks is that the term ‘networks’ is increasingly nebulous, represents a different concept to different people, and is inconsistently applied (Borgatti & Foster, 2003). This section identifies a conceptual definition of networks and connects that concept to the operationalization of the concept in the survey from which the data for this project were drawn.
The notion that several organizations or agencies work together toward achieving a common goal is at the heart of understanding these network arrangements. O’Toole and Meier (2004) argued that networks are an interdependent group of multiple organizations or parts of organizations whereby the connections among the actors are not arranged as superior to each other. This definition suggests that members of the network are considered equal without a hierarchy of reporting. Provan and Kenis (2008) specifically defined networks as “groups of three or more legally autonomous organizations that work together to achieve not only their own goals but also a collective goal” (p. 231). Provan and Kenis (2008) also asserted that network configurations often require lead organizations or network
76


administrators to provide an effective governance mechanism for positive outcomes. The
structure of networks, leadership of the network, and goal-directed behavior are all
considered here. The Laboratory to Combat Human Trafficking, the community partner
connected to this project, defines ‘partnership’ as:
Partnership measures acknowledge that combating human trafficking requires a comprehensive response through the cooperation of multiple sectors.
Partnerships bring together diverse experiences, amplify messages, and leverage resources. For the purpose of this survey, an anti-human trafficking partnership refers to a ... [e.g. a cooperative relationship between two or more organizations established for the purpose of jointly combating human trafficking in some way.]
This language, while utilized specifically for the survey tool, connects to other definitions provided in the network literature and describes ways in which a group of organizations may band together to achieve goals.
More specifically, this research examines various whole public management networks. Public management networks include “agencies involved in public policy making and/or administrative structure through which public goods and services may be planned, designed, produced, and delivered (and any or all of the activities)” (McGuire & Agranoff, 2007). Agranoff and McGuire (2001) indicated that focused and continued efforts to study public management networks remain prevalent (O’Toole, 2014; Saz-Carranza, Salvador Iborra, & Albareda, 2015). The study of networks has arguably developed through three traditions: that of social network analysis in sociology, policy networks in political science, and service provision networks in public management (Berry et al., 2004). Berry and colleagues find contributions across all three traditions and suggest that overall structure matters when it comes to understanding networks, an individuals’ place and the centrality of an organization also matters within the network, and these structures and positions can
77


influence the effectiveness of the network itself (see also, Agranoff & McGuire, 2001; Ansell & Gash, 2008). Research on whole networks examines both independent and dependent variables from the network level; across many network traditions, few studies address whole networks empirically and the effectiveness of whole networks appears to be an area requiring additional attention and consideration (Provan, Fish, & Sydow, 2007).
This article examines whole networks seeking to end human trafficking. Respondents in the networks self-reported on the functioning and future of their own network. The central research question is: How do participants define, describe, and contextualize the networks to which they belong? Additionally, this research considers: How do these perceptions of the network shape the effectiveness and sustainability of the network? Two different datasets, collected from the same survey instrument are analyzed and discussed in the following sections.
Theory
4Ps: Prevention, Protection, Prosecution, and Partnership
The theoretical model under investigation here identifies four components of the movement necessary to end human trafficking: prevention, protection, prosecution, and partnership (4Ps). The concept of the 4Ps was developed by U.S. Department of State’s Office to Monitor and Combat Trafficking in Person’s and was independently validated through the Colorado Plan by the Laboratory to Combat Human Trafficking. In the initial version of federal legislation, the Trafficking Victims Protection Act of 2000, three Ps were identified. The fourth P, partnership, was added as a policy priority in 2010 in response to criticisms from recipients of Bureau of Justice Assistance grants and research organizations like the Laboratory to Combat Human Trafficking and Polaris Project (see also Sheldon-
78


Sherman, 2012). Governments and NGOs alike recognized that prosecuting the perpetrators, protecting victims and preventing the crime may fail to occur in practice without partnerships committed to information sharing, social justice, and community interventions.
Data
The data, provided by the Laboratory to Combat Human Trafficking (http://www.combathumantrafficking.org/), resulted from a multi-year planning and development process. The survey was designed to determine how many efforts are aimed at combating human trafficking, the extent of the success of those efforts, and to identification of promising practices conducted in communities across Colorado and the United States. One hundred and eighty-six respondents completed the national survey and 132 respondents completed the Colorado survey.
The survey instrument included 247 items and several respondents reported feeling survey fatigue as a result of the length. The partnership questions were the final section of the survey, which may mean that respondents who felt fatigued but may not have reported it in the final comment box of the partnership section, were systematically less likely to respond to that section. However, it should also be noted that many of the respondents, due to skip patterns built into the survey, only completed limited sections of the survey pertaining to their organizations focus areas. It remains unclear how this systematic bias may impact findings, but 71% of respondents completed item 191: Is your organization involved in a partnership effort to combat human trafficking (formal or informal)? and 59% of those respondents do not belonging to a partnership. In the national survey, 66% of respondents made it to question 191 and of those respondents, 22% report that they do not belong to a partnership. These descriptive results suggest that while many respondents failed to fully
79


complete the survey, there does not appear to be evidence to suggest the participants who reflected a predisposition to participate in networks versus those who did not.
This particular research utilizes the comment boxes provided for further discussion of the Likert style questions included in the partnership section of the survey. Fifteen questions related to network participation, formation, sustainability, mission, goals achieved, trust, leadership, member turnover, and others were analyzed. These items represent those most closely related to the research questions under consideration. All comments made by respondents to these items are included in the dataset for analysis.
Methodology
The data were examined using a major theme analysis to explore open-ended written responses. The qualitative responses offer an in-depth examination of the network frameworks. Additionally, the captured narratives offer rich quotes that exemplify emergent themes beyond standard Likert scales or other nominal survey responses. Initially, data analysis was conducted separately for state and national respondents. A subsequent analysis compared major themes between national and state responses. The data were analyzed using the holistic approach described by Saldana (2015). First, data were coded for words or short phrase. Second, patterns were identified and transformed into categories. Third, major themes were developed to organize groups of repeating ideas (Auerbach & Silverstein, 2003; Saldana, 2015). The analyses were conducted manually rather than by computer and two raters were used to assist in establishing inter-coder agreement. In the rare cases the coders disagreed the data were re-examined to arrive at consensus.
Results
80


After completing the national analysis then utilizing the state dataset for confirmatory evidence, one key result of the data findings suggests that national respondents and state respondents have slightly different perspectives on the work of their partnerships. While many partnerships report shared goals, reaching consensus on a mission statement, and having a vision for connecting the mission to goals, several networks also suggest that this work is still in progress or has yet to begin. The following sections describe key conceptual themes that arose from both the national and state datasets. The first section details the differences and similarities across both data sets, then the sets are considered together in the following sections. The main themes include: the 4Ps theory; network formation and sustainability; trust and conflict; and the tensions and negotiations that result for organizations partnering as a network. Participants appear to believe greater possibilities are possible when organizations come together to end wicked problems, but the goals achieved by the networks fail to meet expectations. Taken together these themes reinforce the value and utility of the 4P theory for collaborative action.
State and National Differences
The national data set findings indicate that most networks organize their goals around either a single construct or combination of protection, prosecution, and prevention. Many of the partnerships included in the national survey formed as a result of funding from Bureau of Justice Assistance and Office for Victims of Crime (BJA/OVC) task forces. Starting in 2004, funding for partnerships/task forces to combat human trafficking came in grants distributed to law enforcement agencies or victims’ service organizations. These funding streams were initiated in the Bureau of Justice Assistance (BJA) and funneled to victims’ services through the Office for Victims of Crime (OVC) or to law enforcement agencies through the United
81


States Attorney’s Office. Some communities received both funds while other task forces were created with either victim services or law enforcement as the grantee for the task force. Some task forces had mandates to provide comprehensive services while others were focused on providing services to more specialized populations or subsets of victims. These funded groups are referred to as Enhanced Collaborative Model Task Forces and they were meant to work toward combating all forms of human trafficking in collaborative ways. The need for funding often identified by respondents in the survey refers to the loss of the BJA/OVC monies in many regions. These grant monies fluctuate in communities making them unreliable and, for many leaders, makes the task force unsustainable over the long-term. Without the funding, one respondent noted, “unfortunately, dedicated funding is the best method of sustainability. The loss of funding for two dedicated leaders may see the demise of the coalition’s strong presence.” Another mentioned “funding is a big determining factor [of sustainability] but several partners are committed to this task regardless of funding.” These comments signal an important consideration related to funding and network sustainability, effectiveness, and durability in the national data set: networks that rely on grant funding or were formed by the BJA/OVC funding may not be as sustainable as networks formed through prior collaboration on harm reduction or shared service provision. Based on this evidence, it is unclear if government funding is necessarily a positive or negative factor for the success of networks.
State respondents focus more intensely on service provision for survivors as well as prevention in the form of awareness and education campaigns. The state under consideration had a rise in awareness of the crime due to Bureau of Justice Assistance funding to two different networks during the time of the survey. This situation may have caused the results
82


to vary somewhat from the overall national sample. The state respondents report that without funding, the sustainability of the networks decreased and note that the networks were experiencing rapid increases in membership. These structural changes lead to uncertainty about shared definitions of the wicked problem and increasingly diverse views of a solution to end trafficking. One respondent, for example, suggested that, “the core group has a good understanding of HT [human trafficking], but there are always so many new members that one cannot say that all members have a good understanding.” The positive outcomes related to increasing prevention may have the effect of making partnerships more challenging to manage. This finding is important for states and local communities across the country. Empirical findings suggest that as the size of networks increases, the ability of members to participate, engage, and act may decrease (La Due Lake & Huckfeldt, 1998; Provan & Kenis, 2008). Network size, leadership, duration, structure, and funding impact the effectiveness of networks (Provan & Milward, 1995). As awareness of the crime increases, seemingly rapidly in the national discourse as well, networks may require different management structures to remain viable and effective.
Comparing the national and state qualitative data show at the state level, networks tended to focus on prevention and protection through direct service provision or raising awareness in the community and the partnership serves the role of disseminating information. In the national survey, the purpose of partnerships appears to be more comprehensively focused on combating human trafficking. Responses common in the state survey include: “1) figure out coordinated efforts for services, 2) education, 3) provide coordinated efforts for services.” In national responses, the median response appears akin to “discussing trends, issues, and gaps in combating human trafficking statewide.” Both state and national
83


respondents report efforts to end trafficking attempt to be more coordinated, to more effectively deliver victim services, and to encourage prosecution. Overall, state respondents focus more on those efforts while the national survey respondents had a slightly broader approach.
Both national and state respondents reflect on similar frustrations faced by the network. A central theme of frustration was that the network serves only as an avenue for information dissemination rather than action. One state respondent stated, “we would have to be part of a real partnership not just publicizing activities of another organization” to have a sustainable partnership and another suggests that the partnership does not manage trust, instead “meetings are held but there is a sense that real information is not really being share.,” A national respondent noted the partnership is, “limited to information sharing, so there’s not much opportunity to actually collaborate on tasks.” While the genuineness of the partnerships might be questioned in some of the networks, respondents appeared to believe in an idealized form that was rarely achieved. Participants noted, for example, that “we’ve all made efforts to inform each other so there are no surprises, although there has sometimes been disappointment at the decisions made by one side or another.” While it is unclear from this comment alone what it means to be on one side or another, the respondent reported that the partnership fails to achieve the goal of communication sharing. Respondents stated that commitments from all agencies in the network, more variety of stakeholders, funding, and shared goals are required for sustainable and successful networks, but this was not always realized. Setting realistic goals and acknowledging the times when the network was less successful as expected appears to be an important factor for both trust building and managing conflict.
84


Prevention, Protection, Prosecution, and Partnership (the 4Ps)
Networks can be helpful for information dissemination and participants view the goals of the network as protection, prosecution, prevention, and partnership (4Ps). The 4P model provides a useful frame for understanding the work done by member organizations in the network, but the majority of networks lack comprehensive missions for achieving all 4 of the Ps. One respondent noted the purpose of their partnership is “to harmonize state and federal law enforcement, non-profits, city services, social services and the community in a holistic anti-trafficking and victim assistance effort.” Similarly, another stated that they seek to “build community and share information and resources as well as build bridges to know who is doing the work and how together we can help make a difference in our state.” Common themes included the topics of raising awareness, providing services to victims, developing educational and training materials, and coordinating response/prosecution in identified cases of trafficking. While these types of actions connect directly to the 4p model, there appears to be some missing (p)ieces of the model. Many of the respondents, for example, argued that identifying victims, securing funding, and establishing legislation are priorities for partnerships. Identifying victims, protection tasks, may also be conducted by prosecutors and law enforcement and, as such, might best be considered as a shared goal for partnerships. Simply because identifying victims fails to clearly fit as a particular P in the model may mean more specific goals are required of partnerships to achieve particularly if they receive OVC/BJA funding.
Network Formation and Sustainability
The partnerships seeking to end human trafficking under consideration formed under three conditions: community agencies recognizing a need, networks or partnerships that
85


already worked collaboratively in other capacities (i.e., homelessness, child protection, or
harm reduction), and funding from the federal grants. An interesting finding related to the
formation of the partnerships is that the respondents whose networks formed as a result of
federal funding reported that the sustainability of the partnership depended upon continued
funding. Respondents who had worked in other capacities share similar viewpoints:
Participants do more than just human trafficking. Trafficking, like poverty, substance abuse, mental illness, is just one more cause of homelessness that we must understand better to serve our constituents.
Compassion of the participating members and relationships that have formed in the process [contribute to sustainability]. We are non-funded, non-mandated.. .therefore, participation is based on value.
A history of 10+ years working together to serve victims of domestic violence. The fact that more survivors have been identified as a result of the collaboration and education.
Federal funding appears to make the long-term viability of the partnership more tenuous than if the partnership formed in either of the other two ways. Funding from BJA or OVC may also create an added layer of power inequity across partnerships between the agency directly receiving the funds and the others joining the task force. When asked about the leadership of the network, numerous respondents who received funding clearly identified that partner as the network leader and suggested that without continued funding the leadership change may result in the end of the partnership. The goals of the partnership may also be driven by the lead partner - the BJA/OVC partnerships are directly tied to achieving either the comprehensive or specialized services required in the grant whereas partnerships formed in response to community needs may be able to more effectively change direction, population served, or shift to or from prevention, prosecution, or protection as needed. This finding is somewhat contrary to Klijn and Koppenjan’s (2000) suggestion that government is an
86


effective leader of network arrangements due to high levels of trust, legitimacy, and unique resources. When the resources (grants) are no longer available, it appears that the legitimacy of a government network leader are unable to sustain the network and the resources may skew the network toward goals fail to match community needs.
Trust and Conflict
Developing trust and managing conflict sections in national and state responses highlight challenges faced by networks combating human trafficking. Trust is regularly cited as a necessary component of network management as it is indicated as contributing to network effectiveness and must be shared across organizations within networks not just among dyads (Agranoff & McGuire, 2001; McGuire, 2006; O’Toole, 1997; Provan & Kenis, 2008). Building trust was achieved in networks in three ways: regular meetings, effective communication, and shared goals with a consistent level of commitment from members. A lack of trust from members resulted when little attention was paid to creating an environment of honesty, openness, or relationship building. From both perspectives of building trust or lacking trust in the network, respondents commented that trust is achieved when “we have all been honest about our individual viewpoints and goals so that we can negotiate and conceive programming from a place of integrity and honesty.” Trust was lost when respondents view the network as insincere in collaborative efforts, unwilling to openly or honestly address conflict, varied viewpoints, or lack of action, and when “there is a sense that real information is not really being shared.” Trust in the networks had a clear and distinct connection to conflict in the network.
87


Table 4.1: Conflict in Network Arrangements
Managing Conflict Avoiding Conflict

The way the coalition manages conflict is by attacking the problematic issue honest and openly fsicl Overt conflict seems to be avoided, and left to fester. This affects the overall trust among members and the ability to be productive
Communications is the #1 tool in our success. It’s not about one person but the coalition as a whole Hasn’t been much overt conflict, probably because meetings don’t focus too much on real issues (cases, etc.). When there has been overt conflict, usually one of the parties involved simply stops participating
Good facilitation is priceless Don’t think conflict is managed - it is ignored
Open, fair, honest and respectful communication Avoid conflict - do not address conflict (which is a problem)
For the most part, the partnership has been able to work through conflicts by regular meetings and retreats Some individuals who were unable to work through the conflicts chose to dedicate their efforts in other arenas
The co-chairs do a good job of resolving conflicts Difficult issues are avoided
Direct communication with those involved in the conflict I feel that right now this [conflict] has not been addressed
Talk through it. Most members are committed to the cause and can get past philosophical differences
Address the issues openly in meetings
There could be more trust and openness in general, conflict is unpacked and managed reasonably
Conflicts in the networks were handled primarily in two ways: avoided or handled directly and openly. Table 4.1 highlights the responses across both national and state surveys discussing the patterns of response related to conflict. Several respondents suggested that when conflict becomes problematic in networks, those organizations, agencies, or individual representatives opt to leave the network. While this is not a large group of respondents, Hirschman’s (1970) notion of exit, voice, and loyalty appears to again provide meaningful options for network participants. Conflict resulting in exit may, however, mean that the ability of the network to provide services, reach diverse populations, or achieve other shared
88


goals decreases as a result of those exits. Voice, consistently discussed by those respondents who felt conflict was managed effectively, appears to be a necessary condition for conflict resolution. Voice, in conjunction with effective management practices like genuinely sharing important information, providing contexts for feedback, and effective facilitation, may be both necessary and sufficient conditions for effective conflict resolution in networks.
The responses in conflict management appear to reinforce the notion that trust and direct communication are directly related to the sustainability and effectiveness of networks. Sustainability seems to be impacted by the composition network membership. If agencies that disagree or have conflicts with the philosophy, goals, or actions of the network are essentially forced out as a result of the disagreement, the network may lose out on key actors necessary to achieve the goal of combating human trafficking. Additionally, decreasing network effectiveness occurring as a result of an inability to openly handle conflict will have real impacts on service provision to victims, community awareness and education programs, and the ability to prosecute criminals.
Tensions and Negotiations
Several underlying tensions appear to compete for attention in networks. These tensions include the trade-offs between the internal training, education, and information sharing needs of the network and the work of external facing programs, goals, and campaigns and focusing on proactive community work and reactive service provision to victims. These tensions seem to develop as a result of the 4P model. Prevention and protection activities, while seemingly complementary, both require an extensive amount of resources to be conducted effectively. Prevention, the activities related to education, training, awareness, advocacy, and community engagement are mentioned as both goals and activities of
89


networks as are protection goals like comprehensive victim services, referrals, responses to tips or calls, resources like shelters, counseling or education programs, and identification of survivors. Prosecution, while noted as part of the mission or goal of the network, appeared to be conducted apart from prevention and protection activities; this is unsurprising as the network likely had few members able to conduct investigations, charging, or prosecution. Network members reported that while the mission or shared goals of the network include work spanning the 4Ps, when they describe the most significant accomplishments of the network, the goals they achieve are in only one P. One network suggests this as the mission, “utilizing a victim-centered philosophy, to ensure the protection of victims, the prosecution of offenders, and the prevention of human trafficking and slavery through an effective, coordinated partnership,” and reports the significant achievements as “The partnership has strengthened coordination and communication, and has made the provision of services more seamless.” The mission was comprehensive yet the achievement focuses on the partnership and protection aspects. Another respondent describes the mission as, “through advocacy, partnership, and direct service, our mission is to abolish all forms of human trafficking and free victims from the grip of slavery,” and the goals achieved include “developing strong relationships with one another even when professional perspectives are different; Building awareness and advocating on behalf of human trafficking to the public, with one another, and with our respective colleagues.” Again, this network has a comprehensive mission but was only able to achieve goals primarily in partnership and prevention areas. The negotiations these networks face in prioritizing their work may be viewed as tradeoffs between external and internal facing trainings and programs - those aimed at prevention versus partnership. Additionally, these tradeoffs appear as proactive or reactive actions for ending trafficking -
90


protection versus prevention. Since many of these networks are newly formed it is likely that they felt demands to focus in one particular area of strength to conduct their work, but these tradeoffs demonstrate a tension that the partnerships will need to manage effective if the are to achieve their mission and reach their goals.
Discussion
Several key findings suggest implications for improving the capacity of networks to end wicked social problems. First, network leaders may want to consider the ways in which new members are introduced to network participation. Many respondents across both the national and state surveys suggested that awareness of trafficking and human slavery is growing, increasing the likelihood that additional members will be required to make a significant improvement in protection, prevention, or prosecution. New member trainings, particularly trainings that acquaint joiners with methods for conflict resolution, the mission and goals of the partnership, group norms and responsibility expectations, and the ways in which the group communicates appear particularly important for network effectiveness and sustainability.
Another consideration for network effectiveness includes funding structures. All nonprofit organizations highly prioritize funding, but network funding structures and opportunities appear more tenuous. While the Trafficking Victims Protection Reauthorization Act passed as an attachment to the 2013 Violence Against Women Act reauthorization (Public Law 113-4), and the legislation specifically stated that, “officials of the United States Government, shall promote, build, and sustain partnerships between the United States Government and private entities, including foundations, universities, corporations, community-based organizations, and other nongovernmental organizations,” the partnerships
91


Full Text

PAGE 1

! ! ! NETWORK APPROACHES TO WICKED PROBLEMS by ANNIE BLYTHE MILLER B.S., Colorado State University , 200 4 M.S., Miami University , 200 7 M.A., University of Colorado Boulder, 2010 A thesis submitted to the Faculty of the Graduate School of the University of C olorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Public Affairs Program 201 7

PAGE 2

! ii © 2017 ANNIE BLYTHE MILLER ALL RIGHTS RESERVED

PAGE 3

! iii This thesis for the Doctor of Philosophy degree by Annie Blythe M iller has been approved for the Public Affairs Program by Paul Teske , Chair Mary Dodge , Advisor AJ Alejano Steele John Ronquillo Date: May 13, 2017

PAGE 4

! iv Miller, Annie Blythe ( Ph.D , Public Affairs Program ) Network Approaches to Wicked Problems Thesi s directed by Professor Mary Dodge ABSTRACT Wicked problems are policy and social issues where policy makers and citizens agree that action must be taken but the solutions are contested. Wicked problems are often cross jurisdictional, connected to moral an d ethical public policy debates, and appear to require collaborative efforts to solve. This dissertation is a community based project that supports efforts by the Laboratory to Combat Human Trafficking (LCHT) to reach their mission by analyzing how network s (equivalent to partnerships conceptually) act to achieve effectiveness together. This research explores comparisons across 186 networks attempting to solve the same wicked problem, which may give researchers leverage to effectively understand many of the complexities of public management organizational arrangements. The first chapter provides historical analysis of the network and wicked problem literatures and the following chapters address different dimensions of networks and network effectiveness. The second chapter examines network membership and describes which organizations may need networks to accomplish their mission. Chapter 3 focuses on network effectiveness by testing a model proposed by Turrini et al. (20 10 ) , which suggests that structural, con textual, and network determinants lead to network effectiveness. The final chapter utilizes qualitative techniques to summarize and contextualize the network effectiveness findings. The form and content of this abstract are approved. I recommend its publ ication. Approved: Mary Dodge

