The Pennsylvania State University The Graduate School College of Information Sciences and Technology Humanitarian Information Management Network Effectiveness: An Analysis at the Organizational and Network Levels A Dissertation in Information Sciences and Technology by Louis-Marie Ngamassi Tchouakeu Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy August 2011
238
Embed
Humanitarian Information Management Network Effectiveness ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The Pennsylvania State University
The Graduate School
College of Information Sciences and Technology
Humanitarian Information Management Network
Effectiveness:
An Analysis at the Organizational and Network Levels
A Dissertation in
Information Sciences and Technology
by
Louis-Marie Ngamassi Tchouakeu
Submitted in Partial Fulfillment
of the Requirements
for the Degree of
Doctor of Philosophy
August 2011
ii
The dissertation of Louis-Marie Ngamassi Tchouakeu was reviewed and approved* by
the following:
Carleen Maitland
Associate Professor of Information Sciences and Technology
Dissertation Advisor
Chair of Committee
Andrea Tapia
Associate Professor of Information Sciences and Technology
Lynette Kvasny
Associate Professor of Information Sciences and Technology
Wenpin Tsai
Professor of Business Administration
Mary Beth Rosson
Professor of Information Sciences and Technology
Graduate Director for Information Sciences and Technology
*Signatures are on file in the Graduate School
iii
ABSTRACT
Massive international response to humanitarian crises such as the South Asian Tsunami
in 2004, the Hurricane Katrina in 2005 and the Haiti earthquake in 2010 highlights the
importance of humanitarian inter-organizational collaboration networks, especially in
information management and exchange. Despite more than a decade old call for more
research on the effectiveness of inter-organizational networks in the nonprofit context, to
date limited work has been done. The objective of this dissertation is to develop a theory
that provides a better understanding of organizational and network effectiveness in the
humanitarian relief field. The study deals with two broad research questions. The first
research question focuses on the relationship between network structural characteristics
and network effectiveness. The second research question concerns organizational
effectiveness and focuses on the relationship between organizational internal
characteristics (and especially the availability of information technology), ego-network
characteristics, network structural characteristics and effectiveness. To answer these
research questions, I used a multi-method research design that applies social network
analytic techniques in combination with statistical analyses (correlation and regression)
and content analysis to analyze data collected through multiple sources including a web-
based survey, semi-structured interviews, and database search. At the network level of
analysis, my findings extend a previous model for assessing network effectiveness in the
humanitarian relief field. At the organizational level of analysis, my research proposes
an integrated approach for assessing effectiveness that takes into account the
characteristics of organization but also those of the network in which the organization is
embedded. My study also highlights the catalytic role of information technology on
organizational effectiveness in humanitarian information management and
exchange. The dissertation concludes by highlighting both theoretical and practical
contributions and by suggesting directions for future research.
iv
TABLE OF CONTENTS
LIST OF FIGURES .............................................................................................................VII
LIST OF TABLES ............................................................................................................... IX
ACKNOWLEDGEMENTS .................................................................................................... X
3.2.2.1 GOAL MODEL .......................................................................................... 26 3.2.2.2 SYSTEM RESOURCE MODEL ........................................................................ 27 3.2.2.3 INTERNAL PROCESSING MODEL ................................................................... 28 3.2.2.4 MULTIPLE CONSTITUENCIES MODEL ............................................................. 29
3.3 INTER-ORGANIZATIONAL NETWORK EFFECTIVENESS .................................................................. 31 3.3.1 DEFINING NETWORK EFFECTIVENESS ............................................................................ 31 3.3.2 MODEL OF NETWORK EFFECTIVENESS .......................................................................... 33
3.3.2.1 PERFORMANCE GAP MODEL ....................................................................... 33 3.3.2.2 PROVAN & MILWARD MODEL .............................................................................. 33
3.3.2.3 PRINCIPLES AGENTS MODEL ................................................................................. 34
3.3.2.4 STUCTURALIST PERSPECTIVE MODEL ..................................................................... 35
4.5 RESEARCH MODELS AND HYPOTHESES .................................................................................. 54 4.5.1 NETWORK CHARACTERISTICS AND EFFECTIVENESS ................................................................... 54
5.5 DATA COLLECTION ............................................................................................................ 75 5.5.1 SURVEY DATA ........................................................................................................................ 75
5.5.2 INTERVIEW DATA ................................................................................................................... 77
5.5.3 DATABASE DATA ................................................................................................................... 78
5.6 DATA ANALYSIS TECHNIQUES .............................................................................................. 79 5.6.1 SOCIAL NETWORK TECHNIQUES .............................................................................................. 79
7 CONCLUSIONS AND DIRECTIONS FOR FUTURE RESEARCH ...................................... 167
7.1 INTRODUCTION ..............................................................................................................167 7.2 SUMMARY OF THE LITERATURE ..........................................................................................167 7.3 KEY FINDINGS .................................................................................................................168 7.4 CONTRIBUTIONS .............................................................................................................174 7.5 LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH ...........................................................179
APPENDIX A: INFORM CONSENT FORM FOR SOCIAL SCIENCE RESEARCH .............................................199 APPENDIX B: LETTER-EMAIL SENT TO POTENTIAL SURVEY PARTICIPANTS .............................................201 APPENDIX C: SURVEY QUESTIONNAIRE ........................................................................................202 APPENDIX D: INTERVIEW GUIDE .................................................................................................226
vii
LIST OF FIGURES
Figure 1: Level of NGOs coordination .................................................................................... 17
Figure 2. A Preliminary model of network effectiveness ...................................................... 34
Figure 3. Relationships between network effectiveness at different levels of network
analysis and influence by key stakeholders ..................................................................... 35
Figure 4: Research Model for Network Level of Analysis ..................................................... 64
Figure 5: Research Model for Organizatioanl Level of Analysis ............................................ 64
Figure 6: Global Symposium Project Collaboration Sub-Networks ....................................... 70
Figure 7: Global Symposium Advice Sub-Networks .............................................................. 70
Figure 8: United Nations Agencies Network Structure ........................................................... 77
In the particular realm of NGO coordination, initiatives are seen as a solution to
duplication of efforts in assistance projects, badly planned and implemented relief efforts,
and the lack of knowledge among humanitarian organizations on the actual situation in
which they operate. These initiatives entail developing strategies, determining
objectives, planning, sharing information, the division of roles and responsibilities, and
mobilizing resources. They are also concerned with synchronizing the mandates, roles
and activities of the various stakeholders and actors at higher organizational levels. In a
nutshell, NGO coordination is intended to ensure that priorities are clearly defined,
resources more efficiently utilized and duplication of effort minimized; the ultimate goal
being to provide coherent, effective and timely assistance to those in need (Harpviken et
al., 2001).
Coordination among NGOs, as well as between NGOs and other humanitarian actors,
takes place at different levels. Harpviken et al., (2001) identify these levels as
international, national, regional and local. At the international level, the formulation of
policy, general guiding principles and strategies are of concern. At the national level,
coordination typically revolves around program development and policy articulation. At
this level, local groups are typically less involved, while United Nations agencies,
government departments and NGOs representatives assume a central role. Coordination
at the local level usually takes place between representatives from NGOs, United Nations
agencies, and local communities. It is at the local level where humanitarian priorities can
be most readily identified and articulated. Figure 1 below depicts these different levels of
coordination, within which inter-organizational relationships may vary, depending on the
level of coordination pursued. My study focuses on coordination at the international
level.
17
Local level
National level
International level
Type of Coord.
Project/Program
coordination
Coordinated
activities
Actors:
IGOs,
NGOs,
main offices
CBs, UN
agencies,
UN
Type of Coord.
Program/ policy
coordination
(Standards)
Actors:
IGOs,
NGOs,
HQ,
Int.CBs,
UN,
Donors
Type of Coord.
Policy and norms
Actors:
IGOs,
NGOs,
CBs, UN
agencies,
UN, Donors
IGOs = Inter-governmental Organizations
CBs = Coordination Bodies
UN = United Nations
HQ = headquarters
Figure 1: Level of NGOs coordination
Source: Author adaptation from Harpviken et al., (2001)
Inter-organizational Coordination Forms: Identifying and classifying the various forms of
inter-organizational coordination has been a subject of research in both the for- and non-
profit domains. Research on for-profit organizations has identified two general structures
of coordination (Malone, 1987; Thompson et al., 1991). The first is a hierarchical
coordination structure, characterized by long-lasting relationships with fixed rules of
behavior and clear authoritative relationships. Put simply, one organization has control
over the other(s). The second is a “market” coordination structure, in which all
organizations are fully autonomous and make decisions in their own interest.
In the non-profit domain, research has similarly identified multiple structures (Donini &
Niland, 1999). The first is "coordination by command," in which the lead NGO has
authority to pursue coordination through the use of carrots or sticks and possesses strong
leadership abilities. In such a situation, a central authority has the power to define the
agenda, instigate preferences and enforce sanctions. Power can come in the form of
control of information or resources, but also the institutionalized legal means, through
which preferences might be implemented. The second form is "coordination by
18
consensus". In this form, organizations develop agreed-upon guidelines and standards to
achieve similar goals, and there is no authority to enforce compliance. The last form,
"coordination by default" describes ad-hoc coordination in which a division of labor is
generally the only exchange of information among actors. Obstacles to inter-
organizational coordination may vary depending on these various forms of coordination.
Alternatively, research on coordination structures in the humanitarian sector finds that
structure within NGOs themselves. Enjorlas (2008) argues that collectively NGOs on
their own serve as coordination structures. Due to the nature of their individual
governance structures, they reinforce the norm of reciprocity; making possible the
pooling of resources and, because of these features, thereby facilitate collective action
oriented toward public or mutual interest as well as advocacy. Moreover, this nonprofit
governance structure is also compatible with other types of coordination mechanisms,
and thus NGOs are able to operate in complex environments, mobilizing resources from
market operations, governmental subsidies, or from reciprocity (Enjorlas, 2008).