PAGE 5

! v TABLE OF CONTENTS INTRODUCTION ................................ ................................ ................................ .................... 1 Network Approaches to Wicked Problems ................................ ................................ ........... 3 Human Trafficking: International Attention and U.S. Intervention ................................ ..... 6 Partnerships and Networks ................................ ................................ ................................ . 10 Networks and Effectiveness ................................ ................................ ................................ 12 Perceptions of Networks ................................ ................................ ................................ ..... 15 Research Orientation ................................ ................................ ................................ ........... 17 Partnership in Practice ................................ ................................ ................................ ........ 18 Data Collection and Research Design ................................ ................................ ................. 20 Limitations and Considerations ................................ ................................ .......................... 23 Organization and Structure ................................ ................................ ................................ . 24 Conclusions and Findings ................................ ................................ ................................ ... 26 JOINERS: A QCA ANALYSIS OF NECESSARY AND SUFFICIENT CONDITIONS LEADING TO ORGANIZATIONS JOINING NETWORKS ................................ ............... 27 Abstract ................................ ................................ ................................ ............................... 27 Introduction ................................ ................................ ................................ ......................... 27 Theory ................................ ................................ ................................ ................................ . 30 Data and Measures ................................ ................................ ................................ .............. 33 Methodology ................................ ................................ ................................ ....................... 35 Calibration ................................ ................................ ................................ ....................... 36 Publicness ................................ ................................ ................................ ....................... 37 Funding ................................ ................................ ................................ ........................... 39

PAGE 6

! vi Coproduction ................................ ................................ ................................ ................... 40 Theoreti cal Model ................................ ................................ ................................ ........... 42 Analysis ................................ ................................ ................................ ............................... 43 Necessary and Sufficient Conditions in Combatting Human Trafficking Networks ..... 43 Suf ficient Conditions ................................ ................................ ................................ ...... 45 Limited Diversity ................................ ................................ ................................ ............ 47 Missingness ................................ ................................ ................................ ..................... 48 Conclusion ................................ ................................ ................................ .......................... 50 NETWORK EFFECTIVENESS: COLLABORATING TO COMBAT HUMAN TRAFFICKING ................................ ................................ ................................ ...................... 52 Abstract ................................ ................................ ................................ ............................... 52 Introduction ................................ ................................ ................................ ......................... 52 Wicked Problems and Human Trafficking ................................ ................................ ..... 54 Whole Networks ................................ ................................ ................................ ............. 57 Theory ................................ ................................ ................................ ................................ . 59 Methodology ................................ ................................ ................................ ....................... 60 Factor Analysis ................................ ................................ ................................ ............... 61 Results ................................ ................................ ................................ ................................ . 66 Network Effectiveness ................................ ................................ ................................ .... 69 Conclusion ................................ ................................ ................................ .......................... 70 PERCEPTIONS OF WHOLE NETWORKS: ATTEMPTS TO END WICKED PROBLEMS ................................ ................................ ................................ ........................... 73 Abstract ................................ ................................ ................................ ............................... 73

PAGE 7

! vii Practitioner Points ................................ ................................ ................................ ........... 73 Introduction ................................ ................................ ................................ ......................... 74 Partnerships and Networks ................................ ................................ ................................ . 76 Theory ................................ ................................ ................................ ................................ . 78 4Ps: Prevention, Protection, Prosecution, and Partnership ................................ ............. 78 Data ................................ ................................ ................................ ................................ ..... 79 Methodology ................................ ................................ ................................ ....................... 80 Results ................................ ................................ ................................ ................................ . 80 State and National Differences ................................ ................................ ....................... 81 Prevention, Protection, Prosecution, and Partnership (the 4Ps) ................................ ...... 85 Network Formation and Sustainability ................................ ................................ ........... 85 Trust and Conflict ................................ ................................ ................................ ........... 87 Tensions and Negotiations ................................ ................................ .............................. 89 Discussion ................................ ................................ ................................ ........................... 91 Conclusion ................................ ................................ ................................ .......................... 92 REFERENCES ................................ ................................ ................................ ................... 95 APPENDIX A: Terms ................................ ................................ ................................ ....... 106 APPENDIX B: LCHT Survey Description ................................ ................................ ....... 108 APPENDIX C: Factor Analysis Robustness Details ................................ ........................ 109 Appendix C.1: Principle Factor Graphs ................................ ................................ ........ 109

PAGE 8

! viii LIST OF TABLES TABLE 1.2: Determinants of Whole Network Effectiveness ................................ .............................. 14 1.3: Data Collection and Analysis Phases ................................ ................................ ............... 21 2.1: Organizational Conditions Influencing Network Membership ................................ ....... 42 2.2: Theory, Coproduction, Publi cness and Funding Truth Table ................................ .......... 46 2.3: Theory, Coproduction, and Publicness Truth Table ................................ ........................ 49 3.1: Hypothesized Factor Loadings ................................ ................................ ........................ 63 3.2: Iterated Principle Factors Analysi s ................................ ................................ .................. 67 3.3: Iterated Principle Factor Items ................................ ................................ ......................... 68 4.1: Conflict in Network Arrangements ................................ ................................ .................. 88

PAGE 9

! ix LIST OF FIGURES FIGURE 3.1: Turrini et al. (2010) Network Eff ective ness Framework OverviewÉÉÉÉÉÉÉÉ. 59

PAGE 10

! " ! CHAPTER 1 INTRODUCTION Problem, Purpose , and Key Questions Networks often are implemented as a way to leverage multiple agencies or organizations to achieve greater success than a single entity. Networks, coalitions, and partnerships are not a particularly new phenomenon in the social sector. While the field of networks continues to receive attention, Provan and Milward (2001) called for greater attention on the study of network level outcomes. At its broadest level, this research is framed by the following research question: What factors contribute to network level effectiveness? First, out of concern for selection effects when studying networks, the research examines organizations that opt into networks and compares them to organizations that do not. Selection effects may guide positive or associative gains made by networks and this truncation may be leading researchers and practitioners to make inappropriate policy or resource allocation decisions in public service goods provision. The research focus in Chapter 2 is: What are the necessary and sufficient conditions leading to network membership ? Additionally, do organizations that fail to join networks have structural advantages that rarely require network membership? Af ter determining the factors that contribute to network membership, Chapter 3 explores determinants of network effectiveness. Conceptually network effectiveness is the ability of networks to achieve goals and make an impact on a social, political or cultura l issue . N etwork effectiveness , however, has yet to be operationalized and while theory exists to sugge s t which factors may contribute to effectiveness, no specific construct or measure for

PAGE 11

! # ! network effectiveness exists . Some of the hypothesized indicators of network effectiveness include trust, cooperation, goal achievement , stability, effective communication, conflict resolution, and network duration. Applying the unified framework for network effectiveness proposed by Turrini et al. (20 10 ), the chapter e stablishes two factors that appear to capture network effectiveness and network dysfunction. In order to reach a nuanced understanding of network effectiveness , Chapter 4 concludes this dissertation with a qualitative exploration of the survey data. Near ly every question in the network battery utilized in this research included comment boxes, many with rich and deeply informative narrative s about the state of the network. One respondent, for example, had this to say when asked about the sustainability of the partnership : "Goals set by members' individual agencies often determine/derail their goals in the partnership." Another respondent reported that the network allows them "to harmonize state and federal law enforcement, nonprofits, city services, social services, and the community in a holistic anti trafficking and victim assistance effort." The final chapter is a thematic and content analysis that provides rich description supporting the empirical findings of previous chapters. Exploring network function and dysfunction, operationalized as effectiveness, contribute s to the empirical study in the public management and network literatures. This dissertation represents more than scholarly contributions. The findings here have an immediate and direct impact o n the work done by the Laboratory to Combat Human Traffickin g . Additionally, the contribution goes beyond the human trafficking awareness movement; trafficking is one of many wicked problems that plague communities across the globe. Resolving conflict acti vely in partnerships, seeking out incentives for membership in network for organizations that may not perform activities related to prevention or protection

PAGE 12

! $ ! of survivors , and establishing rules and norms within networks that contribute to the identified in dictors of effectiveness will help improve the functioning of many types of partnerships . Th is chapter provides an introduction to wicked problems, network determinants, network effectiveness as well as the specific issue of human trafficking. Network App roaches to Wicked Problems Wicked problems are generally considered policy and social issues where policy makers and citizens agree that action must be taken but the solutions are contested. Citizens may disagree about the root causes or identification, an d the problem persists after numerous attempts to combat the issue. Human trafficking is one example of a contemporary wicked problem with global implications. As borders become more porous, information about the quality of life in other places is easily s hared, and the pace of economic globalization grows, so does the demand for labor, personal safety, or freedom. Trafficking in humans increases in conflict areas and vulnerable populations are especially at risk. Trafficking spans various types of crimes a nd impacts adult men as well as women and children. Too often, bureaucracies may oversimplify the problem as a sex crime only affecting youth and children. Head (2008) suggested this simplification is a direct result of the growth in needing to establish t argets and goals that can be met by bureaucrats and managers in the public sector. Simplification allows for quantification of some aspects of this crime, but may lead to negative externalities for other potential victims and survivors of trafficking. Wick ed problems like human trafficking appear intractable because they are complex, require participation from multiple sectors, and solutions vary among communities. Human trafficking is particularly problematic because combating it requires solutions, collab orations, and partnership across many levels of governance and jurisdictions. A truly

PAGE 13

! % ! wicked problem, according to Roberts (2000), is one where "nothing really bounds the problem solving process Ð it is experienced as ambiguous, fluid, complex, political, and frustrating as hell" (p. 2). Current efforts to end trafficking, particularly those funded by the United States government and the United Nations, have political will to end child sex trafficking. L abor trafficking, border smuggling, drug running, and other forms of coercion, however, while less flashy, also require time, attention, funding, and political will to achieve reduction and elimination of the problems . Trafficking prevalence shifts geographically as conflict situations, black markets, and lab or needs change globally. When a single municipality, county, state, or country decreases trafficking, the crime may easily shift into jurisdictions that are less vigilant just as a markets shift with changes in supply and demand. This situation further co mplicates problem identification, jurisdictional authority, and potential solutions. Ultimately, human trafficking is both wicked and frustrating. Many communities recognize the need to combat this problem, to provide attention and energy to the root cause s by finding ways to prevent the crime, protect survivors, and prosecute the offense. Networks are arrangements that allow organizations and individuals from different sectors, problem orientations, and service provision areas to come together and work to end wicked problems from multiple angles. Wicked problems are plagued with two types of uncertainty, cognitive and strategic, that lead agencies to take different approaches to potential solutions ( van Bueren, Klijn, & Koppenjan, 2003). Cognitive uncertai nties include a lack of scientific understanding about the root causes of a problem or an inability to identify the causal relationships leading to the problem. Strategic uncertainty occurs when there is disagreement on the solution of a wicked problem. St rategic uncertainty, particularly in networks, occur s because different actors may

PAGE 14

! & ! opt to take different approaches to problem resolution, have high transaction costs for information sharing, and disagree about the root causes as a result of cognitive unce rtainty. In the trafficking movement, for example, many agencies utilize a harm reduction model when confronting the issue. If an agency can reduce an individual's vulnerabilities by providing housing, workforce training, health care, food, or other recour ses, the supply of available humans vulnerable to traffick ing will be reduced. This supply side solution contrasts with demand side attempts to end trafficking. Agencies working on the demand side seek to increase penalties for johns, create educational pr ograms aimed at reshaping the thought patterns that lead pimps and johns to abuse people, and to shame individuals who use trafficked labor. It may appear unproblematic that agencies take each of these perspectives and work from both directions; however, o ne of the fundamental reasons that organizations adopt a network approach to solving wicked problems is to share scarce resources. The gains from the network can be lost as a result of the cognitive and strategic uncertainty. Networks often are identified as managerial and social arrangements that may help end some forms of wicked problems (van Bueren, Klijn, & Koppenjan. 2003). Trafficking, which is considered a form of slavery, occurs locally but ending this servitude requires collaborative, multi level, multi jurisdictional efforts. Networks, beyond markets and hierarchies, are an organizational form that can allow agencies to share scarce resources . Also, organizational networks can support employees and individuals to collaborate more effectively, may foster learning and knowledge, and provide efficiencies through interdependencies in supply chains (Podolny & Page, 1998; Powell, 2003). Network forms of organization have grown in prominence because they more effectively reflect empirical observations of the world. Borgatti and Foster (2003) describe d the prominence of networks

PAGE 15

! ' ! as "part of a general shift, beginning in the second half of the 20th century, away from individualist, essentialist and atomistic explanations toward more relational, contextual a nd systemic understandings" (p . 991). Organizations across the globe must learn how to work together, share resources and knowledge, and realize their connectedness across jurisdictional boundaries. The next section provides context and history to the wick ed problem of human trafficking. The remainder of the chapter discusses the dependent variable under consideration in this study, network effectiveness . Additionally, the chapter examines perceptions of networks and how they are utilized; focuses on the re search questions attended to throughout the dissertation; describes the research and methodological approach; and introduces the remaining chapters. Human Trafficking: International Attention and U.S. Intervention Fighting human trafficking, as a modern s ocial movement, began 15 years ago, but as awareness grows, it is clear this is not a new phenomenon. The United Nations recognized the need for the protection of individuals from force, fraud, or coercion many decades earlier. The Declaration of Human Rig hts (UN General Assembly, 1948), in articles three, four, and five, state that individuals should retain the rights to personal liberties, the right to life, freedom from slavery, security of their own person, and the ability to live free from inhumane, cr uel, or tortuous treatment. Trafficking is broadly considered an act where individuals are forced into some action, fraudulently led to behave or act in a particularly way, or coerced by another individual (Alejano Steele, 2013; Trafficking Victims Protect ion Act of 2000). Hart (2009) suggested that human trafficking is the modern term for any form of slavery. Many people think that human trafficking is primarily sex trafficking or coercion into sex acts; however, trafficking is not limited to prostitution or sex crimes and the victims

PAGE 16

! ( ! of this crime are not limited to children or women. According to Alejano Steele (2013), "although greater attention has been given to human trafficking as a global human rights violation that exploits women as commodities, the re continues to be a lag in applying these lenses domestically to our sisters and daughters in the United States" (p. 148). Many of the challenges associated with focusing on the issue domestically connect directly to the complex web of national legislatio n and jurisdictional authorities engaged in combating trafficking. The complexity of human trafficking as a crime warrants the involvement of multiple agencies, and highlights why network effectiveness and the ability of agencies to collaborate is the mean s to ending wicked social problems like trafficking. The Office to Monitor and Combat Trafficking in Persons, part of the U.S. Department of State, identifies several forms of servitude as trafficking including : debt bondage, forced labor, sex traffickin g, bonded labor, involuntary domestic servitude, child soldiers, forced child labor, and child sex trafficking. Early discourses related to trafficking of persons tended to focus primarily on trafficking children and women for sex (Gallagher, 2010). The Un ited States Congress, through passing the Trafficking Victims Protection Act of 2000, expanded the definition of trafficking to include labor, migrant trafficking, and noted that while trafficking often crosses boarders, the problem must be addressed both internationally and domestically. The laws aimed at combating human trafficking in the United States are a complex mix of Acts and Reauthorizations folded into several different pieces of legislation. Table 1 .1 provides an overview of each piece of legi slation and the major contribution made through adoption. In the United States, trafficking protections, in their most current form are a legal component of the Violence Against Women Reauthorization Act of 2013 ( VAWA,

PAGE 17

! ) ! Public Law 113 Ð 4) and The Trafficki ng Victims Protection Reauthorization Act , which passed as an attachment to the 2013 VAWA reauthorization. This legislation established protections such as temporary U Visas for survivors of trafficking in the United States. T his legislation as the title s uggests may not ultimately protect male labor and sex trafficking victims in the same capacity as protections offered for women and children. Table 1 .1 : United States Legislation to Combat Human Trafficking Legislation Year Action The Mann Act 1910, Reauthorize d 1978, 1986 Criminalizes sex trafficking especially as it relates to crossing state or international lines. The Tariff Act 1930 Goods cannot be imported if indentured laborers, forced laborers or slaves, made them. Racketeer Influenced and Corrupt Organizations Act (RICO) 1961 RICO gave law enforcement agencies a way to prosecute organized crimes especially groups organized to force individuals to participate in felonious activities or fraudulent practices. Trafficking Victims Prot ection Act 2000 The first comprehensive legislation on human trafficking that designated human trafficking as a federal offense. Established the first 3 Ps prevention, protection and prosecution and the Office to Monitor and Combat Trafficking in Persons (State Department). Created the T visa, a legal status that allows trafficking survivors to remain in the U.S. legally. Prosecutorial Remedies and Other Tools To End the Exploitation of Children Today (PROTECT) Act 2003 Created enhanced penalties for individuals engaged in tourism practices to participate in sexual acts with children. Also established the Amber Alert System to find lost, runaway, and kidnapped children.

PAGE 18

! * ! Table 1 .1 : United States Legislation to Combat Human Trafficking (Cont.) U.S. Leadership on HIV/AIDS, Tuberculosis, and Malaria Act 2003 Restricts ai d and international funding for countries and organizations that do not have specific anti trafficking policies. Trafficking Victims Protection Reauthorization Act 2005 Authorizes additional grant aid to combat trafficking and establishes a pilot program to shelter minors. Grant aid provided both domestically and internationally. Trafficking Victims Protection Reauthorization Act 2008 Expands the federal jurisdiction to prosecute trafficking and expands the definition of the crime of sex trafficking. Authorizes DOJ to create a model for state legislation. Customs and Facilitations and Trade Enforcement Act 2009 Prohibits the import of goods made by trafficked or indentured individuals. Trafficking Victims Protection Reaut horization Act 2013 Funds anti trafficking efforts until 2017 as an amendment to the reauthorized VAWA. Strengthens collaboration with state and local law enforcement. Seeks to prevent child marriage. National Defense Authorization Act 2013 Preven t and limit government contracts with agencies in foreign countries that may engage in trafficking. *Compiled on August 22nd, 2015 based on analysis from Polaris Project and the Office to Combat and Monitor Trafficking in Persons. Women and children are most vulnerable to sex and labor trafficking and represent approximately 54% of victims int ernationally (International Labour Organization, 2015); however, 46% of victims are male and may not receive appropriate interventions, support, or resources. Additionally, the Reauthorization Act specifically addresses international trafficking apart from trafficking in the United States. International trafficking efforts are

PAGE 19

! "+ ! centralized through the U.S. Department of State while the Department of Justice, Department of Health and Human Services, and Department of Labor distribute grants and oversee activi ties related to trafficking domestically. Many agencies and many agendas intersect to combat trafficking in the United States. Partnership formation is specifically encouraged as part of the 2013 VAWA reauthorization. The legislation specifically states: In coordination and cooperation with other officials at the Department of State, officials at the Department of Labor, and other relevant o fficials of the United States Government, shall promote, build, and sustain partnerships between the United States Government and private entities, including foundations, universities, corporations, community based organizations, and other nongovernmenta l organizations, to ensure that: (1) United States citizens do not use any item, product, or material produced or extracted with the use and labor from victims of severe forms of trafficking; and (2) such entities do not contribute to trafficking in person s involving sexual exploitation. The T rafficking V ictims A ct of 2000 was amended in the VAWA R 2013 A ct to reflect the need that partnerships, especially across the regional bureaus within the State Department set goals for decreasing trafficking in partn ership with international government partners and that the bureaus create metrics for assessing the ability of organizations to meet those goals. This dissertation addresses the effectiveness of partnerships across the United States and the State of Colora do to determine which characteristics, contexts, and functions lead to positive outcomes toward ending trafficking. Partnerships and Networks The language of partnerships and networks is interchangeable in this study. Many terms in the public administratio n and management literature appear s to describe the network concept under consideration here. One indictment of the broadly and quickly expanding literature on networks is that the term Ônetworks' is increasingly nebulous, represents a

PAGE 20

! "" ! different concept to different people, and is inconsistently applied (Borgatti & Foster, 2003). This section identifies the author's conceptual definition of networks and connects that concept to the operationalization of the concept in the survey from which data for this pro ject w ere drawn. The notion that several organizations or agencies work together toward achieving a common goal is the heart of understanding these network arrangements. O'Toole and Meier (2004) argued that networks are an interdependent group of multiple organizations or parts of organizations whereby the connections among the actors are not arranged as superior to each other. This definition suggests that members of the network are not subordinate or report to one another. Provan and Kenis (2008) specific ally defined networks as "groups of three or more legally autonomous organizations that work together to achieve not only their own goals but also a collective goal" (p 231). Provan and Kenis (2008) also argued that network configurations often require lea d organizations or network administrators to provide an effective governance mechanism for positive outcomes. The structure of networks, leadership of the network, and goal directed behavior are all considered here. The Laboratory to Combat Human Trafficki ng, the community partner connected to this project, defines Ôpartnership' research as: Partnership measures acknowledge that combating human trafficking requires a comprehensive response through the cooperation of multiple sectors. Partnerships bring toge ther diverse experiences, amplify messages, and leverage resources. For the purpose of this survey, an anti human trafficking partnership refers to a É [e.g. , a cooperative relationship between two or more organizations established for the purpose of joint ly combating human trafficking in some way.] This language, while utilized specifically for the survey tool, suggests ways in which a group of organizations may join together to achieve their stated goals.