2.4 Humanitarian Information Management and Exchange
My research explored inter-organizational networks in the Global Symposium, a
community of interest in humanitarian information management and exchange
spearheaded by the United Nations Office for the Coordination of Humanitarian Affairs
(UNOCHA). UNOCHA initiated a Global Symposium in recognition of the centrality of
information management to effective and timely response to humanitarian disasters.
Timely and accurate information is recognized as integral to humanitarian action in both
natural disasters and complex emergencies. The international community's ability to
collect, analyze, disseminate, and act on key information is fundamental to an effective
response. Better information, leading to improved responses, directly benefits affected
populations. Over time, improved assessment of impacts and responses through better
data collection and management contributes to a more complete global database on
disaster impacts, leading to better risk assessment and prevention and preparedness
activities.
19
The goal of the Global Symposium is to foster collaboration on humanitarian information
management projects and to disseminate best practices of information exchange. This
community also aims to (i) sensitize its members on the critical aspects of humanitarian
information management preparedness, (ii) facilitate headquarter-field partnerships and
(iii) advocate for more funding from donors for humanitarian information management
related projects. My study focuses on members which actively participate in the activities
in the community and have developed collaborative relationships with other members
within the community.
The Global Symposium held a series of conferences and workshops, organized by
UNOCHA. The series began in 2002 as a meeting of humanitarian information
management professionals and was followed by a series of regional meetings intended to
bring humanitarian information management principles (Table 1) and best practices to a
wider range of humanitarian organizations and in particular bring together practitioners in
the field, as opposed to only headquarters staff. The second meeting of the Global
Symposium was held in October 2007 and included three days of working group
meetings, designed to update the principles and best practices and identify an agenda for
further development of humanitarian information management (HIM).
Principle Description
Accessibility Humanitarian information and data should be made accessible to all humanitarian actors by applying easy-to-use formats and by translating information into common or local languages when necessary. Information and data for humanitarian purposes should be made widely available through a variety of online and offline distribution channels including the media.
Inclusiveness
Information management and exchange should be based on a system of collaboration, partnership and sharing with a high degree of participation and ownership by multiple stakeholders, especially representatives of the affected population.
Inter-operability All sharable data and information should be made available in formats that can be easily retrieved, shared and used by humanitarian organizations.
Accountability Users must be able to evaluate the reliability and credibility of data and information by knowing its source. Information providers should be responsible to their partners and stakeholders for the content they publish and disseminate.
Verifiability Information should be accurate, consistent and based on sound methodologies, validated by external sources, and analyzed within the proper contextual framework.
20
Principle Description
Relevance Information should be practical, flexible, responsive, and driven by operational needs in support of decision-making throughout all phases of a crisis.
Objectivity Information managers should consult a variety of sources when collecting and analyzing information so as to provide varied and balanced perspectives for addressing problems and recommending solutions.
Humanity Information should never be used to distort, to mislead or to cause harm to affected or at-risk populations and should respect the dignity of victims.
Timeliness Humanitarian information should be collected, analyzed and disseminated efficiently, and must be kept current.
Sustainability Humanitarian information and data should be preserved, cataloged and archived, so that it can be retrieved for future use, such as for preparedness, analysis, lessons learned and evaluation.
Table 1: Principles of Humanitarian Information Management and Exchange.
The Geneva 2002 meeting was followed by a series of regional workshops in Bangkok
(2003), Panama (2005) and Nairobi (2006). While the issues confronting the
humanitarian community are global in scope, there are regional differences in both the
types of problems as well as the appropriate solutions. Each workshop focused on
information initiatives and tools in their regional context, each region with its different
vulnerabilities and response capacities. The goals of these workshops were to (i) bring
together regional information management professionals in order to strengthen the
professional community of practice, (ii) discuss the principles and best practices in
information management, especially those which have been developed at the regional
level, and (iii) deepen understanding of the regional issues and priorities that will help
build a plan for improving information exchange in the region. The recommendations
from these workshops reinforced the need for attention to the promotion of standards,
user requirements, quality of information, appropriate responses, tools and technology,
and strong partnerships.
2.5 Humanitarian Collaboration and Coordination Challenges
Research on barriers to inter-organizational coordination and collaboration has been
undertaken in both general organizational contexts (e.g. Burbridge & Nightingale, 1989;
Xiao, 2006; Quarantelli, 1982; Thompson, 1967), as well as among organizations in the
nonprofit context (e.g. Bennett, 1995; Bui et al., 2000; Foster-Fishman et al., 2001; Saab
et al., 2008; Uvin, 1999; Van Brabant, 1999). After an analysis of the literature,
Ngamassi et al., (2011) found a fairly consistent set of eight coordination and
collaboration barriers (Table 2). They include (i) bureaucratic and turf-protection, (ii)
divergent goals and conflicting interests, (iii) resource dependency, (iv) coordination
cost, (v) information and communication issues, (vi) assessing and planning joint
activities, (vii) competition for resources, and (viii) emergency response time.
Bureaucratic barriers and turf-protection refer to the desire to maintain autonomy and
thus avoid having individuals in other organizations interfere within one's own
organization. Burbridge & Nightingale (1989) note a common fear among organizations
is that coordination may somehow result in a take-over or a loss of decision-making
autonomy. Furthermore, the discipline of coordination can limit maneuverability, and
hence poses a major challenge (Uvin, 1999). Coordination may be perceived as
increasing bureaucracy, generating institutional resistance among bureaucratically
burdened NGOs (Van Brabant, 1999).
A common problem in inter-organizational collaboration is that divergent goals or an
over-emphasis on individual organizational goals as opposed to those of beneficiaries
may lead to conflicting interests (Bennett, 1995; Bui et al., 2000; Quarantelli, 1982; Saab
et al., 2008; Van Brabant, 1999). Goal conflicts occur when a party seeks divergent or
incompatible ends. Further, divergent goals may also lead to an exacerbation of turf
issues or other coordination problems (Bui et al., 2000).
Resource dependency is both a motivation for and barrier to coordination (Crowston,
1997; Dawes et al., 2004; Thompson, 1967). Interdependencies, whether of the pooled,
sequential or reciprocal type, require coordination (Thompson, 1967). However, at the
same time they can create problems for coordination and constrain the efficiency of task
performance (Crowston, 1997). One of these problems is the associated cost of
22
coordination, as to be effective it is time and staff intensive and the benefits must
outweigh these costs (Aldrich, 1972; Bennett, 1995; Van Brabant, 1999).
Coordination cost is yet another barrier that hampers coordination among organizations.
Inter-organizational coordination is believed to limit an organization because scarce
resources and energy have to be invested in the maintenance of relationships with other
organizations. Negotiation of resources allocation can lead to difficult bargaining among
parties engaged in coordinated activities. Usually, organizations find it difficult to
allocate scare resources (Bui et al., 2000). Aldrich (1972) argued that it is costly for
organizations to initiate and/or maintain linkages with other organizations. For example,
the costs can be seen as in term of additional staff-time necessary to attend a joint board
of directors’ meeting; or the additional funds necessary to participate in joint database.
According to Uvin (1999), the high cost in time and money that effective co-ordination
entails constitute one of the major barriers to inter-organization coordination.
Another frequently encountered barrier is related to the availability and the quality of
information. This is usually due to the inconsistency in data collection and management
across organizations and to the mismatch between the informational demands and
supplies (De Bruijn, 2006; Fisher & Kingma, 2001). According to Bui, et al., (2000),
there are varying levels of mistrust, misrepresentation of facts, and incomplete
information exchange among organizations. Further, the high level of uncertainty in
humanitarian operations likely requires greater amounts of information to be processed
among decision makers (Galbraith, 1977).
General assessment and planning of joint activities can lead to disagreement about the
means and the ends of a coordinated activity (Bui, et al., 2000). Situations tend to worsen
when organizations are unsure of their role, and act independently, without consulting or
coordinating with others. Joint activities must also confront problems of understanding,
which emanate from the fact that participants in inter-organizational relationships are
accustomed to different structures, cultures, functional capabilities, cognitive frames,
terminologies, and management styles and philosophies (Vlaar et al., 2006).
23
In addition to the resources related to coordination itself, competition for scarce resources
in general may inhibit the initiation of inter-organizational coordination generally (Uvin,
1999; Van Brabant, 1999). Given the increasing numbers of NGOs, combined with
decreasing overseas development assistance budgets, competition for funding between
organizations is heating up (Salm, 1999; Van Brabant, 1999).
Finally, response time is considered yet another obstacle to coordination among
organization. Coordination is often perceived as increasing response time especially in
case of emergency. According to Van Brabant (1999), there is the fear that the
coordination effort will cause delays in providing relief. Comfort (1990) observed that
coordination activities generated delays in response in the four events she analyzed.
Thus, inter-organizational coordination between international humanitarian NGOs will
seek to share information, resources and responsibilities that through more efficient use
of resources and minimization of duplicate activities will provide effective and timely
assistance to those in need (Harpviken et al., 2001). This coordination can occur at
multiple levels and may be carried out through one of several forms, including command,
consensus or default. Whatever the form, it must contend with a wide range of
challenges.
Barriers Issues Authors
Bureaucratic and turf protection
Desire to maintain autonomy and thus avoid having individuals in other organizations interfere within one's own organization
Burbridge and Nightingale (1989) (Uvin, 1999). (Van Brabant, 1999).
Divergent goals and Conflicting interests
Divergent goals or an over-emphasis on individual organizational goals
Bennett 1995; Bui et al, 2000; Quarantelli, 1982; Saab et al, 2008; Van Brabant, 1999.
Resource dependency Interdependencies require coordination but at the same time they can create problems for coordination and hamper performance.
Crowston, 1997; Dawes et al., 2004; Thompson 1967). Aldrich 1972; Bennett, 1995; Van Brabant 1999
24
Barriers Issues Authors
Coordination cost Scarce resources have to be invested in the maintenance of relationships with other organizations.
Bui et al, 2000; Aldrich,1972; Uvin, 1999
Information and communication issues,
Information availability and accessibility,
Information quality,
Information Sharing
Information system quality,
Standards and interoperability
Systems integration
De Bruijn, 2006; Fisher & Kingma, 2001; Bui, et al 2000; Galbraith, 1977.
Assessing and planning joint activities
Disagreement about the means and the ends of a coordinated activity
Bui, et al, 2000; Vlaar et al., 2006
Competition for resources
Competition for scarce resources may inhibit the initiation of inter-organizational coordination
Uvin, 1999; Van Brabant, 1999; Salm, 1999.
Emergency response time
Coordination is often perceived as increasing response time especially in case of emergency
Van Brabant, 1999; Comfort, 1990.
Table 2: Summary of Inter-Organizational Coordination/Collaboration Challenges
25
3 REVIEW OF RELEVANT LITERATURE
3.1 Introduction
As discussed in the introductory chapter, this study is situated in the broader context of
research on inter-organizational networks in the non-profit sector. I investigate the
organizational attributes and network structural characteristics that explain effectiveness.
In this chapter, I review the relevant literature. The chapter is made up of two sections.
The first (Section 3.2) is related to effectiveness at the organizational level of analysis
while the second (Section 3.3) is concerned with effectiveness at the network level of
analysis.
3.2 Organizational effectiveness
3.2.1 Defining Organizational Effectiveness
Although researchers have devoted considerable amount of time investigating
organizational effectiveness, the construct remains elusive. In the literature, there is a
wide range of definitions to this construct (for a review, see Goodman et. al, 1977; Cho,
2007). There is no consensus on the criteria of measuring effectiveness among
researchers (Quinn & Rohrbaugh, 1983; Scott, 1992). Moreover, debates still exit about
the primary factors that constitute organizational effectiveness (Goodman et al., 1977;
Rainey & Steinbauer, 1999) and about the validity of measuring the construct (Goodman
et al., 1983; Steers, 1975). In addition, there is no single theory of organizational
effectiveness (Goodman et al., 1983), rather each paradigm of organizational behavior
generates its own model or criterion of effectiveness (D’Aunno, 1992).
In the literature, there is a wide range of definition to the concept of organizational
effectiveness (for a review, see Goodman et Al., 1977; Cho, 2007). There is no consensus
on the criteria of measuring effectiveness among researchers and no clear classification of
the different levels of effectiveness. For Goodman et al., (1977) organizational
26
effectiveness is measured in terms of the organization's ability to satisfy constraints and
meet organizational goal. Reviewing the literature on organizational effectiveness,
Cameron (1986a; 1986b) describes the concept of effectiveness as theory-bound,
multidimensional, interest-driven, and paradoxical in nature.
3.2.2 Models of Organizational Effectiveness
A variety of different models of organizational effectiveness have been used however,
four major models dominate the literature. They include the goal model, the systems-
resource model, internal processing model, and the multiple constituencies’ model.
Below, I briefly review these different models.
3.2.2.1 Goal Model
The problem of organizational effectiveness has traditionally been studied by means of
the goal approach (Parson 1964; Price, 1971; Cameron & Whetten 1981). The
distinguishing characteristic of the goal model is that it defines effectiveness in terms of
the degree of goal achievement. The greater the degree to which an organization
achieves its goals the greater is its effectiveness. The goal model approach to
organizational effectiveness assumes that organizations are designed to achieve certain
goals, both formally specified and implicit (Perrow 1965; Sowa et al., 2004). The model
also assumes that organizations have goals that are clearly defined and easily measurable
and that data relevant to those measures can be collected, processed and applied in a
timely and appropriate manner (Herman & Renz, 2004a; 2004b). The model views
organizations as a rational set of arrangements oriented toward achieving a goal.
Yuchtman & Seashore (1967) distinguish two components of the Goal Model approach to
organizational effectiveness. The first component is the "prescribed goal approach".
According to the authors, this component focuses on the formal charter of the
organization, or in some category of its personnel as the most valid source of information
27
concerning organizational goals. The second component is the "derived goal approach".
In this component, the researcher derives the goal of the organization from his/her theory.
The Goal Model of organizational effectiveness is suitable for those organizations where
activity is shaped by a focus on output (Cameron & Whetten 1981), and organizational
effectiveness is generally operationalized in term of productivity or efficiency (Scheid &
Greenley, 1997). Organizational effectiveness in organizations with clearly defined and
easily measurable goals may be assessed using the goal model (Cameron & Whetten,
1983).
The main criticism to the Goal Model of organizational effectiveness consistently
identified in the literature especially by the adherents of the System Resource Model, has
been that its proponents have not developed measures of effectiveness which can be used
to study many types of organizations. Adherents of the System Resource approach to
organizational effectiveness make two criticisms of the goal approach (see Price, 1971).
First, they say that the goal approach has provided no means to identify organizational
goals; second, they say that the goal approach uses society, not the organization, as the
basis for the evaluation of effectiveness. The absence of general measures is serious
because it hinders the development of theory. The existence of general measures
promotes measurement standardization; measurement standardization, in turn, facilitates
comparison; and comparison, in turn, furthers the development of theory.
3.2.2.2 System Resource Model
The System Resource Model defines effectiveness, not with respect to the degree of goal-
achievement, but in terms of the ability of the organization to exploit its environment in
the acquisition of scarce and valued resources (Yuchtman & Seashore, 1967; Price,
1971). In this model, organizational effectiveness is the degree to which an organization
can preserve its internal integration, adapt to the environment and therefore survive
(Scheid & Greenley, 1997). Organizational effectiveness is positively related to the
ability of the organization to exploit its environment. According to Sowa et al., (2004),
28
in the system resource model of organizational effectiveness, the inputs into an
organization are more important than their outputs because an organization’s ability to
maintain sufficient resources for survival is the most important indicator of effectiveness.
Steers (1975) found that the most common utilized systems criteria of organizational
effectiveness were organizational adaption and flexibility. Cameron & Whetten (1981)
see systems resource models as best fitting organizations where formalization is low or
when environmental turbulence (uncertainty and complexity) is high, and, hence, system
effectiveness precedes and is a prerequisite for goal effectiveness.
The System Resource Model of organizational effectiveness has also been criticized.
Price (1971) outlined three criticisms of this approach. First, he states that the idea of
"optimization" is an important component of effectiveness as conceptualized the
proponents of the systems approach and yet, according to the author, these same scholars
show little concern for trying to measure optimization. Second, Price argues that the
systems oriented researchers have expressed the need for general measures of
effectiveness, but none have developed these general measures that they claim to be so
necessary. Finally, Price believes that the frame of reference used in the analysis process
by the system researchers is somewhat confused. According to the author, the confusion
centers around the difference between a multidimensional approach to effectiveness with
multiple measures of effectiveness, and a multidimensional approach with multiple
measures of a series of different analytical concepts.
3.2.2.3 Internal Processing Model
The Internal Processing Model conceptualizes organizational effectiveness as the absence
of internal strain and a smooth internal functioning of organizations / networks (Lee,
2006). For Alter & Hage (1993), much of the existing government and foundation
sponsored inter-organizational systems has adopted the internal processing model. They
believe that the choice of this model has been based on the assumption that the outcomes
of the system, the product or service, will be of higher quality if the system functions
29
smoothly and with a minimum of conflict. There are currently several attempts to
evaluate service delivery using the internal processing model (Lee, 2006).
3.2.2.4 Multiple Constituencies Model
The Multiple Constituencies’ Model defines organizational effectiveness as the ability of
organizations to satisfy key strategic constituencies in their environment (D’Aunno,
1992; Zammuto, 1984; Sowa et al., 2004). This approach to organizational effectiveness
began to emerge when researchers focused less on the assessment criteria of abstract
dimensions and more on the concrete expression of stakeholders’ expectation (Connolly
et al., 1980; Zammuto, 1984). The model recognizes that an organization comprises
multiple stakeholders or constituents who are likely to use different criteria to evaluate its
effectiveness (Herman & Renz, 1998). Effective organizations are viewed as those which
had accurate information about the expectation of strategically critical constituents and
adapted internal organizational activities, goals, and values to match those expectations
(Scheid & Greenley,1997). In the Multiple Constituency Model, the emphasis is on the
organizations’ ability to satisfy (or adapt to) divergent preferences. The Multiple
Constituency Model conceives of differing groups of stakeholders, such as clients or
customers, board members, staff, volunteers, and funders, as probably having different
goals and requires that researchers recognize the potential differences in their interests
(Herman & Renz, 1998).
The Multiple Constituencies’ Model of organizational effectiveness spawns a large
number of research (Whetten, 1978; Cameron, 1978; Tsui, 1990). According to Cameron
& Whetten, (1983), researchers using this approach encountered four difficult
methodological challenges including (i) When asked individual stakeholders have
difficulty explaining their personal expectations for an organization; (ii) a stakeholder’s
expectations change sometime dramatically, over the time; (iii) a variety of contradictory
expectations are almost always pursued simultaneously in an organization and (iv) The
expectations of strategic constituencies frequently are unrelated, or negatively related, to
their overall judgments of an organization’s effectiveness.
30
I summarize in Table 3 below, the four model traditionally used to study organizational
effectiveness
Model Definition When Useful? Criticisms
An organization is effective to the extent that
The model is most preferred when
Goal model It accomplishes its stated goals Goals are clear, consensual, time-bound, measurable
No means to identify network goals; Absence of general measures for effectiveness.