PAGE 21

! "# ! More specifically, this study examines various "whole" public management networks. Public management networks can be identified as "agencies involved in public policy making and/or administrative structure through which public goods and services may be planned, designed, produced, and delivered (and an y or all of the activities)" (McGuire & Agranoff, 2007 , p. 1 ). Agranoff and McGuire (2001) first call ed for focused and continued efforts to study public management networks. This need for research was heralded by many scholars and the field has arguably developed through three traditions: that of social network analysis in sociology, policy networks in political science, and service provision networks in public management (Berry et al., 2004). Berry et al. (2004) find contributions across all three tradit ions and suggest that overall structure matters when it comes to understanding networks, an individuals' place and the centrality of an organization also matters within the network, and these structures and positions can influence the effectiveness of the network itself (see also, Agranoff & McGuire, 2001). This dissertation examines which organizations may opt into and out of networks combating human trafficking, identifies structural, contextual, and network determinants of effectiveness and ineffectiven ess. Additionally, the research explores comparisons across multiple networks attempting to solve the same wicked problem, which may give researchers leverage to more effectively understand many of the complexities of public management arrangements. This study focuses on the public administration and management network tradition and contributes to the past and current literature by employing an empirical study of whole networks that provide public goods to prevent human trafficking, protect survivors, and prosecute perpetrators. Networks and Effectiveness

PAGE 22

! "$ ! Networks form when an organization can gain from sharing resources (e.g., information and staff). Sharing resources means reaching new populations, providing resources to clients more effectively or mor e efficiently, partnering to expand services, or acquiring additional funds through mandated collaboration. Scholars agree that complex arrangements among individuals, organizations, and institutions complicate governing (Ansell & Gash, 2008; Provan & Keni s, 2008; Provan & Milward, 2001). Contributions that build network theory may provide greater insight into clari fying, understanding , and managing the complexity that arises because of networked collaborations and governance. Klijn (2005) specifically sugg ested that the network perspective "tries to develop concepts (networks, games, perceptions, etc.) and methods of analyzing (such as formal network analysis, explicating network rules and frames of reference) in order to clarify the complexity of the inter action processes and their outcomes (p. 266)." Berry et al. (2004) argued "the conceptual frameworks and key terms employed across the literature have created a complex and often confusing picture ( p. 539)." Following Berry and colleagues' suggestion that the field of study comes from advances in social network analysis methodology in sociology, studies in political science of policy networks, and examinations of public service provision in public management, this study deepens the literature in the third v ein. This dissertation examines network concepts like structure, function, and context to understand some of the complexity inherent to managing effective public service networks. Public service provision networks continue to expand. Q uestion on the util ity, purpose, and necessity of effectively understanding networks are fully recognized as essential . While , scholars continue to agree that while studying networks matters, there is still uncertainty about the functioning and effectiveness of whole network s (McGuire &

PAGE 23

! "% ! Agranoff, 2007; Provan & Lemaire, 2012; Raab, Mannak, & Cambre, 2015; Turrini et al. 2010). Scholars attribute this conceptual gap to a dearth of studies examining "whole" networks (Kilduff & Tsai, 2003; O'Toole, 2015; Provan, Fish, & Sydow, 2 007; Provan & Lemaire, 2012). Table 1.2 : Determinants of Whole Network Effectiveness Structural Characteristics Contextual Characteristics Network Functioning Composition: Prevent, protect and prosecute External to Network Managerial work Sector: Government, Non profit, Private Grant required partnership Training Members Size Resource Munificence Comprehensi ve Service Provision Formalized Mission or Vision Statement Voice: Includes survivor/ rep from vulnerable pop Goals Local: Includes local community member Perception of network leader Centra lization Broker / leader Provan and Lemaire (2012) argued, "whole network research shifts the focus from the ties that an actor has (an egocentric micro approach) and focuses instead on all of the ties among

PAGE 24

! "& ! a set of actors, which is a macro approach" (p. 639). The macro approach applied here looks at both national and state level networks that seek to end human trafficking. Of the limited studies examining whole networks, the definition of network effectiveness varies (Turrini et al., 2010). Three types of effectiveness: client level, community level, and network level we re identified by Provan and Milward (2001) as indicators of network success (Turrini et al., 2010). Network level effectiveness refers to the sustainability, legitimacy, and maintenance of the network. Ferlie and Pettigrew (1996) examined a single network for duration and effectiveness and several other studies have examined the ability of networks to achieve stated goals, innovate, and remain sustainable, but empirical work has failed to address all of these types of effectiveness along with indicators of ineffectiveness in whole networks. This dissertation builds the network effectiveness literature by applying network concepts like structure, context, and function drawn from Turrini et al.'s (2010) scholarship to determine the factors that lead to effec tiveness and ineffectiveness in whole networks (Provan & Milward, 1995; Turrini et al., 2010). Examining the composition and capacity of these networks may lead to understanding necessary and sufficient conditions for network level effectiveness as well as enhance understanding of how to create more effective network structures to overcome wicked social problems. Perceptions of Networks The increasing attention on the network organizing structure assumes that networks provide gains in service provision thr ough collaboration (Provan & Kenis, 2008) and that networks form through affinity, reputation, or similar goal orientation/attainment (Agranoff & McGuire, 2001; McPherson, Smith Lovin, & Cook, 2001). Although, networks may

PAGE 25

! "' ! actually work against the altruis tic aims of the organizations involved (O'Toole & Meier, 2004). Additionally, networks may be ineffective tools to provide services for trafficking and violence survivors as some evidence suggests that additional trauma and secondary victimization may occu r as a result of reporting incidents to multiple providers (Campbell et al., 1999; Orth, 2002). Another indictment against networks is that overcoming the transaction costs related to knowledge and resource sharing is immensely difficult (Weber & Khademian , 2008). Network managers and organizational leaders that act as collaborative capacity builders may be able to overcome this particular challenge. Chapter 4 explores how determinants like leadership and sustainability of the network may contribute to inef fectiveness. The networks under consideration in this study all seek to end human trafficking. Each organization that joins the network has its own mission, funding structure, and action targets. These organizations may find that joining a network fails t o contribute to their mission, may extend the organization beyond its capacity, or may increase administrative burdens on individuals who represent the organization. This dissertation takes two different approaches to understand what O'Toole and Meier (200 4) refer to as " the dark side " of networks. First, the research compares organizations in the movement who participate in networks with those that do not. Perhaps, organizations that have substantial resources and funding, jurisdictional advantages like pr osecutorial authority, and those already successful in providing victim advocacy in the local community may not need to leverage collaborations to effectively participate in the movement to end human trafficking. Work prior to Grannovetter's (1973) study on the power of weak ties in networks consider ed the negative implications of networks, however since the late 1970s, most

PAGE 26

! "( ! attention focuses on the neutral ties or positive gains made through networks (Labianca, 2014). While this study does not directly ad dress the negative ties between organizations in a network, it does consider which determinants lead to conflict, competition, or derision at the whole network level. This approach is an especially important consideration as organizations are incentivized through federal and state funds to collaborate, share resources, and jointly produce public goods. Research Orientation Different conceptions of mixed methods approaches abound (Creswell & Clark, 2008; Tashakkori & Teddlie, 2010). Tashakkori and Teddlie (2010) provide d perspective on an integrated framework and " meta inferences ." T his epistemology of mixed methods is the most coherent and relevant approach for understanding, describing, explaining, and predicting the success and failure of networks comba ting human trafficking. Truly mixed method research must be a sum greater than its qualitative and quantitative parts; more specifically, "such research can simultaneously address both exploratory and confirmatory questions, thereby gathering information t hat can result in Ômeta inferences' Éthat neither the quantitative nor qualitative perspectives could do alone" (Tashakkori & Teddlie, 2008, p . 101). The research design for this dissertation is an attempt to combine methodological approaches from various research traditions and to pay special attention to the consideration that through the partnership with LCHT this work also is intended to be action research (Lingard , Albert, & Levinson, 2008). The epistemological perspective of Guba and Lincoln (1989), w ho argued that starting first with research question s and then selecting appropriate methods , guides this thesis . T he following section describes the methodology and research design used in this work . The mixed method approach utilized here combines qualit ative

PAGE 27

! ") ! comparative analysis utilizing Boolean algebra , factor analysis, and content analysis to examine organizations' choices for network membership and effectiveness and ineffectiveness of networks working to solve a wicked problem. Partnership in Practi ce The Laboratory to Combat Human Trafficking (LCHT), a lead organization of a network (Provan & Kenis, 2008) based in Colorado, graciously agreed to provide data from a national and statewide survey effort as well as other valuable support. Comparing netw ork effectiveness was possible only through collaboration with their organization as they have the credibility necessary for collecting data on trafficking partnerships nationally and created the survey utilized here. As such, this work is a collaborative and community based project. Without the participation, contributions, and support of LCHT, this study would not be possible. Many of the contributions from this paper are realized through engaged scholarship (Boyer, 1996; Van de Ven, 2007; Hollander & Sal tmarsh, 2000). Engaged scholars seek to work jointly with community partners to achieve better understanding of the problems, assist in clarifying complexities, and create opportunities for social change. LCHT approached the researcher to assist in testing their theoretical model for combating trafficking. The theoretical model developed by LCHT identifies four components of the movement necessary to end human trafficking: prevention, protection, prosecution, and partnership (4Ps). The U.S. Department of State's Office to Monitor and Combat Trafficking in Person's adopted LCHT's 4Ps model, yet the model itself remains untested. The theory driving the hypotheses tested here is that the network composition of actors devoted to protection, prevention, and pro secution influence the ability of partnerships to achieve positive network Ð level outcomes (i.e., effectiveness). By testing the 4Ps model, this study

PAGE 28

! "* ! determines if organizing networks around these functions actually changes the effectiveness of networks combating trafficking across the United States. The data provided by LCHT for this project developed from a multi year planning and development process. The survey was designed to determine how many efforts are aimed at combating human trafficking, the ex tent of the success of those efforts, and to identify promising practices conducted in communities across Colorado and the United States. According the institutional review board submission completed by Alejano Steele (2012): Subjects are law enforcement, prosecutors, government officials, service providers, and community members involved in the anti trafficking movement throughout the United States. These subjects will be asked to take a survey to answer questions about the services available for people wh o have experienced human trafficking, the ways in which anti human trafficking efforts are approached by the criminal justice system, the various prevention efforts for human trafficking and the various partnerships that exist in the anti human traffickin g movement. All subjects will be sent a survey electronically through their email in which they will receive an explanation of the study and an invitation to participate in the study. In the email, they will be directed to a Zoomerang link for subjects to click on to initiate the survey. Service providers from Colorado will additionally complete the Community Needs Assessment Survey, which will also be distributed electronically. All subjects will be directed to a link to the survey in the program Zoomerang that will contain an informed consent. The sample size for this study is 1500. Subjects will be identified through literature reviews, website reviews of anti human trafficking agencies, suggestions made by pre eminent colleagues in the field and thro ugh recommendations from the Laboratory to Combat Human Trafficking's Advisory Board for the Colorado Project. Subjects will be sent an email explaining the study and inviting them to participate two weeks prior to the survey being distributed. They wil l be directed to a link to the survey in Zoomerang that will contain an informed consent. One hundred and eighty six respondents completed the national survey and 132 respondents completed the Colorado state survey. The survey included 247 items and seve ral respondents reported feeling survey fatigue as a result of the length. The partnership questions were the final section of the survey, which may mean that respondents who felt fatigued, but may have failed to report in

PAGE 29

! #+ ! the final comment box of the part nership section, were systematically less likely to respond to that section. H ow this systematic bias may impact findings is unclear , but 71 % of respondents complete d item 191: Is your organization involved in a partnership effort to combat human traffick ing (formal or informal)? O f those respondents 59% indicated they did not belonging to a partnership. In the national survey, 66 % of respondents ma d e it to question 191 and of those respondents, 22 % report ed that they do not belong to a partnership. These descriptive results suggest that while many respondents failed to fully complete the survey, there does not appear to be evidence to suggest the respondents who did complete the network batter reflect a predisposition or selection bias to participate in n etworks or not participate . Data Collection and Research Design At times, research design and methodology is its own form of a wicked problem. The solutions are diverse, contested, and often depend on the epistemological perspective of the researcher (Bra dy & Collier, 2010; Gerring, 2011; K ing, K eohane, & V erba , 1994). The research design proposed here is iterative and occur s in several different phases in order to draw as much leverage as possible on making inferences. King, Keohane, and Verba (1995) argu ed that the distinctions between qualitative and quantitative methods are nearly unimportant; instead, they suggest that the logic of inference ought to apply to all social science studies. Instead of applying a single logic , this research design employs t he logic of inference (K ing, Keohane, and Verba 1994) and the logic of inquiry (Gerring, 2011), to lead to sources of leverage on social science phenomena (Brady & Collier, 2010). In order to achieve the most robust findings, several phases of research are necessary. Table 1.3 identifies the multiple phases, describes the purpose of each, and identifies the method

PAGE 30

! #" ! utilized for that phase of the design. To achieve a mixed method research design, several phases of research were exploratory while others were c onfirmatory. Table 1.3 : Data Collection and Analysis Phases Research Stage Goal Method Phase 1: Specifying variables; Determining variance; Seeking additional data sources if necessary Create all independent and dependent variables; exploratory Exploratory factor Analysis (determine if the network quality battery in LCHT ' s survey matches hypothesized constructs); Collect all necessary data files from LCHT; Assess the size, quality, limitations, and strengths of data sources Phase 2: Test explor atory factor analysis from national data on state level data Determine validity and reliability of Factor Analysis of Dependent Variable; confirmatory Confirmatory factor analysis of national data Phase 3: Identify causal mechanisms Theory refinement; Exp loratory Data Analysis (EDA) Qualitative Comparative Analysis (QCA) Ð specifically, fuzzy set QCA for two different condition sets Phase 4 : Review potential specification problems, Check for validity, accuracy, robustness and misspecification; confirmato ry Kennedy ' s (2010) Diagnostics and Ten Commandments of Applied Econometrics Phase 5 : Compile all comments for qualitative analysis Provide evidentiary support for findings; exploratory and confirmatory Content analysis of items included as comment boxes in the partnership section of LCHT ' s survey; Connect themes drawn from quantitative chapters to improve robustness and validity Phase one of the proposed design is a factor analysis of the national survey. The researcher posits that by conducting testing and exploratory data analysis on the national survey data, the Colorado survey can serve as a confirmatory data set. King, Keohane, and

PAGE 31

! ## ! Verba (1994) suggested that theory development and testing should not occur on the same data so the methodology propose d here seeks to avoid this logic inference problem. In line with the dearth of literature on network ineffectiveness and the expansiveness of the literature on the organizational form of networks, a hypothesis was drawn for an exploratory factor analysis o f one of the batteries included in the partnership section of the survey. Table 1.4 details the proposed factor loadings tested in phase 1. Table 1.4 : Hypothesized Factor Loadings Effectiveness Dysfunction There is a great deal of trust among membe rs Conflict can arise among members because of the different agency/organizational missions Member sometimes socialize together Conflict arises among members because of competing definitions of human trafficking There is good communication among memb ers There is competition among members that work within the same communities The partnership will remain strong if the current leader leaves There are difficulties sharing information among members about victims of trafficking due to confidentiality p olicies There is low turnover of members in the partnership The members of the partnership are able to resolve conflict effectively The partnership has achieved its annual goals over the past year According to Dillon and Goldstein (1984) , expl ora tory factor analysis is a way of determining if items load on similar factors. The intent of phase one factor analysis is to determine if two underlying, or latent, concepts are present. Explanatory factor analysis is

PAGE 32

! #$ ! most commonly utilized when a r esearcher has no strong theoretical priors about how the data may perform around a particular number of factors (DeVellis, 2012; Dillion & Goldstein, 1984). As such, exploratory analysis provides a helpful mechanism to move forward in determining if these items, in fact, represent the two hypothesized concepts of network efficiency and inefficiency. After conducting the exploratory factor analysis, a confirmatory analysis on the Colorado survey w as conducted. The factors did indeed load similarly across th e two different sets of data. The final chapter, a robustness check on the thematic and theoretic chapters preceding, presents a content and thematic analysis of comments provided by respondents in the partnership section of the survey. Limitations and Co nsiderations The study of networks, especially the proposed confirmatory analysis proposed here, has two important limitations. First, the Turrini et al . (2010) model of network effectiveness represents a la r ge amalgamation of various types of indicators. The survey sent out to representatives of organizations in and out of networks does not contain measures or indicators of all the proposed determinants of network effectiveness in their suggested model. As such, several of the determinants are un explored i n this study. Future research in this area is necessary in order to fully test the Turrini et al. (2010) model. An additional consideration is missing data. There is " missingness " present across both surveys and throughout the independent and dependent va riables. Allison (2001) argued there are several ways to manage missing data including listwise and pairwise deletion, imputation, and dummy variable adjustment. In each of the following chapters, the missingness is handled in different ways that best addr ess the more specific research question

PAGE 33

! #% ! under consideration and the selected methodology. In chapter 2, missingness is most appropriately handled by listwis e deletion of cases. Since the Q uali tative Comparative A nalysis (QCA) requires the observation of co mbinations of conditions leading to network membership, imputation or other means of predicting values would be inappropriate. Logical remainders, the cases not empirically observed, are an important part of the QCA , which a lso makes imputation problematic . In chapter 3, missingness is also handled through casewise deletion . However, the missingn ess in this chapter is likely random as opposed to other chapters where datasets and variables are drawn from all four sections of the survey (where some consistent type of missing bias may be observed as respondents opt into and out of the additional sections of the survey) . Chapter 3 utilizes a single battery in the fourth (partnership) section of the survey. In this chapter, 71 cases had full information and were utilized for the factor analysis. Chapter 4 is the qualitative chapter and the researchers opted to examine all comment boxes completed by respondents. If a respondent felt strongly enough to comment, the researchers were compelled to analyze those respons es regardless of completion in other sections or on other items. Organization and Structure The format of the dissertation proposed here is slightly different from a traditionally structured dissertation. The " three article " format is gaining popularity , especially as the expectation for publications rise in the tenure track hiring market . This research follow s that format by structuring each chapter as a stand alone article ready for submission for peer review. This first chapter provide s historical and c ontextual analysis of network literature, and the following chapters address different dimensions of partnerships. The second chapter examines network membership and describes which organizations may need networks to

PAGE 34

! #& ! accomplish their mission and goals and why other organizations may not. This chapter is dedicated to assisting LCHT in their efforts to test the 4P model. Chapter 3 focus es on network effectiveness by testing a model proposed by Turrini et al. (20 10 ) , which suggests that structural, contextual , and network determinants lead to network effectiveness. The intention of the concluding chapter is to enhance conceptual understanding of the networks through the respondents' own words in a qualitative manner; this chapter effectively rounds out the prec eding quantitative chapters into a mixed method dissertation. All of the findings from each chapter support a broad interpretation of network approaches to wicked problems . These network approaches include the consideration of coproduction, the idea that organizations must create outputs directly related to the goals the network seeks to achieve if they are to join a network in the first place. Previous research had not fully illuminated this condition and it may enhance practical work of agencies and netw ork administrative organizations seeking to create effective networks. Second, dysfunction and effectiveness are unique and identifiable constructs at the network level. Effectiveness, comprised most significantly of trust, achieving shared goals, good com munication, and resolving conflict, appears to drive the ability of network level outcomes. Additional qualitative support for the need for conflict resolution, as well as the need for funding and genuine commitment, in networks comes to the front in Chapt er 4. Respondents supported the claims across the network literature suggesting that networks increase education or awareness of a social issue, improve service provision, allow for communication throughout communities, and can produce policy/legislation. These findings expand the network literature and highlight how examinations at the network level will contribute to more effectively understanding client, organization, and community level outcomes.

PAGE 35

! #' ! Conclusions and Findings The following three chapters c umulatively support the development of the network literature in several key ways. First , chapters 2 and 3 contribute to theory building. C hapter 2, the exploration of why organ izations join networks, findings indicate that organizations who need to coprod uce outputs tend to join networks. The condition of coproduction, working together to protect survivors or preventing the crime of human trafficking, is a sufficient condition for network membership. In chapter 3, a handful of key items lik e trust, achievi ng shared goals, good communication and resolving conflict comprise a factor termed network effectiveness. The items difficulty resolving conflict, competing for funding , difficulties sharing information , and competition among service providers in the sa me communities comprise a factor of network dysfunction. These findings suggest that additional empirical work on network effectiveness and dysfunction is now possible. The final chapter in this thesis details the role of conflict in network arrangements . This qualitative chapter includes findings that suggest conflict is particularly detrimental to the effectiveness of networks, and likely related to lack of information sharing, genuine trust, competition and poor communication. Network leaders ought to carefully consider the structures and tools available to partners for conflict resolution as well as techniques for engaging all partners in meaningful and goal oriented discussions. This thesis has policy implications for the formation and sustainability of networks, the management of partnership and network arrangements, and provides a methodological advance of a mu ch larger sample of networks.

PAGE 36

! #( ! CHAPTER 2 JOINERS: A QCA ANALYSIS OF NECESSARY AND SUFFICIENT CONDITIONS LEADING TO ORGANIZATIONS JOINING NETWORKS Abstract Networks achieve gains in service provision primarily through collaboration, strategic planning, and shared resources. This chapter compares multiple service provision networks, aimed at combating human trafficking, to determine how publicness, coproduction, theoretical models and funding influence network membership. Various agencies were surveyed about their efforts to end human trafficking. Respondents reported how the protection of survivors, prevention of trafficking, and pro secution of perpetrators occurs as well as their relationship to a partnership. B y utilizing fuzzy set logic in Qualitative Comparative A nalysis (QCA), aspects of an organization's responsiveness to public demands, theoretical orientation to ending traffi cking, funding, and service provision lead to network membership. Sufficiency conditions suggest that when an organization must coproduce service outcomes, they will join the network. An addit ional combination of conditions: not having a theoretical orient ation to the work, having funding from government grants, and not being constrained by responsiveness to public demands also leads to network membership. While this second combination is a logically feasible outcome, the coproduction pathway explains the m ajority of cases following a sufficient condition analysis. Coproduction is a necessary condition for organizations joining networks. Introduction Networks, an increasingly utilized organizational form, allow independent organizations to work jointly to achieve shared goals. These partnerships permit

PAGE 37

! #) ! organizations to specialize, share information or resources, expand services to fill provisioning gaps, and provide a supportive space to develop new practices aimed at solving wicked problems. This introduct ion describes how network participants view their roles, the challenges associated with networks, and the potential networks have for combating human trafficking. An underlying assumption in the network literature is that organizations that join networks t o produce goods and services achieve gains they could not by working independently. These shared gains produce positive externalities for clients, consumers, or customers. Why then, do some organizations opt out of networks? Why do organizations, especiall y those with high levels of specialization, join networks? Do the advantages of network membership typically outweigh all the additional managerial, communication, complexity , and coordination challenges? Do the resources, knowledge, and time required for effective networks come at a cost too high for smaller, sparsely funded organizations? This chapter explores these tensions and tradeoffs to determine some characteristics of organizations most likely to join networks. The networks under consideration here are all working to end human trafficking. The compositions of the networks vary as does the size, funding, leadership, and formality of the network. Many organizations within the networks describe the positive externalities attained by collaboration . Netw ork members argue that the partnership allows them "to harmonize state and federal law enforcement, nonprofits, city services, social services, and the community in a holistic anti trafficking and victim assistance effort." The partnership is also viewed as "collaborative efforts to enhance flow of information, identification of victims, rescue, service provision, investigations, prosecutions and follow up." These comments Ð drawn

PAGE 38

! #* ! directly from the survey Ð suggest networks and partnerships can create spac e for collaborative advances in the community as well as lead to enhanced service s for clients. Partnership members, however, also report challenges related to network membership. A respondent to the survey had this to say when asked about the sustainabil ity of the partnership, "goals set by members' individual agencies often determine/derail their goals in the partnership." Respondents also indicate that many of the networks form for reasons other than human trafficking. They cite domestic violence coalit ions, housing first programs, or prosecution networks as impetus for recognizing and combating human trafficking. This respondent, for example, reports that, "participants do more than just human trafficking. Trafficking, like poverty, substance abuse, me ntal illness, is just one more cause of homelessness that we must understand better in order to better serve our constituents." Based solely on this comment, the respondent's primary consideration is ending homelessness and he/she recognizes that trafficki ng contributes to the problem of homelessness. Many respondents note that their partnership either openly and honestly addressed conflict or ignored it to the determent of the partnership. Several respondents report that conflict is ignored, avoided or no t discussed; each of these respondents commented that this lack of confrontation appears problematic. Building trust also appeared to be an area where respondents reported some challenge. A respondent suggests, for example , that "I have not seen this addre ssed other than Ôwe have to trust one another' but it sounds like a directive, not followed up with substance." There are often tradeoffs and drawbacks to consider with any organizational form or method for service provision. This work attempts to understa nd how

PAGE 39

! $+ ! and why organizations may opt into collaborative efforts while other organizations do not. Clearly, the opportunities and challenges of network collaboration are complex. Theory There are several ways to conceptualize the formation, structure, an d outcomes of networks. Early conceptions and developments in the network literature focus on shifting the theoretic understandings of organizational arrangements. Many researchers asserted that actors beyond the firm and the government are responsible for providing goods and services in modern economies (Coase, 1937; Ostrom 1972; Williamson, 197 9 ). Third sector entities that manage public and common goods exist and often have important social functions. These functions, such as protecting fisheries, forest s, and access to nature; social service provision; and housing supports, are goods that may be required to end wicked problems. Networks provide a mechanism for the intersection of third sector, government , and private actors. Network participati on, format ion, and membership are most often discussed through the notion that network structure and governance mechanisms create positive externalities over markets, hierarchies, or the bureaucratic forms of organization (Coase, 1937; Jones, Hesterly & Borgatti, 19 97; Ostrom & Ostrom, 1971; Williamson, 1967, 1979). The gains typically associated with networks, like lowering transaction costs for collaboration, are those that allow members of the network to improve service, delivery, production, opportunity, or innov ation. Solving wicked problems, especially those like human trafficking, where traffickers can quickly move to different jurisdictions or transition from sex trafficking to labor trafficking, may require networks comprised of many agencies all working to e nd this crime. Networks, because they have the capacity to overcome collective action problems, are one