System resource model
It acquires needed resources A clear connection exits between inputs and performance
Little concern for trying to measure optimization, a big component of effectiveness; No general measures of effectiveness; Confusion centers around the difference between a multidimensional approach to effectiveness with multiple measures of effectiveness, and a multidimensional approach with multiple measures of a series of different analytical concepts.
Internal processing model
It has an absence of internal strain with smooth internal functioning
A clear connection exits between organizational processes and performance
Strategic constituencies model
All strategic constituencies are at least minimally satisfied
Constituencies have powerful influence on the organization, and it has to respond to demands
A stakeholder’s expectations change sometime dramatically, over the time; A variety of contradictory expectations are almost always pursued simultaneously in a network Expectations of strategic constituencies frequently are unrelated, or negatively related, to their overall judgments of an organization’s effectiveness.
Table 3: Models of Organizational Effectiveness
Source: Author Adaptation from Cameron (1986)
31
3.3 Inter-organizational Network effectiveness
3.3.1 Defining Network Effectiveness
The concept of inter-organizational network effectiveness is discussed at length in the
literature. Much of these discussions highlight the difficulties of defining and assessing
administrative decision making – impersonal methods, personal methods, groups
methods- ; and (v) task integration –sequential pattern, reciprocal pattern, team pattern-.
3.3.2.2 Provan & Milward Model
Provan & Milward (1995) developed a model of inter-organizational network
effectiveness through a comparative study of four community mental health networks.
They investigated the relationship between the structure and context of mental health
networks and their effectiveness. Findings from this research suggest that network
effectiveness could be explained by various structural and contextual factors such as
network integration, system stability and environmental resource munificence. Provan
34
& Sebastian (1998), further developed the model, focusing on clique and clique overlap
in the networks. Their findings suggest that achieving integration across an entire
network of organizations is difficult. Their theory is that most effective networks are
those that are integrated at clique or sub-network level.
Network Structure
Centralized integration
Direct, non fragmented
external control
Network Effectiveness
Network Context
System stability
High resource munificence
Figure 2. A Preliminary model of network effectiveness
Source : Provan & Milward (1995)
3.3.2.3 Principles Agents Model
Provan & Milward (2001) propose another approach to assess inter-organizational
network effectiveness based on the Principles Agents theory. In addition to structure and
context, Provan & Milward (2001) examined network effectiveness at different levels.
These three levels are (i) community, (ii) network, and (iii) organization/participant. The
paper argues that organization/participant and network-level effectiveness criteria can be
satisfied by focusing on community-level effectiveness goals. Weech-Maldonado et al.,
(2003) build upon Provan & Milward’s (2001) network effectiveness framework and
Gamm’s (1998) accountability framework to develop a “stakeholder accountability
approach” in assessing network effectiveness. The stakeholder accountability approach
posits that with each level of analysis (community, network, organizational/participant)
there are different effectiveness criteria reflecting the needs of the various stakeholders.
The authors use this approach to evaluate the effectiveness of community health
35
partnerships. Figure 3 below depicts the relationships between network effectiveness at
different levels of network analysis and influence by key stakeholders.
Network-level
effectiveness
Community-level
effectiveness
Agents
Organization/
participant-level
effectiveness
Key Stakeholders
Principals
Clients
Figure 3. Relationships between network effectiveness at different levels of network
analysis and influence by key stakeholders
Source: Provan & Milward (2001)
3.3.2.4 Stucturalist Perspective Model
Sydow & Windeler (1998) define inter-organizational network effectiveness as viable
and acceptable outcome and practices. They argue that network effectiveness form a
structurationist perspective, is more than embedded in social interactions and structures,
it is social in character. They discuss the concept of inter-organizational network
effectiveness in the light of Giddens’ (1984) duality of structure. For Sydow & Windeler
the meaning of the criteria to assess network effectiveness is not simply given, but
necessarily interpreted and ascribed (signification). Moreover, these criteria are always
interest-related and value-laden (legitimation). And finally, they are powerfully (re-)
produced by individual and collective agents (domination).
36
Sydow & Windeler (1998) identify two levels of analysis of network effectiveness
including (i) the level of the individual network firm and (ii) the level of the total inter-
organizational network. On the level of individual network organizations, they argue that
network effectiveness results from that part of the network effect which a particular
network firm is able to appropriate and eventually to represent in its accounts. In this
sense, network effectiveness contributes to organizational effectiveness. On the level of
the total inter-organizational network, network effectiveness depends upon the
effectiveness of all single network firms and upon the augmentation of resources to be
achieved by the differentiation and integration of the entire network (Sydow & Windeler,
1998). For Sydow & Windeler, network effectiveness on this level of analysis usually
evades conventional calculating and accounting practices by taking the efficacy of
network structures into account.
3.3.2.5 Adaptive Capacity Model
Building upon Sydow & Windeler (1998), Staber & Sydow (2002) propose the concept
of adaptive capacity as an appropriate approach to assess organizational and inter-
organizational network effectiveness in highly volatile and complex environments such
as the case in the humanitarian assistance sector. They define adaptive capacity as the
ability of organizations or networks to cope with unknown future circumstances.
Organizations / networks with high adaptive capacity can reconfigure themselves quickly
in changing environments and consequently are more effective. They argue that adaptive
capacity should thus be viewed in relative and dynamic terms. That is, organizations /
networks have adaptive capacity when learning takes place at a rate faster than the rate of
change in the conditions that require dismantling old routines and creating new ones.
Using Giddens’s structuration theory (Giddens, 1984), Staber & Sydow (2002) discuss
multiplexity, redundancy, and loose coupling as important structural dimensions of
adaptive capacity.
37
These different conceptual models for assessing network effectiveness found in the
literature most of the time borrowed from the four models traditionally used to study
organizational effectiveness that I discussed earlier.
3.3.3 Predictors of Network Effectiveness in Nonprofit
Previous research has also identified important predictors of inter-organizational network
effectiveness (Provan & Milward, 1995; Moscovice, et al., 1995; Wright et al., 1995;
Provan & Sebastian, 1998; Schumaker, 2003; Lemieux-Charles et al., 2005). These
predictors could be grouped into two categories, structural and relational. For example,
several authors (e.g., Provan & Milward, 1995; Moscovice, et al., 1995; Wright et al.,
1995; Provan & Sebastian, 1998) highlight the importance of integration of network
members to network effectiveness.
Provan & Sebastian (1998) argued that achieving integration across an entire network of
organizations is difficult. Their findings also suggest that to be most effective, clique
integration must be intensive, involving multiple and overlapping relationships both with
and across organizations that compose the core of a network. Lerch et al., (2006)
investigated the relationships between formal cluster governance and actual networks of
relationships and between multi-dimensional network integration and innovation
activities. The paper applies a multi-level analysis that distinguishes the cluster level
from network and clique levels and accounts for the recursive interplay between
structural properties of these levels and how agents refer to them in inter-organizational
inter-actions. The paper used longitudinal data which allow for studying network
dynamics. Their results were consistent with those of Provan & Sebastian (1998). They
found that multiplex over-lapping cliques provide not only for a fair amount of network
integration, but also a social context conducive for turning complex knowledge of
research organizations into marketable products.
Lemieux-Charles et al., (2005) examined the effectiveness of four community-based,
nonprofit dementia care networks located in Ottawa, Toronto, Hamilton, and the Niagara
38
region. The research focused on the evolution, structure, and processes of the networks
and on how these networks served the needs of care recipients and caregivers who were
using community-based or ambulatory care services provided by acute-care agencies.
Though the authors studied each network as a whole, they also examined the
relationships that existed among groups of agencies within them. The types of
relationships examined were based on activities related to administrative functions and
service delivery functions. Findings of the study suggest that members perceived higher
administrative and service delivery effectiveness when network members shared multiple
ties with members of different groups within the network as opposed to the sharing of ties
across the network. The centralization of network structure was also found to be related
to the perception of service delivery effectiveness.
Morehead (2008) provides insight into the correlates of effectiveness for a type of health
network, vertically integrated rural health networks. The study uses Provan & Milward’s
(2001) framework for evaluating the effectiveness of public-sector organizational
networks to analyze the effectiveness of twenty three rural health networks. One-to-one
interviews, questionnaires, and archives were used to collect data on the networks
sampled. Findings of the study revealed a few significant predictors for the effectiveness
of vertically integrated rural health networks. Financing was found to be the most
important predictor, as it was significant at both the community and network levels. Both
cohesiveness and the number of problems in the rural environment were also found to be
significant predictors but only at the network level. No significant predictors were found
at the organizational level; however, organizational and network-level effectiveness were
found to be strongly correlated with each other. Overall, networks were found to be more
favorable about their effectiveness at the network and organizational levels.
Similarly, the diversity of network membership is also deemed relevant by several
authors (e.g., Moscovice, et al., 1995; 1996; Schumaker, 2003). Schumaker (2003) for
example found that effectiveness is influenced by external and internal factors that are
39
operationalized through external control, technology, structure, and operational process
variables. In Table 4 below, I present a summary of these studies.
Authors Issues Measures of effectiveness Findings/outcomes
Provan & Milward (1995)
Develop a theory to assess network effectiveness
Perception of solving problems Building social capital Decrease service duplication Improve coordination Goal commitment
Networks are more effective when network integration is centralized, external fiscal control by the state is non-fragmented and direct, resources are sufficient, and the overall system is secure
Moscovice et al., (1995)
Develop an approach to study vertically integrated rural health networks
Benefits and costs of health care provision to network’s clients
Questions for further research
Grusky (1995)
Assess networks effectiveness of mental health care delivery networks
Service quality Coverage Comprehensiveness Coordination
The longer key inter-organizational network agency directors have served the more likely the care system was perceived as effective. The more powerful the lead agency relative to other organizations in the network the more likely the system was perceived as effective.