PAGE 40

! $" ! possible governance and organizational solution s that may increase the capacity of agencies seeking to combat trafficking (Kickert, Klijn, & Koppenjan, 1997; Powell, 2003). Additionally, networks can support information sharing, technical knowledge production, and service delivery. In the context of this study, networks, formed by two or more individual organizations, are a collaborative effort to achieve a shared goal or goals (Provan & Kenis, 2008). Berry, Brower, Choi, Goa, Jang, Kwon, & Word (2004) discussed three different views of networks. These three views, drawn primarily from sociology, political science, and public management disciplines, promo te networks as ways to understand relationships between and among organizations through social network analysis techniques, as groups of policy actors in various subsystems that act to change or amend public policy, and finally as groups coming together, p rimarily after the New Public Management movement, to directly provide public services to citizens. Lecy, Mergel, & Schmitz (2014) similarly viewed the network literature in public administration in three different veins. They argued that networks are conc eptualized and studied as policy formation, governance, or implementation networks. While it is immensely helpful to distinguish how scholars are using the language of networks ; organizations, especially those in the non profit sector, likely participate a s actors in advocacy, lobbying, service provision, and contractual partnerships with government agencies. The distinctions across network traditions are helpful for scholars, but practitioners may not view these concepts as separate types of networks. The public management vein, seemingly a combination of the public management and implementation theoretical constructs, most closely connects to the research question guiding this research, because it describes the ability of organizations to work together to achieve shared goals. This research

PAGE 41

! $# ! focuses on the following question: how do structural and contextual conditions influence network membership? Human trafficking, a contemporary term for slavery of any kind, requires multisector participation and colla boration if efforts to combat it are to be successful. However, it is unclear why some organizations may opt out of network membership. Organizations that can capture the full supply chain of the public good production process may not join networks since t hey do not require other organizations to effectively provide services. Additionally, networks where the success of the consumptive good does not depend upon the production or consumption of other goods may lead organizations to opt of networks (Katz & Sha piro, 1985). Additionally, Su‡rez (2011) argued that organizations that are mandated by grant requirements to form networks, often view these requirements as onerous, counterproductive, or as a dilution of funds that could make a much larger impact if spen t on one specific program or target. Non profit agencies therefore may remain conflicted about network membership. The need for funding may strongly outweigh any of the challenges associated with mandated partnership. An expansive literature exists that discusses the organizational form of networks, the rise of networks, and calls to further study networks (Agranoff & McGuire, 2004; Ferlie & Pettigrew, 1996; Kickert, Klijn & Koppenjan, 1997; McGuire & Agranoff, 2007; Podolny & Page, 1998; Provan & Kenis, 2008; Provan & Lemaire, 2012). There is a much smaller literature that raises concerns about the impacts of network membership for individual organizations (O'Toole, 1997, 2015; O'Toole & Meier, 2004; Raab & Milward , 2003 ). Not all organizations may share in the belief that network membership will allow them to more effectively achieve their goals. Additionally, the conventional assumption that New Public

PAGE 42

! $$ ! Management (Osborne, 1993) strategies lead government organizations to partner and contract may not alw ays ring true. In the realm of human trafficking, many of the key actors are prosecutorial or law enforcement. There is little reason to believe that law enforcement agencies will be incentivized to contract out their jurisdictional authority and the court s will not grant prosecutorial power to other entities. Without incentives to partner, will judicial or prosecutorial entities join networks? These organizations may experience pressure from local advocacy groups or crime victim service providers, but that influence may or may not be enough to drive police departments and district attorneys into joining. The following sections detail the data and measures as well as the QCA methodology utilized to explore the necessary and sufficient conditions that lead to network membership. Data and Measures This research is a collaborative effort with the Colorado based Laboratory to Combat Human Trafficking (LCHT). Community engaged research is both philosophical and participatory as the research seeks to produce knowle dge and insight in partnership with community based organizations and individuals (Anderson et al., 2012; Mikesell, Bromley, & Khodyakov, 2013 ). The goal of this research is to support LCHT's efforts to end human trafficking by identifying some of the cond itions that may lead organizations to join networks. As noted above, the network literature often argues that when organizations can join together they can reap greater collective benefits, improve service provision efforts to vulnerable community members, and enhance the quality of services provided through reduction in redundancies. With this research in mind, LCHT hopes to support projects that may make organizations more able to join networks. Through this research collaboration, this

PAGE 43

! $% ! chapter explores t he relationship between characteristics of an organization and the outcome of that organization opting to join a network (i.e. , network membership). LCHT created and administered a survey to determine the effectiveness and intention of work done national ly and locally to combat human trafficking. The survey was developed as a tool for communities to assess their own efforts to end trafficking and as a project that would systematically capture efforts across the country. Alejano Steele (2013) , co Founder a nd co Director of LCHT , suggests that: as the anti trafficking movement reflects upon 10 years since the passage of the federal Trafficking Victims Protection Act, we see a landscape of scattered efforts, frustrated communities that must cobble resources t o combat a vast and complex problem, victims falling through the cracks, and traffickers escaping punishment. Currently, there is no comprehensive process to gauge the number of anti trafficking efforts underway in the U.S., let alone how effective or succ essful those efforts are in preventing people from being trafficked. The survey was developed a s a tool to support other organizations and networks seeking to end human trafficking. The 4P model was utilized and the survey is structured around prevention, protection, prosecution, and partnerships. A survey description with further details is included in Appendix B . Network Membership One hundred and eighty six respondents completed the national survey. Each organization respondent was asked to answer the following question, " Is your organization involved in a partnership effort to combat human trafficking (formal or informal)?" This question elicited responses of either yes or no and the creation of the dependent variable is a dichotomous value of the cod ed response. Survey fatigue is a concern as several respondents noted that the length and detail of the survey proved challenging. The partnership question, described above is question 191 in the survey and is in the last section of the survey. Most

PAGE 44

! $& ! respon dents did not work through all 191 questions as there were significant skip patterns, but this is something to consider. Of the 186 respondents in the national survey, no information could be collected on nine cases. The respondents agreed to take the surv ey , but failed to complete any responses beyond consent. One hundred and twenty two respondents completed question 191 out of a possible 177. Just over 30 % of the respondents opted out of the partnership section in the survey. There does not appear to be a consistent selection bias on the types or characteristics of organizations that opt not to respond. In this data , this missingness will be characterized as missing at random and those cases will be considered list wise deletions in the case configurations (Allison, 2001). Methodology A qualitative comparative analysis (QCA) was selected for this exploratory data analysis (EDA) because it allows a researcher to view cases jointly and to further determine necessary and sufficient conditions for network mem bership (Ragin, 2008). QCA is a useful tool for exploring the research question posed here because QCA techniques can summarize data succinctly in the form of a truth table, check the coherence of data, check hypothes e s and other existing theories through necessary and sufficient condition testing, test conjectures made based on existing theories, and create opportunities for new theory development based on the results of the case configurations (Rihoux & Ragin, 2009). An example of a particular case utiliz ed here is a non profit organization that subscribes to a theoretical framework for combating trafficking, participates in protection activities but does not engage in advocacy, training, or prevention activities. There are many organizations that share th ese characteristics that responded to the survey. According to Miethe, Hart , and Regoeczi (2008), "through applications of EDA, researchers are often able to uncover underlying structures in

PAGE 45

! $' ! the data, extract trends among variables, detect outliers and oth er anomalies, and develop more parsimonious models" (p. 231). This study uncovers trends in network membership that allows the LCHT to better serve survivors, vulnerable individuals, and organizations working to end human trafficking. The intended goal of this methodology is to determine what conditions may be sufficient and/or necessary for an organization to opt into joining a human trafficking network. The QCA method functions much like a circuit board in a simple electronic device. For example, an orga nization that adheres to a theoretical model for their work has that particular switch in the on position. The most likely path to network membership may or may not be through that particular switch. Network membership, the outcome of interest in this pape r, happened or did not happen in that particular context. Different switch configurations from the conditions of interest may all contribute to membership being flipped on or off. In QCA, a type of set theory, the notion of an independent variable is repla ced with conditions that may be necessary or sufficient to lead to an outcome (Schneider & Wagemann, 2012). The process for determining which conditions lead to the network membership outcome is called calibration. In order to calibrate the set, the condit ions must be assigned values from 0 to 1. In fuzzy set QCA those values determine the level of membership that case has in that condition. Each condition are discussed here to outline the level of membership possible. Calibration Several different charac teristics of organizations may play a role in determining if that organization will opt to join a network. Both contextual and organizational characteristics are important considerations for network membership. Just as network effectiveness includes struct ural and contextual factors (Turrini et al . , 2010), both of these

PAGE 46

! $( ! factors are important considerations for individual organizations when making the calculation if joining the network will be worthwhile. The sector, focus area of the organization (prosecute , prevent, or protect), grant funded organization and a theoretic model are all included as possible attributes. The process of calibrating this fuzzy set follows the technique described by Schneider and Wageman (2012). Calibration requires detailed theor etical and content knowledge and the researcher must determine and fully outline the membership of each case in the condition. The section below highlights each condition and frames the theoretical expectation guiding the selection of that condition. The e nd of the calibration section describes other considerations related to the conditions and calibration technique. Publicness The New Public Management movement in public administration and management heralded the idea that contractual and delegational ar rangements between non profits, private sector entities, and governments would be a cost effective and efficient way to deliver public services (Ferlie, 1996; Hood, 1995). These arrangements also increase complexity, require more intentional communication efforts, and are often structured without effective oversight (Hood, 1991; Kaboolian, 1998; Lane, 2005). Networks providing services to combat human trafficking require shared goal setting, communication, collaboration, and structures that promote informat ion sharing. Organizations residing within the different sectors may have differing incentives to join networks. As previously mentioned, a governmental agency may have full jurisdiction to prosecute or to provide funding for a non profit organization to do the work of prevention; if that is the case, the incentive may not be strong enough to incur the costs of network

PAGE 47

! $) ! membership and will induce non participation. The survey respondents do not report why the y enter into the network and they likely may not know themselves all the reasons for those choices. In order to calibrate the sector condition, there must be a clear dimension on wh ether each case is either more in or more out in terms of members hip . There are a couple dimensions of sectors that may best illuminate the level of membership. The calibration considered here follows in a vein of public management and organizational leadership that views organizations having more and less responsibility to the public broadly. Bozeman (1987) argue d that all org anizations are public and have some level of responsibility to respond to, or at least attend to, political authority and the will of the people. In order to operationalize how the sector of the organization may matter for network membership, the degree of publicness w as assigned for each organization. Some organization types will likely be required to be more directly responsive to political authority or may be less responsive to political authority. Each of the organization types explored here, governmen tal, business, and non profit, all experience varied levels of publicness. The for profit organizations that contribute to this work in dynamic ways are less responsive to the public directly . T heir governance structures require no broad public oversight, and the CEO primarily determines the growth of the company. Businesses are considered entirely out of the public space and rarely respond to political will. Typically , government agencies ought to be viewed as the most responsive to public demands are they have the broadest permissiveness for action in the public sector and are asked to represent the community beyond single issue areas. As such, government agencies working on human trafficking are considered the most public organizations. However, there are some types of government agencies that tend to have little responsiveness

PAGE 48

! $* ! to public demand as they are rather removed from the electoral space and while they are providing a public service, they have strict legal mandates for their work Ð the prosecutoria l and judicial arenas of government. These prosecutorial and judicial organizations are calibrated as being mostly out of the publicness condition as they have little need to respond to the public broadly or to specific types if political will. Non profit organizations, governed by a board of directors and incorporated with a mission to do some kind of philanthropic or public service, are viewed as both public and somewhat responsive to political authority. Since non profits tend to narrow focus to specific types of work of action, they may have a clearly identified or deep commitment to a specific population within the community. This choice, or ability to determine their own notion of Ôpublic,' is viewed here as making them slightly less fully in the condi tion of publicness. Therefore the full condition of publicness is calibrated as government =1, non profit = .75, judicial = .25, and business = 0. Funding Some agencies combating human trafficking in the United States received federal grant money th rough various reauthorizations of the Victims of Trafficking Act of 2000. Additionally, several portions of the 2013 Reauthorization of the Violence Against Women Act (VAWA 2013) provide states, law enforcement agencies and non profit organizations financ ial grants to incentivize combating human trafficking. Some of this grant funding encouraged organizations to form or create networks. The funding provided to combat sex trafficking of minors, for example, authorized block grants to four communities where the population was greater than 5,000,000 people and included certain activities such as providing mental health counseling, shelter, and case management permissible under the grant.

PAGE 49

! %+ ! While this additional burden may place constraints on the agency and foc us their work in an unintended direction, the funding is essential, according to the respondents. As a result, funding for the partnership increase s the likelihood that members join partnerships. This result is derived from the emphasis agencies place on f unding that either carries some sort of encouraging language around network formation or the burdens of reporting and collecting data are best shared across providers doing similar community interventions. While the notion that an agency that receives fund ing will join the network is straightforward, it is worthwhile to explor e the combination of conditions that lead to joining a network when an agency is unfunded for partnership membership. Including this condition allows for combinations of conditions , wh ich may be leading to the outcome of interest. The calibration of this condition is also relatively straightforward. The organizations that receive funding for partnership receive a 1 as they are fully members of that condition. Organizations who do not r eceive funding retain the 0 label as they are fully excluded from that condition. Coproduction LCHT and the United States Department of State (State Department) utilize a framework for understanding the types of activities conducted by organizations seek ing to combat human trafficking. The 4Ps include protection, prosecution, prevention , and partnership as a comprehensive way to view the work they and other agencies conduct. The individual agencies in the sample agree d to complete four different survey s ections; as noted in the introduction, agencies who are able to achieve their goals without coproduction, that is without requiring a consumer, client, or another organizations to assist them in task, project or outcome completion, are likely more burdened by network membership. Coproduction can

PAGE 50

! %" ! be viewed as "an integrating mechanism and an incentive for resource mobilization," as well as place additional administrative burdens on the organization due to the commit ment s required for network membership (Bovi ard, 2007 , p. 846 ). Organizations engaged in coproduction or services may actually need other agencies to produce public goods, or, and perhaps equally meaningful, may feel pressured to join the network if it is a social norm in the community. The requirem ent that other agencies or individuals are required to produce the intended outputs of the organization can increase the expectation that an organization will need the network membership in order to achieve those goals. The result is that agencies may dete rmine that the costs (administrative burdens) of network membership are outweighed by the need to build relationships that will enhance the possibility that other actors will coproduce the desired outcomes. The 4Ps are a helpful tool for calibration when considering the case membership of individual agencies in regard to coproduction. Agencies who do not participate in prevention activit i es or provide services or support to survivors may have full capture of their intended outcome. These organizations are calibrated as 0 for coproduction. Organizations that participate in both prevention and protection are calibrated at 1. Organizations that participate in either prevention or protection are calibrated at .75 because they are more fully in the membership th an out of it. This is an important distinction in QCA as set membership does not allow for cases to remain at the midpoint; the midpoint is a theoretically useless distinction when determining necessary and sufficient conditions (Schneider & Wagemann, 2012 ). Protection and prevention activities tend to be those that require s some form of coproduction, therefore, those cases are closer to full membership in the set than out of it.

PAGE 51

! %# ! Theoretical Model Additionally, LCHT was interested in understanding if harm reduction, demand reduction, behavioral interventions or other types of social science models are articulated in the fight to end sex and labor trafficking. As such, this analysis includes the condition that the agency with a conceptual or theoretical mod el is worth further exploration. Each organization was offered three different types of theoretical models : behavioral, harm reduction, and demand reduction as options . They were also able to select that they do not utilize a theoretical model to guide th eir work. While there is not necessarily a strong theoretical prediction that would stem from an organization's interest in operating from a model and their willingness to join a network, it is a condition worth testing as it may operate in conjunction (as a necessary or sufficient condition) with other conditions more theoretically driven. Cases where respondents selected a theoretical model for combating human trafficking are listed as fully in the set members. The cases where organizations do not use a m odel are calibrated as zeros. Table 2.1 : Organizational Conditions Influencing Network Membership Independent Variable Concept Indicator Publicness Responsiveness to public demands 1 government, .75 non profit, .25 judicial, an d 0 business Funding Organizations have the capacity through federal funding to coordinate or provide services 1 full set membership, 0 no membership

PAGE 52

! %$ ! Table 2.1 : Organizational Conditions Influencing Network Membership (Cont.) Coproduction Organizations increase awareness, advocacy and education towards addressing a c ommunity's systemic vulnerability 1 full set membership, .75 partial set membership, 0 no set membership Theory Agency utilizes a theoretical model to drive their work. 1 full set membership, 0 no membership Analysis This chapter offers bet ter understanding of which organizations may opt into networks. Set theory (Miethe et al., 2008), a type of qualitative case analysis (Ragin, 2000), allows for determining which conditions are likely necessary or sufficient for joining networks. The notati ons of set theory follow from Boolean and fuzzy algebra and utilize logical AND/OR/NOT operations (Schneider & Wageman, 2012). One hundred and eighty six cases were included in the analysis. The researcher utilized the QCA and QCAGUI packages in R (Dusa, 2 007; Dusa & Thiem, 2015) to determine the logical minimization or the analysis of sufficiency (Schneider & Wageman, 2012). The empirical output of logical minimization is a truth table . Table 2.2 details each of the set relations that may lead to network m embership. Necessary and Sufficient Conditions in Combatting Human Trafficking Networks Ragin (2008) suggest ed that a Standard Analysis, one in which three different types of solutions to QCA are explored, consists of: (a) the complex solution, or what we prefer to call the conservative solution, where no assumption about any logical remainder is made; (b) the most parsimonious solution, which is based on simplifying assumptions, that is, those remainders are included into

PAGE 53

! %% ! the logical minimization that c ontribute to parsimony; and (c) the intermediate solution, which relies on easy counterfactuals, that is, only those simplifying assumptions are included that are in line with theory driven directional expectations. (Schneider & Wagemann, 2013, p. 211) Sc hneider and Wagemann (2012, p. 91 92) recommend ed conducting two analysis procedures in order to determine both necessary and sufficient conditions . This statistical procedure includes a top down analysis of sufficiency that produces the causal combinat ions of conditions leading to the outcome with a logical minimization process resulting in only the empirically verifiable combinations and a bottom up analysis of necessity that reveals the assertion that "a logical AND conjunction of two or more conditio ns can only be necessary for Y if, and only if, all single conditions involved in the conjunction are necessary on their own" (p. 92). However, this assumption of necessary conditions independently leading to the outcome is the focus of a new methodologica l debate. This debate in the analysis and description of findings in the QCA technique emerged between Thiem (2016) and Schneider and Wagemann (2012)'s Theory Guided Enhanced Standard Analysis (T/ESA) procedure. Thiem (2016) argue d that when applying T/E SA, utilizing a minimization procedure that distinguishes necessary conditions from sufficient in a two step process (Schneider & Wagemann, 2003; ~ 2012) is erroneous due to faulty assumptions related to the mathematical propositions of Boolean algebra and the combination of a researcher's application of causality to a phenomen on under study. Essentially, Th i e m (2016) suggest ed that the assertion from Schnieder and Wagemann (2012) claiming that the bottom up approach requires each "necessary condition also be an isolated cause of given effects" (p. 480) leads to a conservative solution not intermediate or parsimonious solutions. This result is problematic in the reporting of QCA analyses as it leaves the processes related to both conducting an analysis and r eporting results in a shared

PAGE 54

! %& ! language for peers problematic. As such, the following sections begin with an analysis of sufficient conditions, logical remainders, missingness and an additional robustness test on findings. A bottom up process for finding nec essary conditions was conducted, but no conditions in this dataset appeared to support any necessary condition 1 . As noted below, the presence of coproduction is a sufficient pathway to network membership but cases (very limited numbers) are present that su pport additional pathways to network membership. The empirical observations presented here suggest that the parsimonious and conservative solutions do not differ and presenting a n intermediate solution is not possible based on Thiem's (2016) arguments agai nst T/ESA. Sufficient Conditions The truth table below describes four condition s leading to network membership. F ollowin g the results of Table 2.2 , the primitive expression for all condition sets on the outcome of network membership results in the following: TCFP + TCF~P + TC~F~P + TC~FP + ~TCFP + ~TCF~P + ~TC~FP + ~TCF~P + ~T~CF~P ! Y The process of logical minimization re quires reducing these primitive conditions down to the subset relations between the condition and the outcome. In eight of the nine possible sufficient subset relations, Coproduction values greater than .05 produce the network membership outcome. Thus the subset relations can be logically minimized to: C + t Fp ! Y ! " #$%"&'("()*+,+-$."/01(,2$&+/3"/4"567869:5;<6="$1"&'("/3.>"*$&'?$>"-/31+1&(3&"?+&'"&'("/@&-/)("3(&?/,A")()0(,1 '+*B" 567869:5;<6="-/@.%"0("-/31+%(,(%"$"3(-(11$,>"-/3%+&+/3C"#/?(2(,B"&'("$%%+&+/3$."*$&'?$>"1@DD(1&1"&'$&"-/*,/%@-&+/3"+1"$"1@44+ -+(3&" *$&'?$>"?+&'"'+D'"-/2(,$D(C" "

PAGE 55

! %' ! Table 2.2 : Theory, Coproduction, Publicness and Funding Truth Table Conditions Network Membership Cases Frequency T C F P N 15,27,32,41,46,47,52,70,151,176,183 11 1 1 1 1 1 121,148 2 1 1 1 0 1 2,7,12,50,53,64,65,120 8 1 1 0 0 1 0 1 0 0 0 1 0 1 0 1 1 1 0 1 0 0 1 1 10,40,54,72,83,108,109,123,128,157,174,181 12 1 1 0 1 1 0 1 0 1 0 1 17,21,24,25,26 ,31,49,101,106,177 10 0 1 1 1 1 43,44 2 0 1 1 0 1 141,149,152,165 4 0 1 0 0 1 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 0 1 1 16,20,38,56,57,63,68,71,90,94,100,124,153,163, 179,182 16 0 1 0 1 1 133 1 0 0 1 0 1 Total Cases 66 The condition of the existence of Coproduction, that is, the organization is conducting prevention and/or protection activities , leads to network membership in 65 of the 66 cases. The single case of t Fp comes as a result of a grant writing firm that joined a network (possibly paid to write grants) and responded that they do not coproduce any services related to preventing trafficking or protecting survivors, they are not responsive to public deman ds as they are a business, and yet they do receive funds and seek funds for their work to combat human trafficking. While this is an interesting logical outcome, it is a rare condition leading to network membership. There are two parameters of fit in QCA : consistency and coverage. Consistency is the the way in which a researcher handles combinations of conditions that lead to both the