Provan & Sebastian (1998)
Explore the use of clique analysis for explaining network effectiveness.
client outcomes Effectiveness was negatively related to the integration of full networks. In contrast, effectiveness was positively related to integration among small cliques of agencies when these cliques had overlapping links through both reciprocated referrals and case coordination.
Provan & Milward (2001)
Develop a framework to assess network effectiveness at three levels of analysis (community, network, and organization/participant)
Network membership growth Range of service provided Absence of service duplication Relationship strength (multiplexity) Creation and maintenance of network administrative organization (NAO) Integration/coordination of services Cost of network maintenance Member commitment to network goals
A framework with different effectiveness criteria depending on the level of analysis
Schumaker (2003)
Assess networks outcomes of rural health care delivery networks
Gap between best possible and actual practice
Effectiveness increased with network connectivity, decision making methods, and pattern of service delivery. Centrality and network size decrease together where there is little reliance on vertical sources of funds.
Weech-Maldonado et al., (2003)
Develop an approach to assess network effectiveness (stakeholder accountability approach)
Perceived benefit to the various stakeholders of the network
Use the approach to evaluate the effectiveness of community health partnerships
Lemieux-Charles et al., (2005)
Assess the effectiveness of community-based networks
Facilitate sharing Provide opportunity for share program Facilitate administrative information exchange
Perceived effectiveness increased with multiplexed ties among members of different groups within the network. Perceived effectiveness related to the centralization of network structure.
40
Authors Issues Measures of effectiveness Findings/outcomes
Lerch et al., (2006)
Study the emergence and overlap of organizational cliques in an optics/photonics cluster in Berlin-Brandenburg.
The paper applies a multi-level analysis that distinguishes the cluster level from network and clique levels and accounts for the recursive interplay between structural properties of these levels and how agents refer to them in inter-organizational inter-actions. The paper used longitudinal data which allow for studying network dynamics.
Arya & Lin (2007)
Assess the impact of organization characteristics and network structure characteristics on collaboration outcomes
Ability to obtain funding Ability to enhance reputation Ability to meet clients’ needs
High-status organizations are able to derive critical resources from network involvement
Morehead (2008)
Assess networks effectiveness of rural health care delivery networks
Perceived benefit Number or organizations added Number of service provided Existence of NAO
Financing was found to be the most important predictor, of network effectiveness
Table 4: Inter-organizational Network Effectiveness in the Nonprofit Sector
3.4 Issues Identified in the Literature on Effectiveness
After reviewing the literature on effectiveness at organizational and network level, one
general observation is that while several studies have investigated this concept, and
several other have provided conceptual models to assess effectiveness, limited research
has used these models to empirically analyze the possible antecedents of effectiveness,
particularly for humanitarian inter-organizational networks.
I also observed that each of the four models of organizational effectiveness had a
specific focused perspective of effectiveness. For instance, in the Goal Model -
effectiveness is the ability to excel at one or more output goals - the focus is on the output
of the organizations. The System Resource Model - effectiveness is the ability to acquire
scarce and valued resources from the environment -focuses on the input. Concerning the
Internal Process Approach - effectiveness is the ability to excel at internal efficiency,
coordination, motivation, and employee satisfaction- the focus is on the transformation of
input to output. However, all these different focuses had one thing in common. They all
assess effectiveness based mostly on resources internal to organizational. Not much
attention is paid to external resources.
41
Concerning the inter-organizational network effectiveness, I made the following three
observations. Firstly, in almost all of the various inter-organizational network
effectiveness models the focus was at the whole network level of analysis. Using these
models, it would be difficult to conduct organizational level of analysis. Findings from
empirical work (e.g. Stuart et al., 1999; McEvily & Zaheer, 1999; Stuart, 2000;
Rothaermel, 2001) suggest that inter-organizational relationships play a significant role in
shaping the effectiveness of an organization.
Secondly, the vast majority of studies related to inter-organizational network
effectiveness in the nonprofit field are conducted in the public health sector. Moreover, in
most of them, the level of analysis is either community or network. To my knowledge,
only two studies in the specific field of humanitarian assistance investigate humanitarian
inter-organizational network effectiveness. Those papers are Stephenson (2005) and
Stephenson (2006). Stephenson (2005) identifies some of the reasons for the problems
of inter-organizational coordination faced by humanitarian organizations and suggests
ways to address these problems in order to have more effective humanitarian inter-
organizational networks. Stephenson (2006) contributes to the debate in the humanitarian
community about how to make humanitarian assistance more effective. The author argues
that the problem of power and authority in the environment of humanitarian assistance,
best conceived as an inter-organizational social network, must reconceived.
Thirdly and more importantly, using social network theories in the study of
organizational performance, social network researchers have focused on the
organization’s ego network, which encompasses the focal organization (ego), its set of
partners (alters), and their connecting relationships (Wasserman & Faust, 1994). For
example, by counting the number of alliance partners and measuring structural
equivalence, patent counts, and relative scope, Baum et al., (2000) found that the
composition of alliance networks explains differences in organizational performance.
Ahuja (2000) examined the effects of direct ties, indirect ties, and structural holes on
innovation output. Arya & Lin (2007) found that nonprofit organizations that provide a
broad range of services enhance their effectiveness from collaboration in terms of
42
resource gains. Findings of the study also suggest that high-status organizations are able
to derive critical resources from network involvement.
Other social network studies have explored the effect of network structural characteristics
such as centrality, network density, and clique structure on network-level performance /
effectiveness in terms of outcomes (Provan & Milward, 1995; Provan & Sebastian, 1998;
Lerch et al., 2006). For example, the findings from Provan & Sebastian (1998) suggest
that the most effective networks are those that are integrated at clique or sub-network
level. Their findings also suggest that to be most effective, clique integration must be
intensive, involving multiple and overlapping relationships both with and across
organizations that compose the core of a network. Social network researchers have also
shown that strong ties differ from weak ties in terms of their effect on organizational
performance (Rowley et al., 2000).
However, many of these studies have focused on analysis only at the dyadic or the
network level. There is little research in the inter-organizational networks literature
about organizational-level characteristics that can explain whether or not organizations
can enhance their performance from their network positions. Some studies have looked at
the relationships between organizations’ network ties and these organizations’
performance (e.g., Powell et al., 1999; Stuart, 2000; Lee et al., 2001; Almeida et al.,
2003), but none of these studies have explicitly investigated the relationships between
individual organizational characteristics, ego-net characteristics and network structural
characteristics as antecedents of effectiveness. My dissertation intends to contribute to
reduce this gap in the literature.
I investigate inter-organizational network effectiveness in the humanitarian field. I study
a community of interest in humanitarian information management and exchange. Using a
mixed methods research design, I explored the relationships between the structural
properties of network and network effectiveness in humanitarian information exchange.
Network effectiveness was assessed using three different criteria including one subjective
criteria – Perceived network effectiveness and two objectives criteria – number of funded
43
projects and number of funding partners. My investigation is conducted at two different
levels of analysis, network and organizational levels.
Provan & Sebastian (1998) found that network-level effectiveness can be explained by
intensive integration through network cliques. Building upon this work, I explored
networks of international heterogeneous and geographically dispersed organizations
engaged in humanitarian assistance and disaster relief. I sought to understand the extent
to which Provan & Sebastian’s Model would explain effectiveness in this context. In
addition, unlike Provan & Sebastian who assessed network effectiveness using one
subjective criteria (Patient outcomes), in my study, I explored three effectiveness criteria
including one subjective and two objectives.
At the organizational level, I combined two theoretical lenses including Social Network
and Resource Based View. Network structural characteristics (density, centrality, clique
and clique overlap) have been found to have implications on performance/effectiveness
Hypothesis HO#2 Bridging of structural holes increases organization effectiveness Organization
Hypothesis HN#1 Network effectiveness increases with network density Network
Hypothesis HO#3 Organization effectiveness increases with the density of the network it which it belongs
Organization
Hypothesis HN#2 Network effectiveness increases with the number of cliques in the network
Network
Hypothesis HN#3 Network effectiveness increases with the number of organizations in cliques
Network
Hypothesis HO#4 Organization effectiveness increases with the number of distinct cliques to which it belongs
Organization
Hypothesis HN#4 Network effectiveness increases with the level of overlapping clique in the network
Network
Hypothesis HN#5 Network effectiveness increases with the level of multiplexity in the network
Network
Hypothesis HN#6 Network effectiveness increases with the level of identical cliques in the network
Network
Hypothesis HO#5 The size of an organization is positively associated with its effectiveness
Organization
Hypothesis HO#6 The range of service provided by an organization is positively associated with its effectiveness
Organization
Hypothesis HO#7 The greater the variety of communication media available in an organization, the higher its effectiveness
Organization
Hypothesis HO#8 The greater the variety of collaboration social software available in an organization, the higher its effectiveness
Organization
Hypothesis HO#9 The greater the variety of community social software available in an organization, the higher its effectiveness
Organization
64
Hypothesis HO#10 Organizations that possess a wide variety of communication media will benefit more from high network degree centrality to enhance their effectiveness than those that do not.
Organization
Hypothesis HO#11 Organizations that possess a wide variety of communication media will benefit more from high network density to enhance their effectiveness than those that do not.
Organization
Table 5: Summary of Hypotheses
The two figures (Figure 4 and Figure 5) below, depict my research models.
NETWORK
CHARACTERISTICS
Perceived effectiveness
Number of funded projects
Number of funding partners
EFFECTIVENESS
Network Attributes
Density (+)
Clique (+)
Org in Clique (+)
Clique overlaps (+)
Multiplexity (+)
Identical clique (+)
Figure 4: Research Model for Network Level of Analysis
INTERNAL
CHARACTERISTICS
Number of funded projects
Number of funding partners
EXTERNAL
CHARACTERISTICSEFFECTIVENESS
Organization Attributes
Size (+)
Service provided (+)
Communication media (+)
Collaboration social media (+)
Community social media (+)
Ego-net Attributes
Degree centrality (+)
Structural hole (+)
Number of clique (+)
Network Attributes
Density (+)
Figure 5: Research Model for Organizatioanl Level of Analysis
65
5 METHODOLOGY
5.1 Introduction
This chapter presents the research design and method that I used in my study.
5.2 Research Design
I use a mixed methods research design (Tashakkori & Teddlie, 2003) to explore
multidimensional inter-organizational networks of collaborative relationships among
humanitarian organizations that are members of the Global Symposium. Using mixed
method allowed me to leverage on both the quantitative and the qualitative research
techniques.
Quantitative researchers use postpositivist propositions for developing knowledge, such
as hypotheses and questions, reduction to specific variables, and the test of theories.
Numerical data and statistics are their main instrument (Charles & Mertler, 2002). They
isolate variables and causally relates them to determine the magnitude and frequency of
relationships. In addition, they determine themselves which variables to investigate and
chooses instruments, which will yield highly reliable and valid scores (Ivankova et al.,
2006). Contrary to quantitative research, qualitative research is interpretive or
constructive. It is “an inquiry process of understanding” where the investigator develops
a “complex, holistic picture, analyzes words, reports detailed views of informants, and
conducts the study in a natural setting” (Creswell, 1998, p. 15). In qualitative research,
the investigator makes knowledge claims based on the constructivist (Guba & Lincoln,
1989) or advocacy/participatory (Mertens, 2003) perspectives. In this approach, data is
collected from those immersed in everyday life of the setting in which the study is framed
(Ivankova, et al., 2006). Data analysis is based on the values that these participants
perceive for their world. According to Miller (2000), qualitative research produces an
understanding of the problem based on multiple contextual factors.
66
Combining both quantitative and qualitative in a mixed methods approach, the
investigators develop and build the knowledge on pragmatic grounds (Creswell, 2003;
Maxcy, 2003). They choose approaches, as well as variables and units of analysis, which
are most appropriate for finding an answer to their research question (Tashakkori &
Teddlie, 1998). A major tenet of pragmatism is that quantitative and qualitative methods
are compatible. Thus, both numerical and text data, collected sequentially or
concurrently, can help better understand the research problem (Ivankova, et al., 2006).
Mixed methods involve collecting, analyzing and combining both quantitative and
qualitative data within a single study. According to Creswell (2002), mixed methods help
to understand a research problem more completely. Another argument for mixed methods
is that neither quantitative nor qualitative methods are sufficient by themselves to capture
the details of the situation, such as a complex issue information management and
exchange among organizations engaged in disaster relief. When the quantitative and
qualitative methods are combined, they complement each other and allow for more
complete analysis (Green et al., 1989, Tashakkori & Teddlie, 1998).
While designing my research, I considered the following three important issues: priority,
implementation, and integration (Creswell et al., 2003). Priority refers to which method,
either quantitative or qualitative, is given more emphasis in the study. Implementation
refers to whether the quantitative and qualitative data collection and analysis comes in
sequence or in chronological stages, one following another, or in parallel or concurrently.
Integration refers to the phase in the research process where the mixing or connecting of
quantitative and qualitative data occurs. Quantitative method had the priority in my study.
I first collected quantitative data through a series of three surveys and then conducted
interviews to collect qualitative data. I integrated the qualitative and quantitative data in
the analysis phase. The research took place over a two-year period, encompassing
numerous interactions with various network members and feedback with the research
participants. Data analysis involved multiple levels of social network analysis, statistical
analysis and a combination of inductive and deductive content analysis techniques.
67
5.3 Research Participants
My research explored inter-organizational networks in the Global Symposium, a
community of interest in humanitarian information management and exchange
spearheaded by the United Nations Office for the Coordination of Humanitarian Affairs
(UNOCHA). The research participants were representative of organizations member of
the Global Symposium who attended to at least one of the five Global Symposium
meetings. UNOCHA provided me with the list of all the attendees of the various Global
Symposium meetings. They were almost all high ranked senior staff (e.g., CEO, CIO, IT
Director) in their organizations.
As stated earlier, the Global Symposium is a community of interest in humanitarian
information management spearheaded by UNOCHA. A community of interest as defined
by Arias & Fischer (2000), is made up of individuals from different backgrounds that
come together to solve a particular problem of common concern. The Global Symposium
began its activities in 2002 as a meeting of humanitarian information management
professionals. This community of interest is made up of about 300 information
technology (IT) and information management (IM) professionals from roughly 120
international and national organizations in the field of humanitarian assistance. The goals
of the Global Symposium include (i) to foster collaboration among members on
humanitarian information management related projects, (ii) to disseminate best practices
of information exchange, (iii) to sensitize its members on the critical aspect of
humanitarian information management preparedness and (iv) to facilitate headquarter-
field partnerships and to advocate for more funding from donors for humanitarian
information management related projects.
For both theoretical and empirical reasons, I subdivided the Global Symposium
community into different sub-networks. I identified three sub-networks including the
non-governmental organizations (NGO) subnet, the United Nations agencies (UNA)
subnet, and the governmental organization (GO) subnet. Separating members of a
network into subnets and analyzing how they overlap can be an important means for
68
understanding how the network as a whole is likely to facilitate or constrain certain
actions of these members (Giddens 1984; Sydow & Windeler 1998). Although the
members of the Global Symposium are all interested in humanitarian relief and especially
humanitarian information management and exchange, they theoretical differ on a good
number of characteristics including their missions / goals, their sources of funding and
their mode of governance. The three subnets were also identified based on UNOCHA
categories of organization in the humanitarian relief field. I briefly describe below, the
general characteristics of the organizations member of each of the three sub-networks.
a) NGO
NGOs are “private organizations that pursue activities to relieve suffering, promote the
interests of the poor, protect the environment, provide basic social services, or undertake
community development” (World Bank, 2000). One of the long-established activities of
these organizations is to provide humanitarian assistance. NGOs engage in two broad
types of activities including relief activities and development activities. Relief activities
consist of assisting to victims of natural or manmade disasters. Relief NGOS frequently
specialize in one or more of the five activities that are commonly understood to compose
the relief discipline: food distribution, shelter, water, sanitation and medical care.
Development activities are longer-term assistance, focusing on community self-
sufficiency and sustainability. These activities include establishing permanent and
reliable transportation, healthcare, housing, and food. NGOs’ resources come primarily
from private sources and major donor government contributions. NGOS are governed by
boards of directors that tend to reflect the particular culture, history and mandates of the
organizations concerned. In my study, the NGO subnet was made up of 72 organizations.
b) UNA
The United Nations (UN) plays a vital role in humanitarian assistance. For this endeavor,
the institution operates several major organizations among which five are such visible
players in most complex humanitarian emergencies that describing their functions and
mandates will describe most if not all of the operational work of the entire UN system in
relief operations. They are the World Food Program, the Office of the United Nations
69
High Commissioner for Refugees, the United Nations Children's Fund (UNICEF) and the
United Nations Development Program (UNDP) and the UN Office for the Coordination
of Humanitarian Affairs (UNOCHA). The UN World Food Program functions as the
food aid agency of the UN system, providing a central coordinating role in developing
crop production estimates, food aid requirements and logistics planning for major relief
operations. UNICEF'S special mandate is to focus on the relief and development needs of
women and children, which has made it the focal point among the UN agencies for
emergency medical interventions, mass inoculation campaigns for children, water and
sanitation programs and therapeutic feeding programs for severely malnourished children
in emergencies. UNDP technically has the mandate to manage UN emergency operations
in the field while UNOCHA is charged with the coordination and synchronization of
United Nations humanitarian efforts. Each of these UN organizations depends for funding
on the goodwill of member governments and/or the broader populations of those nations.
In my study, the UNA subnet comprised 25 agencies.
c) GO
Governmental organizations are owned by governments. Governmental organizations
work to achieve the goals set by the government. These goals are often set for political
reasons. The managers of these organizations are appointed by the government. The
government also provides the necessary resources to these organizations. In my study, the
GO subnet was made up of 53 organizations.
Subdividing the Global Symposium community in sub-networks also had an empirical
justification. Using social network block model I found that these sub-networks
presented diversified patterns of inter-organizational relationships. The level of inter-
organizational relationships (measured as network density) ranged from 0.076 to 0.193
for project collaboration dimension and from 0.025 to 0.074 for the advice dimension.
The United Nations agencies sub-network displayed was the most strongly
interconnected on both dimensions followed respectively by the non-governmental
organizations subnet and lastly the governmental organization subnet. On the project
collaboration dimension for example, approximately twenty percent (19.30%) of all the
possible project collaboration relationships between the organizations in the United
70
Nations agencies sub-network were actually found to exit. In contrast, only about eight
percent (7.6%) of all possible linkages between organizations in the governmental
organizations sub-network were found to exist. On the advice dimension, these
percentages were respectively (7.42%) for the United Nations agencies sub-network,
(2.92%) for the non-governmental organizations subnet and (2.52%) for the
governmental organizations subnet.
GO 0.071
0.193
UNA
NGO
0.0780.076
0.1
360.
128
Figure 6: Global Symposium Project Collaboration Sub-Networks
GO 0.0218
0.0742
UNA
NGO
0.02920.0252
0.0
3110.
0336
Figure 7: Global Symposium Advice Sub-Networks
71
An examination of the level of interaction cross the three subnets also shown a significant
discrepancy for both project collaboration and advice dimensions of relationships. Figure
6 and Figure 7 above depict these differences.
Understanding networks in the field of humanitarian relief can also be enhanced by
considering the different type of relationships that exist among organizations. In my
study, I investigate two types of inter-organizational relationships in the community. I
study the relationship on inter-organizational collaboration on humanitarian project
among members of the Global Symposium and the advice relationship.