PAGE 56

! %( ! outcome and the absence of the outcome. In this particular empirical setting, after missingness is handle d across all four conditions, there are not cases observed where the outcome, network membership , is not present . This finding should not be interpreted to suggest that all cases with this combination of conditions always produces network membership; instead, the consisten cy measure is a value of 1 , meaning that all cases observed that lead to network membership fall into the two sufficient paths detailed above. The Coproduction path, has both high raw and unique coverage. Coverage, a measure of fit that explains the amount of overlap between the subset condition and the outcome of interest, can be explored through both raw coverage and unique coverage measures (Dusa & Theim, 2012; Schneider & Wagemann, 2012) . Coproduction, has a raw coverage of .883 and unique coverage of . 818. Coverage values range from 0 to 1. Clearly, a large amount of the outcome network membership occurs through the Coproduction pathway. The absence of theory and publicness with the presence of Funding (tFp), has a raw coverage of .080 and a unique cove rage of .015. As previously discussed, this path accounts for a very small subset of the outcome. Two individual cases cover both solution terms (cases 43 and 44) and are therefore calculated as contributing to raw coverage for each pathway. This reasserts the finding that tFp is likely a very rare sufficient pathway to network membership. Limited Diversity Limited diversity occurs when logical remainders are present during the analysis (Schneider & Wagemann, 2012). Logical remainders, possible subset rel ations that are not empirically observed in the data, are nonetheless important considerations for researchers. I n this analysis, for example, the logical remainder tcFP would be an organization that fails to subscribe to a particularly theory for the work , does not do protection of survivors or

PAGE 57

! %) ! prevention of the crime , but does receive funding and is responsible to the public. As noted in the introduction, many partners in the human trafficking prevention and reduction space are focused on other types of c rime or vulnerabilities in the community. A non profit day shelter, for example, may be funded through Office of Victim's Assistance and Bureau of Justice Affairs funds, may be required to join a network, and yet may not provide any direct services or effo rts against trafficking. While this scenario was not empirically observed, the pathway to network membership may still be present . A n argument could be made that t F p is a logical remainder itself as there is only a singe case observed . The definition of l ogical remainders from Schneider and Wagemann (2012) is "truth table rows that lack enough empirical evidence to be subjected to a test of sufficiency" (p. 152). Since 66 cases were examined, the concept of arithmetic remainders (too few cases for a single case in all possible combinations) would neither apply nor would the scenario of this being an impossible remainder as it was empirically observed. Clearly , there is no clear guidance on how much empirical evidence is enough to support the existence of th e sufficient pathway. Typically, one single case is enough and as such, tcFP remains included in this analysis as a sufficient pathway to network membership. Missingness Missingness is an additional consideration necessary for a complete analysis of the conditions leading to network membership. Missingness does not appear to be missingness at random (Allison, 2001) on the funding variable. Of the 120 cases dropped from the analysis, over half (63), were missing on both the funding condition and the n etwork membership outcome. The funding variable is the only condition drawn from the partnership section of the survey. The partnership section was the final section (of four total) and it appears that

PAGE 58

! %* ! respondents may have suffered from survey fatigue by t he time they reached question number 191 (the first question of the fourth section). As such, the researcher opted to additionally conduct a second QCA analysis , which dropped this funding variable to compare the results and to create a robustness check on the set relations found in the prior section. While the presence of funding and the non presence of theory, coproduction , and publicness did result in a case empirically, that particular case was a unique condition in the data. Table 2.3 details 93 cases and provides support for two findings in the full four conditions of Table 2.2 (above). Table 2.3 : Theory, Coproduction , and Publicness Truth Table Conditions Network Membership Cases Frequency T C P N 4,10,15,27,32,37,40,41,46,47,5 2,54,58,7 0,71,72,83,108,109,123,126,128,151,15 7,161,167,174,176,179,181,183,186 32 1 1 1 1 2,7,12,50,53,64,65,120,121,148,178 9 1 1 0 1 0 1 0 1 1 0 1 0 0 1 8,14,16,17,19,20,21,24,25,26,31,35,363 8,49,56,57,63,68,75,90,94,97,100,101,1 06,115,118,124 ,125,135,140,143,153,16 3,169,173,175,177,182 40 0 1 1 1 43,44,119,141,149,152,165, 7 0 1 0 1 42,60,87 3 0 0 1 1 111,133 2 0 0 0 1 Total Cases 93 The logi cal minimization of the Table 2.3 : Three Conditions results in the equation: C + tc > Y Coproduction is again found to be a sufficient condition to network membership with a much larger number of cases. Of 93 cases, 88 cases were observed with the presence of

PAGE 59

! &+ ! coproduction leading to network membership. Another sufficient pathway, n o coproduction and no theory (regardless of the presence of absence of publicness), represents five cases; tc has much smaller coverage but again meets the empirical threshold for consideration as a sufficient pathway. No theory coupled with no coproductio n is interesting for interpretation. These organizations may seemingly have no direct motivation for direct membership but each of the five cases here all work to end trafficking through other avenues of support. The case observed and mentioned above is a grant writer, another makes art to raise social justice awareness, yet another views themselves as a funding pass through for special services and the last two work on policy and advocacy for women broadly and intersect with trafficking through gender base d violence reduction. These five cases may suggest motivation for networks to consider partners not directly connected to the explicit mission but those that can provide additional viewpoints, contribute to effectiveness through other avenues, or simply be cause the organization is interested in supporting the collaborative efforts. Conclusion Coproduction appears to be a n important condition for network membership. Network membership is substantially driven by organizations that work toward efforts pro tecting survivors and/or preventing trafficking. These organizations may come to the work with diverse theoretical orientations , from public or private organizations and may or may not lack fund ing to combat trafficking. In many ways , this is an incredibly refreshing and powerful finding. Perhaps this is truly a necessary condition Ð organizations that want to work toward multiple different avenues of end ing trafficking tend to believe that they cannot do that work alone. This finding breathes more life int o the interpretations that while networks may have constraints , limitations, and costs, organizations that are engaged in

PAGE 60

! &" ! service provision recognize that those tradeoffs are worth the time, resources, and positive externalities gained from joining network s. Future studies should further explore and understand the relationships among sufficient pathways to not joining networks. The full analysis of all four conditions did not reveal any pathways to non membership; however, the robustness check of dropping funding, suggests that there are five sufficient pathways to nonmembership. These are worthy of additional consideration in future papers. Prosecution agencies and judicial government agencies are not clearly implicated as joining or not joining networks a s discussed previously as a possible pathway to membership. The notion of responsiveness to the public does not directly create a subset relationship with network membership but it may still be implicated in sufficient pathways to non membership. Further e xamination may provide additional insights into network arrangements that support and do not support agencies with jurisdictional authority or issue capture. Network effectiveness, the ability of networks to sustain collective performance, trust each othe r, work together, share information, and achieve goals, may be impacted by the different pathways to membership. The networks that contain organizations from pathways achieved not through Coproduction may enhance the network effectiveness through additiona l and diverse perspectives or may inhibit and slow networks through tangential missions, lack of attention to issues at the core of trafficking, or perceptions that preventing other forms of violence are more important. Understanding the more complex arran gements of networks and the formation and sustainability of those networks is key to enhancing effectiveness, partnerships seeking to end wicked problems, and the complexities of working together at the organizational level.

PAGE 61

! &# ! C HAPTER 3 NETWORK EFFECTIVENESS: COLLABORATING TO COMBAT HUMAN TRAFFICKING Abstract Networks, collaborations of more than three agencies working to achieve shared goals, hope to attain collective gains through partnerships, sharing resources, and/or providing more comprehensive services. The goal of many of these networks is to tackle wicked problems, those too complex, systemic, or challenging to end without a broader effort. There is still little empirical evidence that these effi ciencies are achieved as often as touted or that network governance is effectively structured to provide public goods and services. Network effectiveness, the ability of multiple organizations to trust each other, work together, share information, socializ e, achieve goals, and resolve conflict is examined across 186 networks that are all working to combat human trafficking. Ending human trafficking, a form of slavery that preys on vulnerable people living with in institutions of inequality, will require effe ctive collaborative governance. Sampling networks seeking to achieve similar goals allows for comparison where similar contextual and structural challenges likely exist. Through factor analysis, two distinct concepts emerged from the data: effectiveness an d dysfunction. The specific items that comprise these factors are useful for inclusion in future research, for more deeply understanding the differences among client, community, and network level effectiveness, and encourag ing network formation and mainten ance that attends to both increasing effectiveness and decreasing dysfunction. Introduction

PAGE 62

! &$ ! Foundations, governments, private entities and non profits are increasingly developing networks that improve service delivery, make a collective impact greater than they could do alone, and share resources (e.g., scarce funding). As agencies increase their interdependencies, they also increase the layers of bureaucracy, administration, and authority. Due to these increasing cognitive challenges related to additio nal layers of personnel as well as competing strategic goals for ending wicked problems, the ability of the networks to achieve goals, resolve conflict, or develop trust may suffer. Some networks seem to overcome challenges more effectively than others or manage to find clarity where uncertainty previously ruled. By developing and testing a theory of network effectiveness, this study enhances understanding of the management and administration of networks. The vast literature on networks can be understoo d as developing in three research traditions and represent conceptually different ways to understand the study of networks. The three traditions of network research : sociology, political science, and public management produced innovations in social network analysis, policy innovation, change, and agenda setting, and development of managerial structures and behaviors as well as network performance and outcomes (Berry et al., 2004). These traditions built on each other, intertwined, and provided powerful insi ghts like the concept of "structural holes" (Burt, 1992), density (Marsden, 1990), capacity to end wicked problems through partnerships (van Bueren, Klijn, & Koppenjan. 2003; Head, 2008), and the capacity for policy innovation and knowledge transfer across a network (Rhodes, 1997; Robinson, 2006; Weber & Khademian, 2008). Through a process of tracing citation relationships, Lecy, Mergel, and Schmitz (2014) argue d that scholars do a poor job of defining the way in which they conceptualize the network under c onsideration, the research tradition influencing the work, or the research

PAGE 63

! &% ! program. The literature tracing process led th e se scholars to structure the public administration literature on networks in three conceptual areas: policy formation, governance , and implementation (Lecy, Mergel, & Schmitz, 2014). The major development provided in this research is in the implementation of networks vein and focuses specifically on comparing the capacity of whole networks to solve wicked problems. Wicked Problems and H uman Trafficking Poverty, human rights abuses, unemployment, homelessness, and violence prevention are all forms of wicked problems. These problems share the attributes of being sustained in communities over long periods of time, having multiple and compe ting definitions, needing public and administrative attention, and requiring solutions that may be contested by different segments of the population (i.e., the solutions are often politicized). These issues plague society because of the immense cognitive a nd strategic uncertainty that surround both the problem and the solutions (van Bueren, Klijn, & Koppenjan. 2003). Roberts (2000) provide d a conceptual distinction around various types of problems. The first type, simple problems, may initially present a c hallenge, but when expertise is applied there is a clear definition of the problem and systematic ways to solve the issue. Complex problems present themselves as ones where there is consensus on what constitutes the problem. Citizens and policy makers can agree that the status quo is the problem and that something needs to be done. Complex problems, however, have multiple solutions. The final type of public problems is considered wicked because both the identification and possible solutions are contested. C itizens and government officials disagree about the root causes or identification, and the problem persists after numerous attempts to combat the issue.

PAGE 64

! && ! Human trafficking is one example of a contemporary wicked problem with global implications. As borders become more porous, information about the quality of life in other places is easily shared, and as the pace of economic globalization grows so does the demand for labor, personal safety, or freedom. Trafficking in humans increases in conflict areas and vu lnerable populations are especially at risk. Trafficking spans various types of crimes and impacts adult men as well as women and children. Too often, bureaucracies may oversimplify the problem as a sex crime only affecting youth and children. Head (2008) suggest ed t his simplification is a direct result of the growth in needing to establish targets and goals that can be met by bureaucrats and managers in the public sector. Simplification allows for quantification of some aspects of this crime, but may lead to negative externalities for other potential victims and survivors of trafficking. Wicked problems like human trafficking appear intractable because they are complex, require participation from multiple sectors, and solutions vary among communities. Human trafficking represents both a criminal offense (federal statute created by the Trafficking Victims Protection Act of 2000) and a violation of an individual's human rights. Trafficking is "defined by the fact that victims" agency and other human rights are violated through force, fraud, or coercion" (Foot, 2016, p. 2). This violation primarily comes as a result of some form of vulnerability. Women and children, the most frequent victims of trafficking, are typically more marginalized in societies across the globe or more likely to be perpetrated against (Roberts, 2008). Men, especially those who feel they must provide for themselves and their families are more frequently trafficked to provide labor. While vulnerable populations are at greater risk to be forc ed, coerced, or fraudulently deceived, trafficking is also viewed today as any modern form of slavery. The root causes of

PAGE 65

! &' ! contemporary slavery are debatable and may vary depending upon the state, region, or country. This complexity leads to the cognitive a nd strategic uncertainty that defines wicked problems. Human trafficking is particularly problematic because combating it require s solutions, collaborations, and partnerships across many levels of governance and jurisdictions. A truly wicked problem, ac cording to Roberts (2000), is one where "nothing really bounds the problem solving process Ð it is experienced as ambiguous, fluid, complex, political, and frustrating as hell" (p. 2). Current efforts to end trafficking, especially those funded by the Unit ed States government and the United Nations, have political will to end child sex trafficking. Labor trafficking, border smuggling, drug running, and other forms of coercion, while less flashy also require time, attention, funding, and political will if th ey are to end. Trafficking prevalence shifts geographically as conflict situations, black markets, and labor needs change globally. When a single municipality, county, state, or country decreases trafficking, the crime may easily shift into jurisdictions t hat are more permissive just as a markets shift with changes in supply and demand. This situation further complicates problem identification, jurisdictional authority, and potential solutions. Ultimately, human trafficking is both wicked and frustrating as hell. Many communities recognize the need to combat this problem, to provide attention and energy to the root causes by finding ways to prevent the crime, protect survivors, and prosecute the offense. Combating the multiple causes and outcomes related to trafficking also demands effective communication across agencies and governments, timely information sharing, and prevention, protection, and prosecution efforts that build on collaboration among all parties. Networks are arrangements that allow

PAGE 66

! &( ! organizati ons and individuals from different sectors, problem orientations, and service provision areas to come together and work to end wicked problems from multiple angles. Human trafficking represents an example of what Roberts (2000) describe d as a truly wicked problem. She identifies three possible strategies for coping with wicked problems: authoritative, competitive, and collaborative. The focus of this research is to examine how whole networks function in collaboration to produce outcomes aimed at ending hum an trafficking. Whole Networks Since networks have the collaborative capacity to combat wicked problems, scholars typically view them through a positive lens (Ansell & Gash, 2008; Klijn & Koppenjan, 2000; Koliba, Meek, & Zia, 2010; Meier & O'Toole, 2003; Padgett & Ansell, 1993; Provan & Milward, 1995). One problematic aspect of the broadly and quickly expanding literature on networks is that the term "networks" is increasingly nebulous, represents a different concept to different people, and is inconsisten tly applied (Borgatti & Foster, 2003). This section identifies a conceptual definition of networks and connects that concept to the operationalization of networks in the survey from which the data for this project w ere drawn. The notion that several organi zations or agencies work together toward achieving a common goal is the heart of understanding these network arrangements. O'Toole and Meier (2004) argued that networks are an interdependent group of multiple organizations or parts of organizations whereby the connections among the actors are not arranged as superior to each other. This definition suggests that members of the network are neither subordinate nor report to one another. Provan and Kenis (2008) specifically defined networks as "groups of three or more legally autonomous organizations that work together to achieve not only their own

PAGE 67

! &) ! goals but also a collective goal" (p . 231). Provan and Kenis (2008) also argued that network configurations often require lead organizations or network administrators to provide an effective governance mechanism for positive outcomes. The structure of networks, leadership of the network, and goal directed behavior are all considered here. Specifically, this article addresses the research questions from the "network as a form of governance' approach as opposed to a "network analytical" approach (Provan & Kenis, 2008). The network as governance focus pays special attention to meta level characteristics of the network. The network leadership, funding, structure, stability, and size are all important considerations in this perspective. The Laboratory to Combat Human Trafficking, the community partner connected to this project, defines partnership research as: Partnership measures acknowledge that combating human trafficking requires a comprehensive response through the cooperation of multiple sectors. Partnerships bring together diverse experiences, amplify messages, and leverage resources. For the purpose of this survey, an anti human trafficking partnership refers to a É [e .g., a cooperative relationship between two or more organizations established for the purpose of jointly combating human trafficking in some way.] This language, while utilized specifically for the survey tool, suggests ways in which a group of organizati ons may collaborate to achieve their stated goals related to human trafficking. More specifically, this study examines various "whole" public management networks. Public management networks include "agencies involved in public policy making and/or admini strative structure through which public goods and services may be planned, designed, produced, and delivered (and any or all of the activities)" (McGuire & Agranoff, 2007). O'Toole (1997) and Agranoff and McGuire (2001) call ed for focused and continued ef forts to study public management networks. While other networks are informal, lack funding mechanisms for combating trafficking, and have little regulatory authority to prosecute or

PAGE 68

! &* ! criminalize trafficking in their local communities. This article addresses the following research question: How can we measure and understand network level effectiveness? Theory A comprehensive theory on network effectiveness is currently nonexistent (Turrini et al., 2010; O'Toole, 2015). However, based on previous work related to network effectiveness (Provan & Milward, 1995; Provan & Sebastian, 1998; Turrini et al., 2010) the structural, functional, managerial and contextual descriptions of networks relate to the performance of networks (Klijn & Koppenjan, 2000; Provan & Milwa rd, 2000; Klijn, 2005; Provan & Kenis, 2008) . T he goal of this research is to explore the relationships among indicators of network performance. The most comprehensive attempt to identify a framework for understanding network effectiveness comes from Turri ni et al. (2010). They posit the following: Figure 3.1 : Turrini et al. (2010) Network Effectiveness Framework Overview (© Public Administration , Wiley Online Library) The Turrini et al. (2010) framework is a way to conceptualize the determinants of netw ork effectiveness.

PAGE 69

! '+ ! Of the limited studies examining whole networks, the definition of network effectiveness varies (Turrini et al., 2010). Provan and Milward (2001) identified three types of effectiveness: client level, community level, and network level as indicators of network success. Network l evel effectiveness refers to the sustainability, legitimacy, and maintenance of the network. Ferlie and Pettigrew (1996) examined a single network for duration and effectiveness and a handful of other studies hav e examined the ability of networks to achieve stated goals, innovate, and remain sustainable, but empirical work has failed to address all of these types of effectiveness. The conceptual definition of network effectiveness utilized for this study is the ne twork level capacity of the actors to work together to achieve goals related to combating human trafficking. Achieving goals is a key tenant in the definition, though working together may not be as easily operationalized. Working together represents the ne twork's ability to communicate, build trust, and resolve conflict. Clearly, effectiveness differs from efficiency and utility. Simon (1965), with interpretation from Baldwin (1985), clarified the notion that efficiency and utility imply a relationship bet ween costs and outputs whereas effectiveness is solely a conception of the benefits, outputs, or outcomes of the work. The current research analysis attends to the observable outcomes of working together and achieving goals, not the costs imposed on the ac tors to achieve those outcomes. Future research will begin to address the relationships among the hypothesized determinants and network performance, but first, an empirical operationalization of network effectiveness is necessary. The following section di scusses the results of a factor analysis of items collected from a national survey of networks seeking to combat human trafficking. Methodology

PAGE 70

! '" ! In 2013, organizations involved in combat ing human trafficking completed a survey conducted by a self identifi ed, organization that seeks to end th is wicked social problem. The Laboratory to Combat Human Trafficking (LCHT) supports research, response, and reporting efforts in Colorado and works closely with organizations across the state developed to prevent the c rime, protect survivors, and prosecute perpetrators. Two surveys were administered: one was a national sample and the other w as distributed to organizations within a single state. 2 Since the surveys were administered simultaneously, this study utilizes onl y national survey data as a sample of the population of networks. The national survey is larger, likely more representative of unique, whole networks, and covers much broader geographic areas. The national respondents were selected through purposive and co nvenience sampling, which lead to a sample of 453 organizations that received the survey. One hundred and eighty six organizations returned surveys and the executive or program director in the agency completed the survey the majority of the time (55 % ). Fur ther details about the methodology of the survey or the types of organizations included in the sample is available at http://coloradoproject.combathumantraff icking.org/resultsandfindings/nationalreport . The following section details the use of factor analysis to determine how the items in the survey represent network effectiveness. Factor Analysis Factor analysis, a technique in which a researcher must deter mine how to rotate and group items tapping a single latent construct (DeVellis, 2012), was utilized to explore survey items believed to influence the functioning, performance, and overall effectiveness of the networks. The survey contained a question batte ry asking the respondent to consider several E " 5/)*.(&("$--(11"&/"1@,2(>"F@(1&+/31B")(&'/%/./D>B"$3%"1$)*.+3D"-$3"0("4/@3%"$&" '&&*GHH-/. /,$%/*,/I(-&C-/)0$&'@)$3&,$44+-A+3DC/,DH,(1@.&1$3%4+3%+3D1H3$&+/3$.,(*/,& "

PAGE 71

! '# ! statements related to the partnership they opted to describe. Respondents who belong to multiple partnerships were asked to think of the most diverse partnership in terms of sector membership. Also, each agency was allowed to complete the battery multiple times for each partnership ; however, only the most diverse partnership is included in this data set. The battery includes items on a five point Likert scale related to statements that refer to the partnership. O ne item states, for example, "there is a great deal of trust among members" and the respondent selected strongly disagree, disagree, neutral, agree, or strongly agree. 3 A total of 1 1 statements were included in the battery (see Table 3.1 for specific surve y items). A factor analysis was conducted to determine if these items m ight effectively represent fewer, underlying constructs (Jolliffe, 2002; Brown, 2015). Based on the statements, generally positively or negatively related to the functioning of the net work, the researcher hypothesized that there would be two distinct factors and that exploratory factor analysis (EFA) through principle axis factoring and rotation is most appropriate. As there is little empirical evidence or measures currently in use for network effectiveness, this approach is a n exploratory factor analysis rather than confirmatory. Brown (2015) suggest ed that confirmatory factor analysis (CFA) is only appropriate when there are specific theoretical priors for the communality estimate, the structure of the variables, matrices and variance, and how the indicators are related to the factor or construct. Exploratory factor analysis is the most appropriate technique based on the state of theory on network effectiveness and the oft unreliable na ture of confirmatory factor analysis ( Borsboom, 2006). 3 A quick note on the missingness in this battery. There are 186 respondents for the larger survey and 122 report on membership in a network. Of those 122, 95 respondents report that they are, indeed, part of a network combatting human trafficking. A factor analysis program will generally use casewise deletion leading to a total N value of 71 for the factor analysis. "

PAGE 72

! '$ ! Specification of the exploratory factor analysis is arguably one of the most important considerations when attempting to identify the factors that represent suspected latent concepts ( Fabrigar et al., 1999; Preacher & MacCallum, 2003). These considerations include selecting the matrix of coefficients to be analyzed, the number of factors to be extracted, the extraction method, if the factors should be rotated, and the interpretation of the resulting fa ctor scores (Thompson, 2004). The following section outlines each of those choice points. Since the statements are generally written in positive terms or negative terms, a theoretical assumption is that the positive items will likely be correlated and the more negative statements will likely also have a correlated relationship. Additionally, trust, communication, and achieving goals together are implicated as ways in which networks can be viewed as effective (Provan & Milward, 2000; Provan & Kenis, 2008; Tu rrini et al., 20 10 ). Table 3.1 shows the expected factor loadings from the survey items : one that is the latent representation of effectiveness and representation of network dysfunction: Table 3.1 : Hypothesized Factor Loadings Effectiveness Dysfunction There is a great deal of trust among members Conflict can arise among members because of the different agency/organizational missions Member sometimes socialize together Conflict arises among members because of competing definitions of human traf ficking There is good communication among members There is competition among members that work within the same communities The partnership will remain strong if the current leader leaves There are difficulties sharing information among members about victims of trafficking due to confidentiality policies