5.4 Data Collection Instruments
I used multiple instruments to collect data including surveys, interviews and online database
search. Collecting data by different methods from different sources produces a wider scope of
coverage and may result in a fuller picture of the phenomena under study than would be achieved
otherwise (Bonoma, 1985). The data captures both the whole network and the individual
organizational perspectives on inter-organizational humanitarian information exchange
relationships among members of the Global Symposium community.
5.4.1 Survey
A survey instrument that also contains network-related questions was my main data
collection instrument (See Appendix B). I conducted a series of three surveys (October
2007, May 2008 and July 2009) and used two different types of survey instrument. The
first survey was paper based and the two subsequent were web based. The electronic
form of the survey instrument was designed based on the quality criteria identified in
(Wright, 2005). It had a simple layout using a straightforward navigation strategy. I kept
graphics and color to a minimum in other to minimize the downloading time. I used the
Survey Monkey software to develop the survey. I chose this software mainly because I had
earlier used it in several research projects with my adviser. I had developed extensive
experience and a high skill set in the use of that software.
72
On the survey instrument, an introduction page including an informed consent preceded the
survey questions. The purpose of the survey introduction page was multiple. First, I wanted
to create a trusting relationship with survey participants by repeating the survey purpose
already explained in the invitation letter. Second, the introduction page was intended to offer
a non-financial incentive – a report of the results, and to guarantee confidentiality and
privacy of research participants. Third, I wanted to provide a third party guarantee of the
survey’s authenticity and credibility by stating the University’s Institutional Review Board
(IRB) approval. The informed consent asked participants to give their permission for the
survey.
Following the introduction page, came the survey questions. Though they were not
completely identical cross surveys, there were significant overlaps especially with
regards to inter-organizational relationship questions. In general, the questions included
the following four categories: (i) respondent’s organization information; (ii) issues on
humanitarian information management and exchange in the Global Symposium
community; (iii) Global Symposium community collaborative benefits and effectiveness,
and (iv) the community inter-organizational networks. For questions concerning the
inter-organizational network, survey participants were provided with the list of members
of the Global Symposium community and were asked to identify (i) those with which
they had collaborated on humanitarian projects and (ii) those with which they had advice
relationships.
Most of the questions were structured using a five point Likert scale (Likert, 1932). For
every question or statement, I provided respondents with five choices representing the
degree of agreement on the question. For the network question, the survey instrument
included the list of all the organizations member of the Global Symposium. With regards
to conducting the Surveys, the paper based survey was administered during the 2007 Global
Symposium meeting. Survey questionnaires were handed to participants. They had to
compete it and turned back by the end of the conference. With regards to the electronic
surveys, the survey invitation was sent through direct email to each participant. The
invitation was a shortened version of the survey introduction page. After reading the
invitation, online community members ignored the post or self-selected to take the survey by
73
clicking on the survey URL. Two follow-up “reminder” invitations were sent approximately
one week apart to the participants. All inquiry email, whether sent as a reply to the invitation
was responded to as soon as I got the mail.
Both the electronic and paper forms of survey instrument have advantages and
drawbacks. By using a combination of the two forms, my intention was to leverage on
their advantages to limit the impact of their disadvantages on my research. As compared
to paper based surveys, electronic surveys present many advantages. They provide a way
to conduct studies when it is impractical or financially unfeasible to access certain
2009). In order to increase the validity of data and the number of cases, I collected data
for a period of ten years (1999-2009). I realized that no data was available in this
database for organizations of the for-profit network. This network would be ranked in the
study using only the perceived network effectiveness criteria.
79
5.6 Data Analysis Techniques
I used a combination of social network and statistical analyses techniques in my study.
5.6.1 Social Network Techniques
Reasons for Using Social Network Techniques
Social network analysis is appropriate for my research for the following four reasons.
First, Social network analysis is a powerful and relatively new research tool which has
developed popularity in recent years (Kilduff & Tsai, 2006; Quatman, 2008). The
network perspective offers some unique advantages to the research process. According to
Quatman (2008), network approaches allow for example for: (1) a concrete vitality for
several difficult-to-define constructs; (2) simultaneous analysis of multiple levels of
relational data thus providing some fluidity between micro-, meso-, and macro- linkages;
and (3) a unique integration of quantitative, qualitative, and graphical data producing an
intuitive, thorough, and rich analysis of phenomena .
Second, social network studies cover a wide range of research contexts. The utility and
applicability of social network analysis is very broad and has been embraced by
researchers in a number of fields (Kilduff & Tsai, 2006; Quatman, 2008). Several papers
provide quite extensive reviews and a variety of contextual examples of the uses of social
network analysis for research purposes such as: (i) Brass et al., (2004) and Parkhe et al.,
(2006) for management and organizational behavior topics; and (ii) Provan & Milward
(1995), Lemieux-Charles et Al. (2005), and Arya & Lin (2007) for health service
delivery.
Third, in a network approach, actors can be characterized by any type of entity embedded
within a larger system of entities (Granovetter, 1985; Wasserman & Faust, 1994; Kilduff
& Tsai, 2006). In the social sciences, the entities of interest are often individual people or
groups of people acting as a unit. In a network approach, researchers also have the
freedom to operationalize the relationships of interest between the actors. For example, a
80
researcher might explore friendship links between employees in an organization or
resource exchange links between organizations in a market.
Fourth, social network approaches also allow researchers to investigate several different
attributes of relational ties between actors (Wasserman & Faust, 1994; Kilduff & Tsai,
2006). Instead of simply considering whether or not a tie is present, a researcher can
examine additional implications from network configurations. For example, a network
investigation can incorporate such things as the intensity (often measured by strength or
frequency of interaction) and direction of ties (often used to represent the direction of
effect). In addition, a single set of network members can also be used to examine the
multiplexity of ties between members in the network. As discussed earlier, the
multiplexity of a tie refers to the extent to which two network members are linked
together by more than one relationship. Moreover, the attributes of the ties (for example
directionality, intensity, and multiplexity) do not have to be considered mutually
exclusively. In a nutshell, a network can be examined from any and all of these
perspectives simultaneously.
Social Network Analysis
In network analysis I first considered measurement of the basic network structural
properties. They included size, density, connectedness and centralization. This allowed
for consideration of whole network behavior as well as an understanding of individuals in
the networks. At the network level, the size of the network is an important consideration
for the potential reach or number of logically possible relationships for the number of
actors in the network. Network size at the individual level considers a number of factors
in relation to the number of adjacent actors.
As I mentioned earlier, the density of a network indicates to the number of recorded links
between network members in proportion to the number of all possible links within a
network. Density calculations illustrate the degree to which a network realizes its
potential, assuming that the optimum is a fully saturated network where everyone
81
contacts everyone else. The organizational context is important in assessing the desired
density. With the humanitarian information management and exchange, high density in
project collaboration may indicate a tightly group who consult frequently to resolve
issues. On the other hand, a high density in advice may indicate many organizations
struggling with how to handle problems.
At the whole network level, another consideration is centralization. Centralization
measures the extent to which a network or group is organized around its central point
(Freeman, 1979). The arrangement of actors in the network affects how quickly and
easily information can be distributed among all the actors (Freeman, 1979; Wasserman &
Faust 1994; Haythornthwaite, 1996). Centralization is a measure of integration or
cohesion of the group. A centralized network may reflect an uneven distribution of
knowledge such that knowledge is concentrated in the focal points of the network. In
addition to matrix calculation, the sociograms illustrating these calculations are
particularly useful for viewing the different networks.
To consider network structural influence on individual actors and identify the variety of
network roles within the various networks, a number of adjacency calculations on the
direct connections from one member to another demonstrated the degree to which an
actor sends or receives information. For graph theorists, there exits four types of network
nodes including (i) isolate, (ii) transmitter, (iii) receiver and (iv) carrier. An isolate
neither sends nor receives information; a transmitter sends information; a receiver
receives information; and a carrier both sends and receives information (Wasserman &
Faust, 1994, p. 128). The outdegree calculated as the sum of connections an actor has to
others, is often used to measure an actor’s influence. In-degree links refer to the number
of actors sending information to the actor in question. Network members that receive a
lot of information may be more powerful, suffer from information overload, or hold a
position of prestige (Hanneman & Riddle, 2005, p. 43). Network members neither
sending nor receiving information either withhold information or fail to contribute to a
network. In this study, I did not take into consideration the isolates. I excluded all the
82
isolated network members from my analysis. I considered the rest (those that had at least
one connection) as carriers.
The presence or absence of subgroups within the social network structures was a prime
consideration for the analysis of the inter-organizational collaboration activities among
members of the Global Symposium community. In some networks, a sub-group forms
when only two actors have a tie to each other. In others, groups of actors demonstrate
more ties to each other than to the other members of a network. This type of sub-group is
termed a clique, defined by Hanneman & Riddle (2005) as “some number of actors (more
than two, usually three is used) who have all possible ties present among themselves” (p.
80). This definition may restrict the concept’s application in many social networks. As a
result, an extension is the concept of the n-clique, where n is the maximum path length at
which members of the clique are considered connected. This extension “is much closer to
people’s everyday understanding of the word clique” (Scott, 1991, p. 115).
The impact of subsets within a social network may depend on the degree to which they
are connected. The examination of bridges, or critical ties between two actors, extends
from the consideration of small groups in a larger network. An actor who provides the
connection or critical tie to another group of actors performs “the liaison role of
connecting two otherwise disconnected cliques” (Kilduff & Tsai, 2003, p. 28). This role
becomes important in considerations of what happens if the connection drops, and the
value of maintaining or continuing to invest in the relationship.
Finally, the construction of sociograms to demonstrate visually some of the properties
across the networks assisted the analysis by highlighting numerous features for
consideration.