PAGE 73

! '% ! Table 3.1 : Hypothesized Factor Loadings (Cont.) Effectiveness Dysfunction There is low turnover of members in the partnership The members of the partnership are able to resolve conflict effectively The partnership has achieved it annual goals over the past year Several different techniques for extracting the factors are available to researchers looking to explore the relationship between multiple variables, which may have a relationship to an underlying construct. Generally, researchers would consider princip le component analysis and common factor analysis (Galbraith, Moustaki, Bartholomew , & Steele, 2002; Jolliffe, 2002; Warner, 2008) Principle component analysis (PCA) is a sum of the variance of the variables Ð the factors are created directly from the origi nal data and represent essentially a sum of variance from the variables included in the factor under consideration (Galbraith, Moustaki, Bartholomew , & Steele, 2002; Warner, 2002). Fabrigar et al. (1999) caution ed that researchers may fail to realize that PCA is a data reduction technique and fails to account for the individual variance and error terms of each variable in the factor analysis. P rinciple factor analysis or common factors, however, is a multiple regression technique that utilizes the shared va riance of a set of variables to determine a communality estimate . T his estimate is the variance that is predicted as a result of the correlation among the variables that comprise the factor (Warner, 2002). This communality value is replaced in the R matrix and several statistical packages can run this estimate over numerous iterations to improve the accuracy and reliability of the estimate (Fabrigar & Wegener, 2012). Another distinction between PCA and the common factoring reported here

PAGE 74

! '& ! is that PCA does not distinguish between common and unique variance. Common factoring is preferable here because it is likely that trust, communication, socializing, and other items that may compose one of the factors lack similar error terms or distributions (Fabrigar & Wege ner, 2012). Therefore, the common factoring is utilized here. Both maximum likelihood extraction and the principle factors extraction techniques estimate parameters of unique variance and factor loadings (common factoring), but maximum likelihood (ML) tech niques allow for statistical testing of correlations and factor loadings ( Fabrigar, Wegener, MacCallum, & Strahan, 1999). Therefore, ML estimations must assume that the included variables are normally distributed and do not violate other normality assumpti ons. Principle factors allow for non normality, but are often a challenge to report and compare the reliability or validity of results (Preacher & MacCallum, 2003). Principle factoring can be iterative; the process utilizes the sample correlation matrix in itially then estimates new communality estimates in the matrix until the process reaches convergence (Preacher & MacCallum, 2003). Based on findings for exploratory factor analysis, only PAF is utilized here as it appears to outperform ML and inferential s tatistics are unnecessary (or reliable) when conducting exploratory analysis (Costello & Osborne, 2009; De Winter & Dodou, 2012). A decision of rotating the axis or not must also be considered (Fabrigar & Wegener, 2012). Rotating the factor analysis allows for the axis of each factor to shift in order to describe a more elegant and simple analysis for interpretation (Fabrigar & Wegener, 2012; StataCorp, 2012). Orthogonal rotation ensures the assumption that the factors that are uncorrelated remain in place , whereas oblique rotation allows for factors to violate that assumption and be correlated. Since this is exploratory factor analysis, oblique rotation was

PAGE 75

! '' ! utilized. There is no evidence to suggest that each of the items in the battery are uncorrelated, an d if fact, it is likely that the ability to resolve conflict effectively is somewhat correlated to the relevance of conflicts that arise out of mission differences or from varying definitions of human trafficking. Additionally, not only could the items be correlated, the dysfunction of networks, one of the hypothesized factors may or not may not be correlated with the effectiveness of networks. Theoretically, a network may remain effective and at the same time have dysfunction. For example, it is not unreal istic to expect that network members may disagree on the definitions of the wicked problem, to have conflicts as a result of differing missions or to have challenges related to sharing information and yet, with dysfunction present, the network will still h ave some level of effectiveness that can be measured and evaluated. The dysfunctions in the network may or may not have a relationship with the effectiveness of the network and further studies will need to explore the relationship between network determina nts, dysfunction, and effectiveness. As such, oblique rotation is the most appropriate selection. Results In order to utilize the iterative option of principle factors, the first step was to examine non iterative models to determine that two factors were likely present. Current theory supports this hypothesis, but it was also necessary to test. In Kaiser criterion, screeplot and parallel analysis, two factors were the best fit. The eigenvalues reported below support both a scree test and parallel analyses of the iterative principle factor extraction. 4 Typically eigenvalues greater than one would suggest that a distinct factor is present and this is often referred to as the Kaiser criterion (Fabrigar & Wegener, 2012; DeVillis, 2012; Dillon & Goldstein, 1984 ). The eigenvalues for factor 1 and factor 2 are greater than one and the scree J " Images of Scree Test and Parallel Analysis provided in Appendix C . "

PAGE 76

! '( ! plot appears to drop off considerably after the second factor. Parallel analysis results also suggest that two factors are present. Table 3. 2: Iterated Principle Factors Analysis Variable Factor1 Factor2 Uniqueness Trust 0.891 0.095 0.176 Socialize 0.453 0.200 0.777 Compete for Funding 0.068 0.587 0.641 Good Communication 0.851 0.116 0.239 Mission Differences lead to Conflict 0.159 0.401 0.798 Definition of HT leads to Conflict 0.397 0.425 0.621 Competition in Same Community 0.014 0.738 0.457 New Leadership 0.300 0.026 0.908 Low Turnover 0.529 0.267 0.683 Difficulty Sharing Information 0.177 0.565 0.673 Resolve Conflict Effectively 0.859 0.151 0.209 Achieved Past Goals 0.694 0.203 0.443 The graph for the screeplot and the parallel analysis are located in Appendix C . Based on all three of these results, the hypothesis is generally supported that the rotated, iterative principle factor model of two distinct factors suggests that while two factors are present, several items may not be a good fit for the resulting factors. Table 3.2 details the parameters of fit for the factor analysis. The loading plot, the item communal ities and an exploration of the Chronbach's alpha leads to the conclusion that several of the included items from the battery fail to strongly load onto either factor 1 or factor 2. The item communalities reported above suggest that socializing, mission di fferences lead to conflict, and a strong network with a leadership change fail to meet general social science guidelines; although a hard and fast cut point is not identified the current research standards suggest that values between .32 and .7 may be

PAGE 77

! ') ! cons idered when the item theoretically fits the factor and items above .7 should generally be included ( Costello & Osborne, 2009; Velicer & Fava, 1998). Table 3 .3 : Iterated Principle Factor Items Factor 1 Factor 2 Does Not Load Effectiveness Dysfunction (Eigenvalue = 3.82 ) (Eigenvalue = 1.55) Proportion of variance attributed to this factor: .68 Proportion of variance attributed to this factor: .35 There is a great deal of trust among members Conflict arises among members because of competing def initions of human trafficking The partnership will remain strong if the current leader leaves There is good communication among members There are difficulties sharing information among members about victims of trafficking due to confidentiality pol icies Conflict can arise among members because of the different agency/ organizational missions There is low turnover of members in the partnership Partners compete for funding Member sometimes socialize together The members of the partnersh ip are able to resolve conflict effectively There is competition among members that work within the same communities The partnership has achieved it annual goals over the past year Uniqueness is generally considered to represent some level of measurement error as well as the communality of a variable in relation to a factor; uniqueness values above .6 (considered

PAGE 78

! '* ! high) are variables that are not closely correlated with the factors (StataCorp, 2012). Based on these criteria, factor 1 was further explored with five items and factor 2 with four items. Table 3 .3 details the items that appear to load on factor 1, factor 2, or do not load on either. Factor 1, titled Network Effectiveness as it appears to represent positive, whole network outcomes, explains approximately two thirds or 68 % of the variance in the matrix. Network dysfunction, factor 2, describes 35 % of the variance in the test battery and the items that appear to load on this factor have lowe r communalities and higher uniqueness values. The less robust findings for factor 2 can be explored in future scholarship and for the remainder of this particular research, network effectiveness, factor 1, is the primary factor under consideration. Effecti veness more strongly demonstrates items that meet appropriate standards for inclusion based on theory, interitem correlations, and uniqueness values. Network Effectiveness Network effectiveness, the central concept under consideration here, is worthy of further, detailed attention. Since prior research on network effectiveness primarily considers client or community level effectiveness (Provan & Milward, 1995, 2001; O'Toole, 2015; Turrini et al., 2010), the following results demonstrate how items load on factor 1, network effectiveness. Tab le 3 .4 highlights the results of a Chronbach alpha correlation test on the items that appear to load on network effectiveness. Based on the results included in Table 3 .4 , most of the item rest correlations among the variables are close to each other meanin g that the item correlates with the other items comprising the factor; however, the low turnover variable appears to be much lower than the other included items ( Velicer & Fava, 1998). When this item is dropped from the factor, the average interitem correl ation increases to .73 and the

PAGE 79

! (+ ! overall alpha increases to .9 2 . 5 .Generally values greater than .6 on the average interitem correlation are preferable and suggest a reliable indicator, but alphas above .8 are often acceptable (Statacorp, 2012). Therefore, dr opping the turnover item may best be considered a theoretical choice for future research. Table 3.4 : Factor 1 (Network Effectiveness) Chronbach Alpha of Items Item Obs Sign item test correlation item rest correlation average interitem correlation alpha Trust 73 + 0.903 0.837 0.528 0.817 Good Communication 73 + 0.889 0.816 0.537 0.823 Low Turnover 73 + 0.602 0.408 0.733 0.917 Resolve Conflict 73 + 0.890 0.817 0.537 0.822 Achieved Past Goals 72 + 0.805 0.687 0.594 0.854 Test scale 0.586 0.876 Conclusion Establishing the construct of network effectiveness and the construct of dysfunction are a useful advancement in the network effectiveness literature for several reasons. First, the two constructs may now be utilized to develop empirical variables. Now that these two distinct latent concepts have been identified through exploratory factor analysis, future studies can further develop the measurement of each factor through confirmatory factor analysis, and ultimately, structural equation modeling techniques can be utilized to explore relationships between contextual, managerial, functional, and structural network determinants that may lead to effectiveness. C onceptually, the two factors reflect the effectiveness of the network as a combination of trust development, achieved shared goals, good communication, low K " See Table 4: Chronbach's Alpha Without Low Turnover in the Appendix for complete details. " "

PAGE 80

! (" ! membership turnover, and conflict resolution. Somewhat surprisingly, a change of leadership did not correlate with effectiveness. Leadership, and the turnover of leadership, may more effectively represent a determinant of effectiveness as hypothesized earlier rather than correlate as a function of effectiveness. Additionally, differences in organizationa l missions and the willingness of participants to socialize together do not appear to correlate strongly enough to comprise the network effectiveness factor either. Socializing, much like leadership change, may instead be predictors of effectiveness rather than a component. Mission differences that create conflict across organizations may fail to represent effectiveness or dysfunction for several reasons. A third (33%) of respondents reported that they were neutral about the statement that conflict can aris e as a result of differing missions Ð it is not unrealistic to expect that respondents are neutral as they might be able to identify agencies within the network that make the network less effective as a result of differing missions as well as agencies who make the network more effective regardless of mission differences. Since it is currently unclear why so many respondent report neutral feelings, this item is not clearly an aspect of network effectiveness. Future work may need to more closely connect n etwork effectiveness to indicators that represent the outcomes of human trafficking networks actions. For example, only achieving past goals represents the ability of these networks to achieve community or client level outcomes. In order to determine how n etwork effectiveness will actually decrease human trafficking, or directly influence the root causes of wicked problems, require s further operationalization of network effectiveness concepts and measures. Additional work is needed to support the construct validity of each factor as this will enhance and guide the development of empirical analysis of network level effectiveness.

PAGE 81

! (# ! Provan and Kenis (2008) argue d that trust and goal consensus appear to be important aspects related to network effectiveness. Addi tionally, Provan and Milward (2001) suggest ed that network level effectiveness must address the ability of organizations to work together, to fund and maintain information sharing as wells as member commitments to network goals. The results provided here s uggest that trust, low turnover, achieving goals , and communicating effectively are all important dimensions of effectiveness. The addition of the concept of conflict resolution is new to the theoretical discussion of effectiveness. Little attention in cur rent literature attends to conflict resolution developments across networks suggesting this is an area fruitful for further application and consideration. O'Toole (2015) assert ed that developing opportunities to study networks as dependent and independent variables is a much needed development for the advancement of network theory. This study directly develops in that vein. Practically, the development of this network effectiveness variable will benefit grantors, grantees, community agencies, and may drive network administrative organizations to pay increasing attention to the items that comprise network effectiveness.

PAGE 82

! ($ ! CHAPTER 4 PERCEPTIONS OF WHOLE NETWORKS: ATTEMPTS TO END WICKED PROBLEMS Abstract Networks require collaboration from multiple partners; some partners may have goals that fail to effectively align with the mission of the network. Other agency members, however, may argue that a network is the only way to accomplish goals relate d to service provision, equity, or solving wicked problems. Therefore, how do participants define, describe and contextualize the networks to which they belong? Additionally, how do these perceptions shape the effectiveness and sustainability of the networ k? A qualitative analysis of 134 whole networks combating human trafficking revealed that networks are typically able to improve service provision, allow for communication throughout communities, increase education or awareness of a social issue, and produ ce policy/legislation. Network concerns included conflict, genuine commitment and the lack of funding for partnerships Ð while these coalitions may demonstrate that they can provide enhanced services, the network's ability to achieve shared goals is often limited. Practitioner Points • ! Respondents suggest that conflict arises in networks and there appears to be two modes for conflict management Ð handling it openly and honestly with all partners or avoiding it altogether. • ! Conflict in networks decreases the attainment of network goals and the sustainability of the partnership. • ! Trust, a key to any organizational setting where teamwork is required, is built in three key ways: repeated interactions like regular meetings and communications, handling conflict, tur nover, and leadership selection in an open and reflective way, and by working together to achieve shared goals. • ! Granting agencies and umbrella non profit organizations may opt to consider incentivizing network partnerships that support opportunities for op en lines of

PAGE 83

! (% ! communication, developing conflict resolution skills in individual participants in the network, and promoting opportunities for creating shared goals across a community. • ! Networks that form to achieve goals beyond information sharing and dissem ination tend to have lower levels of perceived conflict and build trust more effectively. Introduction Networks, interdependent and structurally equal organizational arrangements comprised of two or more agencies seeking to achieve shared goals (Kilduff & Tsai, 2003; Meier, 2004; Provan & Kenis, 2008), are prevalent in communities across the globe. Agencies often recognize that a single entity is unable to meet the needs of clients in a community (Brummel, 2010; Siddiki, 2015). Therefore, these agencies s eek out other resource providers, partners, or groups to help reach organizational metrics. These agencies may or may not need to rely upon the other organizations to reach their own goals and this distinction may have significant impact on their commitmen t, attention, effort, and communication with the network to which they belong. Community level needs may fail to align with the focus of individual organizations. Or, the provision of services may vary based on the political will around a particular issue in the community to the detriment of other services. Exploring the ways in which organizational participants view the effectiveness, leadership, and sustainability of their networks may lend itself to more effectively understanding the nature of network le vel leadership, outcomes from collaborative networks and possible community level impacts. Agencies also may enter a network due to a policy , which mandates membership or network composition (Saz L Carranza, Salvador Iborra , & Albareda, 2015). Lepore (2016) , for example, argued that services for poor children like child protective services come as a result of a lack of service provision for poor women. Public attention and political will is invested

PAGE 84

! (& ! in a sprawling juvenile justice system and departments of c hild welfare while money spent preventing teen pregnancy, training and supporting young women from vulnerable backgrounds, or providing job placement opportunities may effectively support the well being of America's children. Community agencies attempting to decrease the prevalence of family violence or at risk children may recognize this political reality and partner together to support women in vulnerable situations. The child welfare agency's goals may be to evaluate and assess cases and complete risk an alyses, for example, which can be empirically evaluated but fail to address the root cause of family violence. In order to systematically address the the problem, case workers and supervisors may recognize a need to work jointly with prevention partners fo r long term, community level results. These networks, seeking to end interpersonal violence, may have goals that diverge from those of the specific agency, but partnerships may be worthwhile nonetheless. Preventing violence is one of many wicked problems facing communities across the globe. ÔWicked problems' are those that are contested, systemic, frustrating, and may require multiple, contextual solutions to end them (Head & Alford, 2015). Human trafficking is one example of a contemporary wicked problem with global implications. As borders become less defined , information about the quality of life in other places is more easily shared, and the pace of economic globalization grows, so does the demand for labor, personal safety, or freedom. Trafficking in humans increases in conflict areas and vulnerable populations are especially at risk. Trafficking spans various types of crimes and impacts adult men as well as women and children. Too often, bureaucracies may oversimplify the problem as a sex crime only a ffecting youth and children. Head (2008) suggested this simplification is a direct result of the growth in needing to establish targets and goals that can be met by bureaucrats and

PAGE 85

! (' ! managers in the public sector. Simplification allows for quantification of some aspects of this crime, but may lead to negative externalities for other potential victims and survivors of trafficking because the root cause may not be directly addressed. Wicked problems like human trafficking appear intractable because they are com plex, require participation from multiple sectors, and the solutions may vary among communities. Partnerships and Networks The language of partnerships and networks is interchangeable in this study. Many terms in the public administration and management literatures appear to describe the network concept under consideration here. One indictment of the broadly and quickly expanding literature on networks is that the term Ônetworks' is increasingly nebulous, represents a different concept to different people , and is inconsistently applied (Borgatti & Foster, 2003). This section identifies a conceptual definition of networks and connects that concept to the operationalization of the concept in the survey from which the data for this project w ere drawn. The not ion that several organizations or agencies work together toward achieving a common goal is at the heart of understanding these network arrangements. O'Toole and Meier (2004) argued that networks are an interdependent group of multiple organizations or part s of organizations whereby the connections among the actors are not arranged as superior to each other. This definition suggests that members of the network are considered equal without a hierarchy of reporting. Provan and Kenis (2008) specifically defined networks as "groups of three or more legally autonomous organizations that work together to achieve not only their own goals but also a collective goal" (p . 231). Provan and Kenis (2008) also a sserted that network configurations often require lead organiz ations or network

PAGE 86

! (( ! administrators to provide an effective governance mechanism for positive outcomes. The structure of networks, leadership of the network, and goal directed behavior are all considered here. The Laboratory to Combat Human Trafficking, the c ommunity partner connected to this project, defines Ôpartnership' as: Partnership measures acknowledge that combating human trafficking requires a comprehensive response through the cooperation of multiple sectors. Partnerships bring together diverse exper iences, amplify messages, and leverage resources. For the purpose of this survey, an anti human trafficking partnership refers to a É [e.g. a cooperative relationship between two or more organizations established for the purpose of jointly combating human trafficking in some way.] This language, while utilized specifically for the survey tool, connects to other definitions provided in the network literature and describes ways in which a group of organizations may band together to achieve goals. More spec ifically, this research examines various whole public management networks. Public management networks include "agencies involved in public policy making and/or administrative structure through which public goods and services may be planned, designed, produ ced, and delivered (and any or all of the activities)" (McGuire & Agranoff, 2007). Agranoff and McGuire (2001) indicated that focused and continued efforts to study public management networks remain prevalent (O'Toole, 2014; Saz L Carranza, Salvador Iborra , & Albareda, 2015). The study of networks has arguably developed through three traditions: that of social network analysis in sociology, policy networks in political science, and service provision networks in public management (B erry et al., 2004). Berry and colleagues find contributions across all three traditions and suggest that overall structure matters when it comes to understanding networks, an individuals' place and the centrality of an organization also matters within the network, and these structures and positions can

PAGE 87

! () ! influence the effectiveness of the network itself (see also, Agranoff & McGuire, 2001; Ansell & Gash, 2008). Research on whole networks examines both independent and dependent variables from the network level ; across many network traditions, few studies address whole networks empirically and the effectiveness of whole networks appears to be an area requiring additional attention and consideration (Provan, Fish, & Sydow, 2007). This article examines whole netwo rks seeking to end human trafficking. Respondents in the networks self report ed on the functioning and future of their own network. The central research question is: How do participants define, describe, and contextualize the networks to which they belong? Additionally, this research considers: How do these perceptions of the network shape the effectiveness and sustainability of the network? Two different datasets, collected from the same survey instrument are analyzed and discussed in the following section s. Theory 4Ps: Prevention, Protection, Prosecution, and Partnership The theoretical model under investigation here identifies four components of the movement necessary to end human trafficking: prevention, protection, prosecution, and partnership (4Ps). Th e concept of the 4Ps was developed by U.S. Department of State's Office to Monitor and Combat Trafficking in Person's and was independently validated through the Colorado Plan by the Laboratory to Combat Human Trafficking. In the initial version of federal legislation, the Trafficking Victims Protection Act of 2000, three Ps were identified. The fourth P, partnership, was added as a policy priority in 2010 in response to criticisms from recipients of Bureau of Justice Assistance grants and research organiza tions like the Laboratory to Combat Human Trafficking and Polaris Project (see also Sheldon

PAGE 88

! (* ! Sherman, 2012). Governments and NGOs alike recognized that prosecuting the perpetrators, protecting victims and preventing the crime may fail to occur in practice w ithout partnerships committed to information sharing, social justice, and community interventions. Data The data, provided by the Laboratory to Combat Human Trafficking (http://www.combathumantrafficking.org/), resulted from a multi year planning and deve lopment process. The survey was designed to determine how many efforts are aimed at combating human trafficking, the extent of the success of those efforts, and to identif ication of promising practices conducted in communities across Colorado and the Unite d States. One hundred and eighty six respondents completed the national survey and 132 respondents completed the Colorado survey. The survey instrument included 247 items and several respondents reported feeling survey fatigue as a result of the length. T he partnership questions were the final section of the survey, which may mean that respondents who felt fatigued but may not have reported it in the final comment box of the partnership section, were systematically less likely to respond to that section. H owever, it should also be noted that many of the respondents, due to skip patterns built into the survey, only completed limited sections of the survey pertaining to their organizations focus areas. It remains unclear how this systematic bias may impact fi ndings, but 71 % of respondents completed item 191: Is your organization involved in a partnership effort to combat human trafficking (formal or informal)? and 59% of those respondents do not belonging to a partnership. In the national survey, 66% of respon dents ma d e it to question 191 and of those respondents, 22% report that they do not belong to a partnership. These descriptive results suggest that while many respondents failed to fully

PAGE 89

! )+ ! complete the survey, there does not appear to be evidence to suggest the participants who reflect ed a predisposition to participate in networks versus those who did not. This particular research utilizes the comment boxes provided for further discussion of the Likert style questions included in the partnership section of t he survey. Fifteen questions related to network participation, formation, sustainability, mission, goals achieved, trust, leadership, member turnover, and others were analyzed. These items represent those most closely related to the research questions unde r consideration. All comments made by respondents to these items are included in the dataset for analysis. Methodology The data were examined using a major theme analysis to explore open ended written responses. The qualitative responses offer an in dept h examination of the network frameworks. Additionally, the captured narratives offer rich quotes that exemplify emergent themes beyond standard Likert scales or other nominal survey responses. Initially, data analysis was conducted separately for state a nd national respondents. A subsequent analysis compared major themes between national and state responses. The data were analyzed using the holistic approach described by Salda–a (2015). First, data were coded for words or short phrase. Second, patterns were identified and transformed into categories. Third, major themes were developed to organize groups of repeating ideas (Auerbach & Silverstein, 2003; Salda–a, 2015). The analyses were conducted manually rather than by computer and two raters were used to assist in establishing inter coder agreement. In the rare cases the coders disagreed the data were re examined to arrive at consensus. Results

PAGE 90

! )" ! After completing the national analysis then utilizing the state dataset for confirmatory evidence, one key r esult of the data findings suggests that national respondents and state respondents have slightly different perspectives on the work of their partnerships. While many partnerships report shared goals, reaching consensus on a mission statement, and having a vision for connecting the mission to goals, several networks also suggest that this work is still in progress or has yet to begin. The following sections describe key conceptual themes that arose from both the national and state datasets. The first sectio n details the differences and similarities across both data sets, then the sets are considered together in the following sections. The main themes include: the 4Ps theory; network formation and sustainability; trust and conflict; and the tensions and negot iations that result for organizations partnering as a network. Participants appear to believe greater possibilities are possible when organizations come together to end wicked problems, but the goals achieved by the networks fail to meet expectations. Take n together these themes reinforce the value and utility of the 4P theory for collaborative action. State and National Differences The national data set findings indicate that most networks organize their goals around either a single construct or combinatio n of protection, prosecution, and prevention. Many of the partnerships included in the national survey formed as a result of funding from Bureau of Justice Assistance and Office for Victims of Crime (BJA/OVC) task forces. Starting in 2004, funding for part nerships/task forces to combat human trafficking came in grants distributed to law enforcement agencies or victims' service organizations. These funding streams were initiated in the Bureau of Justice Assistance (BJA) and funneled to victims' services thro ugh the Office for Victims of Crime (OVC) or to law enforcement agencies through the United

PAGE 91

! )# ! States Attorney's Office. Some communities received both funds while other task forces were created with either victim services or law enforcement as the grantee fo r the task force. Some task forces had mandates to provide comprehensive services while others were focused on providing services to more specialized populations or subsets of victims. These funded groups are referred to as Enhanced Collaborative Model Tas k Forces and they were meant to work toward combating all forms of human trafficking in collaborative ways. The need for funding often identified by respondents in the survey refers to the loss of the BJA/OVC monies in many regions. These grant monies fluc tuate in communities making them unreliable and, for many leaders, makes the task force unsustainable over the long term. Without the funding, one respondent note d, "unfortunately, dedicated funding is the best method of sustainability. The loss of funding for two dedicated leaders may see the demise of the coalition's strong presence." Another mention ed "funding is a big determining factor [of sustainability] but several partners are committed to this task regardless of funding." These comments signal an i mportant consideration related to funding and network sustainability, effectiveness, and durability in the national data set: networks that rely on grant funding or were formed by the BJA/OVC funding may not be as sustainable as networks formed through pri or collaboration on harm reduction or shared service provision. Based on this evidence, it is unclear if government funding is necessarily a positive or negative factor for the success of networks. State respondents focus more intensely on service provisio n for survivors as well as prevention in the form of awareness and education campaigns. The state under consideration had a rise in awareness of the crime due to Bureau of Justice Assistance funding to two different networks during the time of the survey. This situation may have caused the results

PAGE 92

! )$ ! to vary somewhat from the overall national sample. The state respondents report that without funding, the sustainability of the networks decreased and note that the networks were experiencing rapid increases in me mbership. These structural changes lead to uncertainty about shared definitions of the wicked problem and increasingly diverse views of a solution to end trafficking. One respondent, for example, suggest ed that, "the core group has a good understanding of HT [human trafficking], but there are always so many new members that one cannot say that all members have a good understanding." The positive outcomes related to increasing prevention may have the effect of making partnerships more challenging to manage. This finding is important for states and local communities across the country. Empirical findings suggest that as the size of networks increases, the ability of members to participate, engage, and act may decrease (La Due Lake & Huckfeldt, 1998; Provan & K enis, 2008). Network size, leadership, duration, structure, and funding impact the effectiveness of networks (Provan & Milward, 1995). As awareness of the crime increases, seemingly rapidly in the national discourse as well, networks may require different management structures to remain viable and effective. Comparing the national and state qualitative data show at the state level, networks tended to focus on prevention and protection through direct service provision or raising awareness in the community and the partnership serves the role of disseminating information. In the national survey, the purpose of partnerships appears to be more comprehensively focused on combating human trafficking. Responses common in the state survey include: "1) figure out co ordinated efforts for services, 2) education, 3) provide coordinated efforts for services." In national responses, the median response appears akin to "discussing trends, issues, and gaps in combating human trafficking statewide." Both state and national

PAGE 93

! )% ! respondents report efforts to end trafficking attempt to be more coordinated, to more effectively deliver victim services, and to encourage prosecution. Overall, state respondents focus more on those efforts while the national survey respondents had a slig htly broader approach. Both national and state respondents reflect on similar frustrations faced by the network. A central theme of frustration was that the network serves only as an avenue for information dissemination rather than action. One state respo ndent state d , "we would have to be part of a real partnership not just publicizing activities of another organization" to have a sustainable partnership and another suggests that the partnership does not manage trust, instead "meetings are held but there i s a sense that real information is not really being share.," A national respondent note d the partnership is, "limited to information sharing, so there's not much opportunity to actually collaborate on tasks." While the genuineness of the partnerships might be questioned in some of the networks, respondents appeared to believe in an idealized form that was rarely achieved. Participants note d , for example, that "we've all made efforts to inform each other so there are no surprises, although there has someti mes been disappointment at the decisions made by one side or another." While it is unclear from this comment alone what it means to be on one side or another, the respondent report ed that the partnership fails to achieve the goal of communication sharing. Respondents state d that commitments from all agencies in the network, more variety of stakeholders, funding, and shared goals are required for sustainable and successful networks, but this was not always realized. Setting realistic goals and acknowledging the times when the network was less successful as expected appears to be an important factor for both trust building and managing conflict.

PAGE 94

! )& ! Prevention, Protection, Prosecution, and Partnership (the 4Ps) Networks can be helpful for information disseminati on and participants view the goals of the network as protection, prosecution, prevention, and partnership (4Ps). The 4P model provides a useful frame for understanding the work done by member organizations in the network, but the majority of networks lack comprehensive missions for achieving all 4 of the Ps. One respondent note d the purpose of their partnership is "to harmonize state and federal law enforcement, non profits, city services, social services and the community in a holistic anti trafficking and victim assistance effort." Similarly, another state d that they seek to "build community and share information and resources as well as build bridges to know who is doing the work and how together we can help make a difference in our state." Common themes included the topics of raising awareness, providing services to victims, developing educational and training materials, and coordinating response/prosecution in identified cases of trafficking. While these types of actions connect directly to the 4p model, there appears to be some missing (p)ieces of the model. Many of the respondents, for example, argue d that identifying victims, securing funding, and establishing legislation are priorities for partnerships. Identifying victims, protection tasks, may also be conducted by prosecutors and law enforcement and, as such, might best be considered as a shared goal for partnerships. Simply because identifying victims fails to clearly fit as a particular P in the model may mean more specific goals are required of pa rtnerships to achieve particularly if they receive OVC/BJA funding. Network Formation and Sustainability The partnerships seeking to end human trafficking under consideration formed under three conditions: community agencies recognizing a need, networks or partnerships that

PAGE 95

! )' ! already worked collaboratively in other capacities (i.e., homelessness, child protection, or harm reduction), and funding from the federal grants. An interesting finding related to the formation of the partnerships is that the responde nts whose networks formed as a result of federal funding reported that the sustainability of the partnership depended upon continued funding. Respondents who had worked in other capacities share similar viewpoints: Participants do more than just human traf ficking. Trafficking, like poverty, substance abuse, mental illness, is just one more cause of homelessness that we must understand better to serve our constituents. Compassion of the participating members and relationships that have formed in the process [contribute to sustainability]. We are non funded, non mandatedÉtherefore, participation is based on value. A history of 10+ years working together to serve victims of domestic violence. The fact that more survivors have been identified as a result of th e collaboration and education. F ederal funding appears to make the long term viability of the partnership more tenuous than if the partnership formed in either of the other two ways. Funding from BJA or OVC may also create an added layer of power inequity across partnerships between the agency directly receiving the funds and the others joining the task force. When asked about the leadership of the network, numerous respondents who received funding clearly identified that partner as the network leader and suggested that without continued funding the leadership change may result in the end of the partnership. The goals of the partnership may also be driven by the lead partner Ð the BJA/OVC partnerships are directly tied to achieving either the comprehensive or specialized services required in the grant whereas partnerships formed in response to community needs may be able to more effectively change direction, population served, or shift to or from prevention, prosecution, or protection as needed. This finding is somewhat contrary to Klijn and Koppenjan's (2000) suggestion that government is an

PAGE 96

! )( ! effective leader of network arrangements due to high levels of trust, legitimacy, and unique resources. When the resources (grants) are no longer available, it appears t hat the legitimacy of a government network leader are unable to sustain the network and the resources may skew the network toward goals fail to match community needs. Trust and Conflict Developing trust and managing conflict sections in national and stat e responses highlight challenges faced by networks combating human trafficking. Trust is regularly cited as a necessary component of network management as it is indicated as contributing to network effectiveness and must be shared across organizations wit hin networks not just among dyads (Agranoff & McGuire, 2001; McGuire, 2006; O'Toole, 1997; Provan & Kenis, 2008). Building trust was achieved in networks in three ways: regular meetings, effective communication, and shared goals with a consistent level of commitment from members. A lack of trust from members resulted when little attention was paid to creating an environment of honesty, openness, or relationship building. From both perspectives of building trust or lacking trust in the network, respondents c omment ed that trust is achieved when "we have all been honest about our individual viewpoints and goals so that we can negotiate and conceive programming from a place of integrity and honesty." Trust was lost when respondents view the network as insincere in collaborative efforts, unwilling to openly or honestly address conflict, varied viewpoints, or lack of action, and when "there is a sense that real information is not really being shared." Trust in the networks had a clear and distinct connection to con flict in the network.

PAGE 97

! )) ! Table 4.1 : Conflict in Network Arrangements Managing Conflict Avoiding Conflict T he way the coalition manages conflict is by attacking the problematic issue honest and openly [sic] O vert conflict seems to be avoided, a nd left to fester. This affects the overall trust among members and the ability to be productive C ommunications is the #1 tool in our success. It's not about one person but the coalition as a whole H asn't been much overt conflict, probably because meeting s don't focus too much on real issues (cases, etc.). When there has been overt conflict, usually one of the parties involved simply stops participating G ood facilitation is priceless D on't think conflict is managed Ð it is ignored O pen, fair, honest and respectful communication A void conflict Ð do not address conflict (which is a problem) F or the most part, the partnership has been able to work through conflicts by regular meetings and retreats S ome individuals who were unable to work through the conflic ts chose to dedicate their efforts in other arenas T he co chairs do a good job of resolving conflicts D ifficult issues are avoided D irect communication with those involved in the conflict I feel that right now this [conflict] has not been addressed T alk through it. Most members are committed to the cause and can get past philosophical differences A ddress the issues openly in meetings T here could be more trust and openness in general, conflict is unpacked and managed reasonably Conflicts in th e networks were handled primarily in two ways: avoided or handled directly and openly. Table 4. 1 highlights the responses across both national and state surveys discussing the patterns of response related to conflict. Several respondents suggest ed that whe n conflict becomes problematic in networks, those organizations, agencies, or individual representatives opt to leave the network. While this is not a large group of respondents, Hirschman's (1970) notion of exit, voice, and loyalty appears to again provid e meaningful options for network participants. Conflict resulting in exit may, however, mean that the ability of the network to provide services, reach diverse populations, or achieve other shared

PAGE 98

! )* ! goals decreases as a result of those exits. Voice, consiste ntly discussed by those respondents who felt conflict was managed effectively, appears to be a necessary condition for conflict resolution. Voice, in conjunction with effective management practices like genuinely sharing important information, providing co ntexts for feedback, and effective facilitation, may be both necessary and sufficient conditions for effective conflict resolution in networks. The responses in conflict management appear to reinforce the notion that trust and direct communication are di rectly related to the sustainability and effectiveness of networks. Sustainability seems to be impacted by the composition network membership. If agencies that disagree or have conflicts with the philosophy, goals, or actions of the network are essentially forced out as a result of the disagreement, the network may lose out on key actors necessary to achieve the goal of combating human trafficking. Additionally, decreasing network effectiveness occurring as a result of an inability to openly handle conflict will have real impacts on service provision to victims, community awareness and education programs, and the ability to prosecute criminals. Tensions and Negotiations Several underlying tensions appear to compete for attention in networks. These tensions include the trade offs between the internal training, education, and information sharing needs of the network and the work of external facing programs, goals, and campaigns and focusing on proactive community work and reactive service provision to victims . These tensions seem to develop as a result of the 4P model. Prevention and protection activities, while seemingly complementary, both require an extensive amount of resources to be conducted effectively. Prevention, the activities related to education, t raining, awareness, advocacy, and community engagement are mentioned as both goals and activities of

PAGE 99

! *+ ! networks as are protection goals like comprehensive victim services, referrals, responses to tips or calls, resources like shelters, counseling or educatio n programs, and identification of survivors. Prosecution, while noted as part of the mission or goal of the network, appeared to be conducted apart from prevention and protection activities; this is unsurprising as the network likely had few members able t o conduct investigations, charging, or prosecution. Network members reported that while the mission or shared goals of the network include work spanning the 4Ps, when they describe the most significant accomplishments of the network, the goals they achieve are in only one P. One network suggests this as the mission, "utilizing a victim centered philosophy, to ensure the protection of victims, the prosecution of offenders, and the prevention of human trafficking and slavery through an effective, coordinated partnership," and reports the significant achievements as "The partnership has strengthened coordination and communication, and has made the provision of services more seamless." The mission was comprehensive yet the achievement focuses on the partnership and protection aspects. Another respondent describes the mission as, "through advocacy, partnership, and direct service, our mission is to abolish all forms of human trafficking and free victims from the grip of slavery," and the goals achieved include "de veloping strong relationships with one another even when professional perspectives are different; Building awareness and advocating on behalf of human trafficking to the public, with one another, and with our respective colleagues." Again, this network has a comprehensive mission but was only able to achieve goals primarily in partnership and prevention areas. The negotiations these networks face in prioritizing their work may be viewed as tradeoffs between external and internal facing trainings and program s Ð those aimed at prevention versus partnership. Additionally, these tradeoffs appear as proactive or reactive actions for ending trafficking Ð

PAGE 100

! *" ! protection versus prevention. Since many of these networks are newly formed it is likely that they felt demands to focus in one particular area of strength to conduct their work, but these tradeoffs demonstrate a tension that the partnerships will need to manage effective if the are to achieve their mission and reach their goals. Discussion Several key findings s uggest implications for improving the capacity of networks to end wicked social problems. First, network leaders may want to consider the ways in which new members are introduced to network participation. Many respondents across both the national and state surveys suggested that awareness of trafficking and human slavery is growing, increasing the likelihood that additional members will be required to make a significant improvement in protection, prevention, or prosecution. New member trainings, particularl y trainings that acquaint joiners with methods for conflict resolution, the mission and goals of the partnership, group norms and responsibility expectations, and the ways in which the group communicates appear particularly important for network effectiven ess and sustainability. Another consideration for network effectiveness includes funding structures. All non profit organizations highly prioritize funding, but network funding structures and opportunities appear more tenuous. While the Trafficking Victim s Protection Reauthorization Act passed as an attachment to the 2013 Violence Against Women Act reauthorization (Public Law 113 Ð 4), and the legislation specifically stated that, "officials of the United States Government, shall promote, build, and sustai n partnerships between the United States Government and private entities, including foundations, universities, corporations, community based organizations, and other nongovernmental organizations , " the partnerships

PAGE 101

! *# ! rarely receive funding for participating in network arrangements. Partnerships and networks represent an increasing unfunded mandate to many local government entities and non profits. This additional administrative burden will need some form of financial support if it is to be sustainable. Commun ity agencies like the United Way, Ford Foundation, or the Gates Foundation, who act as umbrella organization funders through RFP processes, may want to consider structuring funding opportunities toward effective network structures. Incentivizing networks w ith a shared mission, proven methods of open conflict management and communication, and that provide support for the time spent on goals will improve network level outcomes and effectiveness. Funding for effectively supported partnerships may be the future of working to end wicked problems across communities. The final implication drawn from both the national and state surveys suggests that state and national legislation identifying, defining, and codifying trafficking are somewhat dissimilar. These diffe rences in legislation appear to make it difficult for partnerships to operate effectively. This is an old phenomenon as it occurs across child welfare, interpersonal violence, and other offenses in the judicial system, but laws related to prostitution, hom elessness, and loitering appear intricately linked to the crime of trafficking. Working with prosecutors and law enforcement officers to determine the circumstances related to homelessness, prostitution, smuggling, and trespassing may improve prevention an d protection outcomes. Trainings for service providers done jointly with state and local law enforcement as well as district attorneys with the federal, state and municipality statutes related to these topics could improve identification of survivors and b etter support networks seeking to end human trafficking. Conclusion

PAGE 102

! *$ ! Two central concepts for further research include the way in which federal funding influences sustainability of mandated partnerships and the distinctions between networks and cross secto r collaborations particularly as the outcomes of those groups related to achieving shared goals. Many of the networks created as a result of federal funding opportunities reported lower sustainability or concern for the future of the partnership without c ontinued funding. Funding did appear as a concern for other respondents, however, many reported that they were seeking and cultivating local or state sources for funding. Funding issues were reported as typically affecting specific organizations or agencie s when the partnership formed as a result of preexisting relationships or a demonstrated community need. Federal agencies like BJA/OVC may want to consider funding already established networks in the future and allowing those groups to maintain their previ ously established leadership structures. This finding appears to contradict prior research on network funding being an important, positive relationship related to sustainability and effectiveness (Provan & Milward, 1995). The distinction between federal, s tate, and local funding streams to partnerships and the relationship between formation and sustainability as it relates to start up funding for the network require additional consideration as the federal government continues to view task forces, partnershi ps, and collaborations in local communities necessary tools for prosecution and crime prevention. Future studies related to network effectiveness and ending wicked problems will need to carefully consider the distinction between cross sector collaboration s and networks. Most of the definitions guiding the network and whole network literature indicate that networks provide a context for multiple organizations to work together to achieve shared goals. Many

PAGE 103

! *% ! respondents argued that while they did belong to net works, the networks merely worked together to share each other's information, disseminate activities occurring in the community, or even claimed that the real work of the network was conducted in some other context beyond the regular network gatherings/mee tings. Although additional consideration of this notion is necessary, it is worth arguing that shared goals must go beyond information sharing if the network is to have any functional meaning to participants. Bryson, Crosby and Stone (2006) described cross sector collaborations as "the linking or sharing of information, resources, activities and capabilities by organizations in two or more sectors to achieve jointly an outcome that could not be achieved by one organization is sectors separately (p. 44)." Cr eating a clearer distinction between cross sector collaborations and networks may help illuminate the differences between seeking to complete shared goals versus collaboration in the forms of information sharing, activities, or resource distribution. The c oncept of Ôshared goals' that networks act together to achieve may need to be further defined as community level comprehensive service provision, funding, educational and training programs, legislation, needs or asset assessments, protocols for reporting a nd prosecution, and crime prevention.

PAGE 104

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! *& ! REFERENCES Agranoff, R., & McGuire, M. (2001). Big questio ns in public network management research. Journal of Public A dministration R esearch and T heory, 11(3), 295 326. Alejano Steele, A. (2013). Friends and Enemies in the Crime of Sex Trafficking. In Harre, R. & Moghaddam, F. (2013). Psychology of Friendship and Enmity, The: Relationships in Love, Work, Po litics, and War . Santa Barbara, Ca: Praeger. Allison, P. D. (2001). Missing data (Vol. 136). Thousand Oaks, CA: SAGE Publications. Ansell, C. (2006). Network institutionalism. The Oxford H andbook of P olitical I nstitutions. Oxford , U nited K ingdom : Oxf ord University Press . Ansell, C., & Gash, A. (2008). Collaborative governance in theory and practice. Journal of P ublic A dministration R esearch and T heory, 18(4), 543 57 1. Anderson, E. E., Solomon, S., Heitman, E., D uBois, J. M., Fisher, C. B., Kost, R. G., Lawless, M. E., Ramsey, C., Jones, B., Ammerman, A. & Ross, L. F. (2 012). Research ethics education for community engaged research: A review and research agenda. Journal of Empirica l Research on Human Research Ethics , 7 (2), 3 19. Auerback, C. F., & Silverstein, L. B. (2003). Qualitat ive data: An introduction to coding and analysis . New York, NY: New York University Press. Baldwin, D. A. (1985). Economic Statecraft . Princeton, NJ: Princeton U niversity Press. Balkundi, P., & Kilduff, M. (2006). The ties that lead: A social network approach to leadership. The Leadership Quarterly , 17 (4), 419 439. Berry, F. S., Brower, R. S., Choi, S. O., Goa, W. X., Jang, H., Kwon, M., & Word, J. (2004). Three traditions of network research: What the public management research agenda can learn from other research communities. Public Administratio n Review , 64(5), 539 552. Borgatti, S. P., & Foster, P. C. (2003). The network paradigm in organizational research: A review and typology. Journal of M anagement , 29 (6), 991 1013. Borgatti, S. P., Mehra, A., Labianca, G. J., & Brass, D. J. (Eds.). (2014). Contemporary p erspectives on o rganizational s ocial n etworks (Vo l. 40) . Bingly, United Kingdom: Emerald Group Publishing. Borsboom, D. (2006). The attack of the psychometricians. Psychometrika , 71 (3), 425 Ð 440. http://doi.org/10.1007/s11336 006 1447 6 .

PAGE 105

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! *' ! Bovaird, T. (2007). Beyond engagement and participation: Use r and community coproduction of public services. Public Administration R eview , 67 (5), 846 860. Bozeman, B. (1987). All organizations are public: Bridging pu blic and private organizational theories . New York: Jossey Bass Inc. Brady, H. E. , & Collier, D . (2010). Rethinking s ocial i nquiry: Diverse tools, shared standards . Lanham, MD: Rowman & Littlefield Publishers. Brass, D. J., Galaskiewicz, J., Greve, H. R., & Tsai, W. (2004). Taking stock of networks and organizations: A multilevel perspective. Academy of Management Journ al , 47 (6), 795 817. Brown, T. A. (2015). Confirmatory factor analysis for applied research . New York, NY: Guilford Publications. Brummel, R. F., Nelson, K. C., Souter, S. G., Jakes, P. J., & Williams, D. R. (2010). Social learning in a policy mandated collaboration: Community wildfire protection planning in the eastern United States. Journal of Environmental Pl anning and Management , 53 (6), 681 699. Bryson, J. M., Crosby, B. C., & Stone, M. M. (2006). The design and implementation of Cross L Sector collaborations: Propositions from the literature. Public Administration Review , 66 (s1), 44 55. " Burt, R. S. (1992). Structural Holes . Cambridge, MA: Harvard University Press. Campbell, R., Sefl, T., Barnes, H. E., Ahrens, C. E., Wasco, S. M., & Zaragoza Diesfeld, Y. (1999). Community services for rape survivors: enhancing psychological well being or increasing trauma? Journal of Consulting and Clinical P sychology , 67 (6), 847. Castells, M. (2000). The Rise of the Network Society: Economy, Society and Culture . Cambridge , Ma : Blackwell Press. Coase, R. H. (1937). The nature of the firm. Economica , 4 (16), 386 405. Costell o, A. B. & Osborne, J. W. (2009). Best practices in exploratory factor analysis: Four recommendations for getting the most from your analysis. Pan Pacific Management Review , 12 (2), 131 146. Creswell, J. W., & Plano Clark, V. L. (2007). Designi ng and conducting mixed methods research. Thousand Oaks, Ca: SAGE Publications . Dawes, S. S., Cresswell, A. M., & Pardo, T. A. (2009). From "need to know" to "need to share": Tangled problems, information boundaries, and the build ing of public sector knowledge networks. Public Administration Review , 69 (3), 392 402.

PAGE 106

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! *( ! De Winter, J. C., & Dodou, D. (2012). Factor recovery by princip al axis factoring and maximum likelihood factor analysis as a function of factor pattern and sample size. Journal of Applied Statistics , 39 (4), 695 710. DeVellis, R. F. (2012). Scale development: Theory and applications (Third Edition). Thousand Oaks, CA: SAGE Publications Dillon, W. R., & Goldstein, M. (1984). Multivariate analysis: Methods and applications (Vol. 45). New York: Wiley. Dusa, A. (2007). User manual for the QCA(GUI) package in R. Journal of Business Research , 60(5), 576 586. Dusa, A., & Thiem, A. (2012). Qualitative c omparative a nalysis with R . New York, NY: Springer New York. -----. (2015). Enhancing the Minimization of Boolean and Multivalue Output Functions With eQMC. Journal of Mathematical Sociolog y , 39 (2), 92 108. Emerson, K., Nabatchi, T., & Balogh, S. (2012). An integrative framework for collaborative governance. Journal of Public Administration Research and Theory , 22 (1), 1 2 9. Fabrigar, L. R., & Wegener, D. T. (2012). Exploratory factor analysis . Oxford, United Kingdom : Oxford University Pr ess. Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating the use of exploratory factor analysis in psychological research. Psychological M ethods , 4 (3), 272. Ferlie, E. (1996). The new public management in action . Oxford, United Kingdom: Oxford U niversity P ress. Ferlie, E., & Pettigrew, A. (1996). Managing through networks : some issues and implications for the NHS. British J ournal of M anagement , 7 (s1), S81 S99. Foot, K. (2016). Collaborating against human trafficking; Cross sector challenges and Practices . Lanham, Md: Rowan and Littlefield . Galbraith, J. I., Moustaki, I ., Bartholomew, D. J., & Steele, F. (2002). The analysis and interpretation of multivariate data for social scientists . Boca Raton, FL: CRC Press. Gallagher, A. T. (2010). The international law of human trafficking . Cambridge, MA: Cambrid ge University Press.

PAGE 107

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! *) ! Gazley, B., & Brudney, J. L. (2007). The purpose (and perils) of government nonprofit partnership. Nonprofit and Voluntary Sector Quarterly , 36 (3), 389 415. Gerring, J. (2011). Social s cience m ethodology: A unified framework . Cambridge, MA: Cambridge University Press. Goodale, M. , & Merry, S. E. (2007). The P ra ctice of Human Rights: T racking law between the global and the local . Cambr idge, MA: Cambridge University Press. Granovetter, M. (1973). The strength of weak ties. American Journal of S ociology , 78 (6), l 360 1380 . Gr anovetter, M. (1985). Economic Action and Social S tructure: the problem of embeddedness. American Journal of S oc iology, 91(3) , 481 510. Guba, E. G., & Lincoln, Y. S. (1989). Fourth generation evaluation . Thousand Oaks, CA: SAGE Publications. Haesebrouck, T. (2015). Pitfalls in QCA's consistency measure. Journal of Comparative Politics , 8 (2), 65. Hart, J. (2009). Human trafficking . New York, NY: The Rosen Publishing Group. Head, B. W. (2008). Wicked problems in public policy. Public Policy , 3 (2), 101. Head, B. W., & Alford, J. (2015). Wicked p roblems i mplications for p ublic p olicy and m anagem ent. Administration & Society , 47 (6), 711 739. Hirschman, A. O. (1970). Exit, voice, and loyalty: Responses to decline in firms, organizations, and states (Vol. 25). Harvard university press. Hollander, E. L., & Saltmarsh, J. (2000 ). The E ngaged U niversity. Academe , 86 (4), 29. Hood, C. (1991). A public management for all seasons. Public administration , 69 (1), 3 19. ----. ( 1995). The "New Pu blic Management" in the 1980s: Variations on a T heme. Accounting, Organizations and S ociety , 20 (2), 93 109. International Labour Organization. (Retrieved August, 22, 2015) website: http://www.ilo.org/global/top ics/forced labour/policy areas/statistics/lang -en/index.htm Jolliffe, I. (2002). Princ ipal component analysis . New York: Springer. Jones, B. D., Baumgartner, F. R., & True, J. L. (1998). Policy punctuations: US budget authority, 1947 Ð 1995. The Journal of Politics, 60(01), 1 33 .

PAGE 108

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! ** ! Jones, C., Hesterly, W. S ., & Borgatti, S. P. (1997). A General Theory of Network G overnance: Exchange conditions and social mechanisms. Academy of Management R eview , 22 (4), 911 945. Kaboolian, L. (1998). The New Publ ic M anagement: Challenging the boundaries of the management vs. administration debate. Public Administration Review , 58(3), 189 193. Katz, D., & Kahn, R. L. (1966). Organizat ions and the system concept . In Shafritz, J., Ott, J., & Jang, Y. (2015). Classics of Organization Theory . New York, NY: Cengage Learning. Katz, D., & Kahn, R. L. (1978). The s ocial p sychology of o rganizations . New York: Wiley . Katz, M. L., & Shapiro, C. (1985). Network e xternalities, c ompetition, and c ompatibility. The American Economic Review , 75(3), 424 440. Kickert, W. J., Klijn, E. H., & Koppenjan, J. F. M. (Eds.). (1997). Managing complex networks: Strategies for the public sector . Thousand O aks, CA: SAGE Publications . Kilduff, M., & Tsai, W. (2003). Social networks and organizations . Thousand Oaks, CA: SAGE Publications . Kim, J. O., & Mueller, C. W. (1978). Factor analysis: Statistical methods and practical issues (Vol. 14). Thousand Oaks, CA: SAGE Publications . Kim, J. O., & Mueller, C. W. (1978). Introduction to factor ana lysis: What it is and how to do it (No. 13). Thousand Oaks, CA: SAGE Publications . King, G., Honaker, J., Joseph, A., & Scheve, K. (2001). Analyzing incomplete political science data: An alternative a lgorithm for multiple imputation. American Political Science Association 5 ( 1 ) , 49 69). King, G., Keohane, R. O., & Verba, S. (1994). Designing social inquiry: Scientific inference in qualitative research . Princeton, NJ: Princeton University Press. Kin g, G., Keohane, R. O., & Verba, S. (1995). The importance of research design in political science. American Political Science Review , 89 (02), 475 481. Klijn, E. H. (2005). Networks and inter organizational management: C hallenging, steering, evaluation, and the role of public actors in public management. In Ferlie, E., Lynn, L.E., & Pollitt, C. (Eds.), The Oxford handbook of public management , ( 257 281 ) . Oxford, United Kingdom: Oxford University Press.

PAGE 109

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! "++ ! Klijn, E. H., & Koppenjan, J. F. (2000). Public ma nagement and policy networks: F oundations of a network approach to governance. Public Management an International Journal of Research and Theory , 2(2), 135 158. Koliba, C., Meek, J. W., & Zia, A. (2010). Governance Networks in Public Administration an d Public Policy . New York, NY: CRC Press, Taylor & Francis Group. Labianca, G. J. (2014). Negative t ies in o rganizational n etworks . In Borgatti, S. P., Mehra, A., Labianca, G. J., & Brass, D. J. (Eds.). (2014). Contemporary perspectives on organizational social networks (Vol. 40). Bingley, U nited K ingdom : Emerald Group Publishing. Lane, J. E. (2005). Public a dministration and p ublic m anagement: T he principal agent perspective . New York, NY: Taylor & Francis. La Due Lake, R., & Huckfeldt, R. (1998). Social capital, social networks, and political participation. Political Psychology , 19 (3), 567 584. Lecy, J. D., Mergel, I. A. , & Schmitz, H. P. (2014). Networks in Public Administration: Current scholarship in review. Public Management Review , 16 (5), 643 665. Lingard, L., Albert, M., & Levinson, W. (2008). Grounded theory, mixed methods, and action research. Bmj , 337 , 459 461. Marsden, P. V. (1990). Network data and measurement. Annual Review of Sociology , 16, 435 463. McGuire, M., & Agranoff , R. (2007, October). Answering the big questions, asking the bigger questions: Expanding the public network management empirical research agenda. In Public Management Research Association Conference, Tucson, AZ . McGuire, M., & Silvia, C. (2009). Does leadership in networks matter? Examining the effect of leadership behaviors on managers' perceptions of network effectiveness. Public Performance & Management Review , 33 (1), 34 62. McPherson, M., Smith Lovin, L., & Cook, J. M. (2001). Birds of a feather: Homophily in social networks. Annual R eview of S ociology , 27(1), 415 444. Meier, K. J. & Hill, G. C. 2005. Bureaucracy in the twenty first century. In Ferlie, E., Lynn, L.E., & Pollitt, C. (Eds.), The Oxford handbook of public management , (51 71). Oxford, United Kingdom: Oxford University Press. Meier, K. J., & O'Toole, L. J. (2001). Managerial strategies and behavior in networks: A model with evidence from US public education. Journal of Public Administration Research and Theory , 11 (3), 271 294.

PAGE 110

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! "+" ! Meier, K. J., & O'Toole, L. J. (2003). Pu blic management and educational performance: The impact of managerial networking. Public Administration Review , 63 (6), 689 699. Miethe, T. D., Hart, T. C., & Regoeczi, W. C. (2008). The conjunctive analysis of case configurations: An exploratory method for dis crete multivariate analyses of crime data. Journal of Quantitative Criminology , 24 (2), 227 241. Mikesell, L., Bromley, E., & Khodyakov, D. (2013). Ethical community engaged research: a literature review. American Journal of P u blic H ealth , 103 (12), e7 e14 . Netemeyer, R. G., Bearden, W. O., & Sharma, S. (2003). Scaling procedures: Issues and applications . Thousand Oaks, CA: SAGE P ublications . O'Toole Jr, L. J. (1997). Treating networks seriously: Practical and research based agendas in public administration. Public Administration Review , 57 (1), 45 52. O'Toole, L. J. (2015). Networks and n etworking: The Public a dministrative a ge ndas. Public Administration Review , 75 (3), 361 371. O'Toole, L. J., & Meier, K. J. (2004 a ). Desperately seeking Selznick: Cooptation and the dark side of public management in networks. Public Administration Review , 64 (6), 681 693. O'Toole, L. J., & Meier, K. J. (2004 b ). Public management in intergovernmental networks: Matching structural networks and managerial networking. Journal of Public Administration Research and Theory , 14 (4), 469 494. O'Toole Jr, L. J., & Meier, K. J. (2006). Networking in the penumbra: Public management, cooptative links, and distributional consequences. International Public Management Journal , 9 (3), 271 294. Orth, U. (2002). Secondary victimization of crime victims by criminal proceedings. Social Justice Research , 15 (4), 313 325. Osborne, D. (1993). Re inventing government. Public Productivity & Management Review , 16(4), 349 356. Ostrom, V., & Ostrom, E. (1971). Public choice: A different approach to the study of public administration. Public Administration Review, 31(2), 203 216 . Ostrower, F. ( 2005). The reality underneath the buzz of partnerships . Stanford, CA: Stanford Social Innovation Review . Padgett, J. F., & Ansell, C. K. (1993). Robust Action and the Rise of the Medici, 1400 1434. American Journal of Sociology, 98(6), 1259 1319.

PAGE 111

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! "+# ! Podo lny, J. M., & Page, K. L. (1998). Network forms of organization. Annual Review of S ociology , 24(1), 57 76. Powell, W. (2003). Neither market nor hierarchy. In Handel, M. J. The sociology of organizations: classic, contemporary, and critical readings , 1 04 117. Thousand Oaks, Ca: SAGE Publications. Preacher, K. J., & MacCallum, R . C. (2003). Repairing Tom Swift's electric factor analysis machine. Understanding statistics: Statistical I ssues in Psychology, Education, and the Social Sciences , 2 (1), 13 43. Provan, K. G., Fish, A., & Sydow, J. (2007). Interorganizational ne tworks at the network level: A review of the empirical literature on whole networks. Journal of Management , 33 (3), 479 516. Provan, K. G., & Kenis, P. N. (2008). Modes of network governance: Structure , management, and effectiveness. Journal of Public Administration Research and Theory , 18(2), 229 252. Provan, K. G., & Lemaire, R. H. (2012). Core concepts and key ideas for understanding public sector organizational networks: Using research to inform scholarship and practice. Public Administration Review , 72 (5), 638 648. Provan, K. G., & Milward, H. B. (1995). A preliminary theory of interorganizational network effectiveness: A comparative study of four community menta l health systems. Administrative S cience Q uarterly , 40(1), 1 33. Provan, K. G., & Milward, H. B. (2001). Do networks really work? A framework for evaluating public L sector organizational networks. Public A dministration R eview , 61 (4), 414 423. Provan, K. G., & Sebastian, J. G. (1998). Networks within networks: Service link overlap, organizational cliques, and network effectiveness. Academy of Management Journal , 41 ( 4), 453 463. Raab, J., & Milward, H. B. (2003). Dark networks as problems. Journal of P ublic A dministration R esearch and T heory , 13 (4), 413 439. Raab, J., Mannak, R. S., & CambrÂŽ, B. (2015). Combining structure, governance, and context: A Configurational approach to network effectiveness. Journal of Public Admini stration Research and Theory , 25 (2), 479 511. Ragin, C. C. (2000). Fuzzy set social science . Chicago, Il: University of Chicago Press. ------(2008). Redesigning social inquiry: Fuzzy sets and beyond (pp. 190 212). Chicago , Il : University of Chicago Pr ess.

PAGE 112

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! "+$ ! Rhodes, R. A. (1997). Understanding governance: policy networks, governance, reflexivity and accountability . London , United Kingdom : Open University Press. Rihoux, B., & Ragin, C. C. (2009). Configurational comparative methods: Qualitative compara tive analysis (QCA) and related techniques . New York: Thousand Oaks, CA: SAGE Publications . Riker, W. H. (1962). The theory of political coalitions (Vol. 578). New Haven: Yale University Press. Roberts, D. (2008). Human insecurity: G lobal structures of violence . New York: Palgr ave Macmillan. Roberts, N. (2000). Wicked problems and network approaches to resolution. International P ublic M anagement R eview , 1 (1), 1 19. Robinson, S. E. (2006). A decade of treating networks seriously. Policy Studies Journal , 34 (4), 589 598. Salda–a, J. (2015). Thinking qualitatively: Methods of mind . Thousand Oaks, CA: SAGE Publications . Sandstršm, A., & Carlsson, L. (2008). The performance of policy networks: the relation between network structure and network performance. Policy Studies Journal , 36 (4), 497 524. Saz L Carranza, A., Salvador Iborra, S., & Albareda, A. (2015). The power d ynamics of m andated network administrative o rganizations. Public Administration Review , 76(3), 449 462. Schneider, C. Q., & Wagemann , C. (2003). Improving i nference with a ' t wo step' a pproach: Theory and l imited d iversity in fs/QCA. European University Institute Working Paper SPS 2003/7. Florence: Italy. ----. (2010). Standards of good practice in qualitative comparative analysis (QCA) and fuzzy sets. Comparative Sociology , 9 (3), 397 418. ----. (2012). Set theoretic methods for the social sciences: A guide to qualitative comparative analysis . Cambridge, Ma: Cambridge University Press. ----. (2013). Doing justice to logical remainders in QCA: Moving beyond the standard analysis. Political Research Quarterly , 66(1), 211 220. Scott, J. (1988). Social network analysis. Sociology , 22 (1), 109 127.

PAGE 113

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! "+% ! Sheldon Sherman, J. A. (2012). The Missing'P': Prosecution, Prevention, Protect ion, and Partnership in the Trafficking Victims Protection Act. Penn State Law Review , 117 (443). Siddiki, S. N., Carboni, J. L., Koski, C., & Sadiq, A. A. (2015). How Policy Rules Shape the Structure and Performance of Collaborative Governance Arrangeme nts. Public Administration Review , 75 (4), 536 547. Simon, H. A. (1947). Administrative behavior (Vol. 1). New York: Free Press. ------. (1983). Reason in human affairs . Malibu, Ca: Stanford University Press. StataCorp, L. P. (2012). Stata base reference manual (Vol. 13). College Station, Tx: Stata Press Publication. Su‡rez, D. F. (2011). Collaboration and professionalization: The contours of public sector funding for nonprofit organizations. Journal of Public Administration Research a nd Theory , 21 (2), 307 326 . Tashakkori, A., & Teddlie, C. (2008). Quality of inferences in mixed methods research: Calling for an integrative framework. In Bergman, M. M. (Ed.), Advances in Mixed Methods Research , (101 119). Thousand Oaks, CA: SAGE Public ations. ----. (Eds.). (2010). Sage handbook of mixed methods in social & behavioral research . Thousand Oaks, CA: SAGE Publications. Thiem, A. (2016). "Standards of Good Practice and the Methodology of Necessary Conditions in Qualitative Comparative A nalysis." Political Analysis 24 (4):478 84 . Thompson, B. (2004). Exploratory and confirmatory factor analysis: Und erstanding concepts and applications . New York: American Psychological Association. Turrini, A., Cristofoli, D., Frosini, F., & Nasi, G. (2010). Networking literature about determinants of network effectiveness. Public Administration , 88 (2), 528 550. UN General Assembly, Universal Declaration of Human Rights , 10 December 1948, 217 A (III), available at: http://www.refworld.org/doci d/3ae6b3712c.html [accessed 24 June 2015] . U.S. Department of State. (2010). Trafficking in Persons Report . Washington, DC: United States Department of State .

PAGE 114

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! "+& ! U.S. Government Accountability Office. (2006). Human trafficking: Better data, strategy, and reporting needed to enhance U.S. anti trafficking efforts abroad (Publication No. GAO 06 825). Report to t he Chairman, Committee on the Judiciary and the Chairman, Committee on International Relations, House of Representatives, Washington DC. V an Bueren, E. M., Klijn, E. H., & Koppenjan, J. F. (2003). Dealing with wicked problems in networks: Analyzing an environmental debate from a network perspective. Journal of Public Administration Research and Theory , 13 (2), 193 212. Van de Ven, A. H. (2007). Engaged scholarship: A guide for organizational and social research: a guide for organizational and social research . Oxford, UK: Oxford University Press. Velicer, W. F., & Fav a, J. L. (1998). Affects of variable and subject sampling on factor pattern recovery. Psychological M ethods , 3 (2), 231.

PAGE 115

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! "+' ! A PPENDIX A: Terms Human Trafficking: The recruitment, harboring, transportation, provision, or obtaining of a person for one of the three following purposes: labor or service through the use of force, fraud, or coercion for the purpose of subjection to involuntary servitude, peonage, debt bondage, or slavery; or a commercial sex act through the use of force, fraud, or c oercion; or if the person is 18 years of age, any commercial sex act, regardless of whether any form of coercion is involved. These definitions do not require that a trafficking victim be physically transported from one location to another. Network: Prov an and Kenis (2008) specifically defined networks as "groups of three or more legally autonomous organizations that work together to achieve not only their own goals but also a collective goal" (p 231). This term is considered equivalent to partnership in this study. Prevention : Prevention measures increase awareness, advocacy and education towards addressing a community's systemic vulnerability to a continuum of exploitation, including human trafficking. A good prevention plan recognizes that exploitation and human trafficking are symptoms of root causes like poverty, gender inequality and other forms of oppression which create vulnerable populations in the first place. Prosecution : Prosecution measures ensure the creation and implementation of laws which address the continuum of labor exploitation and the pursuit of criminal punishments for such cases, treating human trafficking as exploitation of victims rather than recruitment/transportation of workers or people in prostitution.

PAGE 116

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! "+( ! Protection: Protection m easures ensure that human victims of human trafficking are provided access to: (at a minimum) health care, legal aid, social services and education in ways that are not prejudicial against victims' rights, dignity, or psychological well being. Protection a lso means creating an environment (social, political and legal) that fosters the protection of victims of trafficking. Partnership: Partnership measures acknowledge that combating human trafficking requires a comprehensive response through the cooperation of multiple sectors. Partnerships bring together diverse experiences, amplify messages, and leverage resources. For the purpose of this survey, an anti human trafficking partnership refers to a É [e.g. a cooperative relationship between two or more organi zations established for the purpose of jointly combating human trafficking in some way.] Please use this definition as you answer the questions in this section.

PAGE 117

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! "+) ! APPENDIX B: LCHT Survey Description Subjects are law enforcement, prosecutors, government offi cials, service providers, and community members involved in the anti trafficking movement throughout the United States. These subjects will be asked to take a survey to answer questions about the services available for people who have experienced human tra fficking, the ways in which anti human trafficking efforts are approached by the criminal justice system, the various prevention efforts for human trafficking and the various partnerships that exist in the anti human trafficking movement. All subjects wil l be sent a survey electronically through their email in which they will receive an explanation of the study and an invitation to participate in the study. In the email, they will be directed to a Zoomerang link for subjects to click on to initiate the sur vey. Service providers from Colorado will additionally complete the Community Needs Assessment Survey, which will also be distribute d electronically. All subjects will be directed to a link to the survey in the program Zoomerang that will contain an informed consent. The sample size for this study is 1500. Subjects will be identified through literature reviews, website reviews of an ti human trafficking agencies, suggestions made by pre eminent colleagues in the field and through recommendations from the Laboratory to Combat Human Trafficking's Advisory Bo ard for the Colorado Project. Subjects will be sent an email explaining the stu dy and inviting them to participate two weeks prior to the survey being distributed. They will be directed to a link to the survey in Zoomerang that will contain an informed consent.

PAGE 118

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! "+* ! A PPENDIX C: Factor Analysis Robustness Details Appendix C.1: Principle Factor Graphs 0 1 2 3 4 Eigenvalues 0 5 10 15 Number Scree plot of eigenvalues: IPF

PAGE 119

,-../.0!12345!6789:,;!<==,:<>?7@!8:!9A>;7B!=,:CD7E@ ! ! ""+ ! 0 1 2 3 4 Eigenvalues 0 5 10 15 Factor Factor Analysis Parallel Analysis Parallel Analysis: IPF Table C.2: Chronbach's Alpha Without Low Turnover Item Obs Sign item test correlation item rest correlation average interitem correlation alpha Trust 73 + 0.92 0.85 0.71 0.88 Good Communication 73 + 0.92 0.85 0.71 0.88 Resolve Conflict 73 + 0.91 0.84 0.71 0.88 Achieved Past Goals 72 + 0.83 0.71 0.81 0.93 Test Scale 0.73 0.92