Unit of Analysis in Social Network
In network analysis there are four units of analysis that are frequently used. They include
dyads, triads, egocentric networks and whole networks (Wasserman & Faust, 1994). In
83
my study, I am concerned with two units of analysis, the egocentric network and the
whole network. The egocentric network level has been used primarily to study network of
individuals. For example, Laumann (1973) analyzed friendship networks among urban
men; Granovetter (1974) studied how information about job is transmitted; and Minor
(1983) examined personal relationships among former heroin addicts. One problem with
the egocentric network level analysis when applied to organizational networks is that it
carries a connotation of individual relations, with a psychoanalytic orientation. For this
reason, I will refer to the unit of analysis as organizational-centered networks, implying a
focus on organizations as opposed to individuals.
This level of analysis consists of each organization along with all other organizations
with which it has a relationship. Generally, this unit of analysis is used to examine
attributes and characteristics of the relationships which exist between each organization
and all other organizations in its organizational-centered network. Each organizational-
centered network can be described by the number, magnitude, type and other
characteristics of it linkages with others in the network (Knoke & Kuklinski, 1982;
Streeter, 1989).
The network-level of analysis has also been intensively used in inter-organizational
network research. At the network-level of analysis, researchers look at the composition of
the networks (e.g., network size, network heterogeneity, mean frequency of contact) and
the structure of these networks (e.g., density of links among alters). According to
Wellman & Frank (2000), such analyses seek to understand how the properties of
networks affect what happens in them and to them. Provan et al., (2007) provide an
extensive and comprehensive review of inter-organizational network research conducted
at the network level of analysis. According to this review, research at network level has
mainly been conceptual, anecdotal, or based on single, descriptive case studies performed
at one point in time. Also most of the research reviewed by Provan et al., (2007) was
done in the health sector.
84
5.6.2 Content Analysis
Content analysis “is a well-established set of techniques for making inferences from text
about source, content, or receivers of information” (Schamber, 2000, p. 735). Organizing
and properly coding data is critical to content analysis. Coding is the process of combing
the data for themes, ideas and categories and then marking similar passages of text with a
code label. Coding the data makes it easier to search the data, to make comparisons and
to identify any patterns that require further investigation. The process of coding is an
iterative and cyclical process of constant discovery. Seidel (1998) developed a model
(figure 9) to explain the basic process of qualitative data analysis. The model consists of
three parts: Noticing, Collecting, and Thinking about interesting things. These parts are
interlinked and cyclical. For example while thinking about things you notice further
things and collect them. Noticing interesting things in the data and assigning ‘codes’ to
them, based on topic or theme, potentially breaks the data into fragments. Codes which
have been applied to the data then act as sorting and collection devices.
Figure 11: Qualitative data analysis coding process (Seidel, 1998)
In my research, I coded the transcribed interviews both deductively and inductively
(Epstein & Martin, 2004). In the deductive coding approach, the codes are developed
before data collection. I developed my set of codes based on my research questions.
Usually the deductive approach is used when researchers may be seeking to test existing
theories or, as it is the case in my research, expand on them.
85
During the coding process I also let some inductive codes emerge from the data. The
inductive approach reflects frequently reported patterns used in qualitative data analysis.
Inductive coding begins with close readings of text and consideration of the multiple
meanings that are inherent in the text. The researcher then identifies text segments that
contain meaning units, and creates a label for a new category into which the text segment
is assigned. Additional text segments are added to the category where they are relevant.
At some stage the researcher may develop an initial description of meaning of category
and by the writing of a memo about the category (e.g., associations, links and
implications). The category may also be linked to other categories in various
relationships such as: a network, a hierarchy of categories or a causal sequence. Coding
inductively, researchers are likely to create new codes, they therefore need to go back and
check the units of data they coded previous to creating this code. This is to check if there
is any more data that should be coded at the newly created node. The diagram below
shows how I applied new codes to previously coded data.
While organizing and coding qualitative data, it is important to carefully read and
recognize data prior to the coding process. Mason (1996) suggests there are three main
epistemological reading schemes; literal, interpretive and reflexive. In literal reading, the
researcher is interested in the literal form of the data, whether it is the content, structure,
style, and layout. Researchers do not make interpretation of the data. They look at the
data as it is presented. Most qualitative researchers argue that a literal version of reading
data could not yield desirable results for this kind of data organization as it might direct
our attention from the whole to details and style. In interpretive reading, researchers look
beyond literal form of the data, and try to get to the underlined or implied meanings.
Interpretive reading brings the researcher's own opinions into play. In reflexive readings,
the researcher will look at him/herself as part of the data they generated, and will seek to
explore his or her role and perspective in data generation and data interpretation. The
interpretive and reflexive reading puts emphasis on construction and documentation of
the meaning of data rather than the literal structure of it. In my research, I used the
interpretative reading scheme.
86
Mason (1996) also highlights two different methods for coding qualitative data including
(i) cross-sectional or categorical coding and non-cross-sectional or contextual coding.
Cross-sectional coding consists to consistently code the whole data set according to some
sets of common principles in a very systematic way. The simplest form of cross-sectional
coding is serial coding, which is to insert subheadings at relevant points in the text data.
Some of the advantages of this method are that (i) it makes sorting and retrieval easier (ii)
it gives a holistic view or understanding of the data set (iii) it can index the locations of
interpretive, conceptual and theoretic themes within the data, and (iv) it can provide
analytic handles of different parts of the data set for cross-comparison. According to
Mason (1996), there are three main limitations to the cross-sectional coding approach. Its
categories might be too broad to be very useful. Second, a section of text is likely to be
related to more than one concept, thus serial coding might be inappropriate. Third, serial
coding is unlikely to work well if the data is not of a uniform layout. Cross-sectional
indexing can be done very easily, if the data is mostly textual information. For instance it
might not cater to some relevant comparisons across categories. In addition, it tends to be
less useful for interview transcripts, particularly when the interview is either semi-
structured or unstructured.
Non-cross-sectional coding, on the other hand, relates to a totally different idea. The
researcher using this method reads over the data set and constructs a different lens for
each document by examining the documents individually. The principal advantage
behind this approach is that it involves the evaluation of each document in its entirety
inclusive of the context of the data generation. Since this approach helps the researcher
build a case out of each examination, it is also referred to as the case study method of
organizing data. According to Mason (1996), one of the driving reasons to perform non-
cross-sectional data organization is related to the fact that the researcher can identify and
analyze deeply the ideas inherent in each document and how these ideas are interwoven.
In my research, I used the cross-sectional coding method.
87
5.6.3 Statistical Analysis
In my study, I used correlation and multiple regression for statistical analysis. Multiple
regression is a statistical technique that allows to predict the value of a dependent
variable on the basis of the values of several independent variables. Multiple regressions
and multiple correlations deals with the relationship of one variable compared with a
number of other variables. In this research, the multiple regression and multiple
correlation was used to compare different predictors of organizational effectiveness in
humanitarian information management networks.
5.7 Methodological Issues
The issues of external validity involve the degree to which the results of the research
study apply to other communities. These issues can occur at different levels including the
level of theories and methods used as well as the level of the findings. At the level of
theories at methods, threats to external validity occur if inappropriate concepts,
instruments, or methods are applied to a research study. At the level of findings, external
validity concerns the extent to which the results of the study hold true for similar
populations. My literature review, the concepts and methods that I used were appropriate
and indicate no threat by the external validity concerns to my study.
External reliability relates to the ability of other independent researchers to “discover the
same phenomena or generate the same constructs as an original researcher if they did
studies in the same or similar settings” (Schensul, Schensul, & LeCompte, p. 275).
Clearly, identifying the research steps and detailing the analysis process and
interpretation of results elevates external reliability. I believe that it is not possible to
duplicate the research setting and results of my study. However, other researchers can
duplicate the research process.
88
5.7.1 Social Network Analysis Issues
Validity
According to Wasserman & Faust (1994), construct validity is the extent to which the
questions really assess what they purport to measure. Wasserman & Faust state that “very
little research on the construct validity of measures of network concepts has been
conducted” (p. 58). However, they do acknowledge that “the construct validity of social
network measures can be studied by examining how these measure behave in a range of
theoretical propositions” (p. 58). The construction of questions (both survey and
interview) relied on validated questions I had used on numerous occasions (Cross &
Parker, 2004) in the humanitarian field, to “uncover important network relationships” (p.
147) of information exchange and collaboration.
Measurement Errors
A discrepancy between what is measured and the “true” value of a concept constitutes
measurement error (Wasserman & Faust, 2004, p. 59). In my research, the response rate
of the three surveys that I conducted had an impact on the measurement of network
structural properties, reciprocity in particular. Higher response rate would have been
better. One way to counter threats to internal validity and measurement error is to use
member checking. Presenting the data results to the members of the Global Symposium
community improved the degree to which the responses obtained reflected their
perception of network interactions. The data acquired through interviews provided
additional confirmation.
Reliability
Assessing reliability of measures of sociometric data relates to the success of achieving
the same estimates from repeated measurements. In the words of Wasserman & Faust
(1994) “this assumption is likely to be inappropriate for social network properties, since
social phenomena cannot be assumed to remain in stasis over any but the shortest spans
of time” (p. 58). In this study, sociometric data collection relied on three datasets
collected at three different points in time. Admittedly, this introduces an element of error.
89
5.7.2 Content Analysis Issues
Validity
Validity concerns the extent to which instruments are accurate and dependable, and the
degrees to which the research results make sense to the people studied and generalize to
other similar populations (Schensul & LeCompte). In my research, content validity
reflects the high correlation between what users described in their own words and
examples of those expressions in the research literature. The frequency and redundancy
of these descriptions supports the degree to which the research results apply to the people
studied.
Reliability
The issue of reliability concerns whether another researcher using the same methods can
replicate the research results. Internal reliability for content analysis relates to the match
between the constructs identified and the data sets that generated the constructs. One way
to increase the reliability of these matches is for at least one other researcher to review
the data sets and the constructs to see if there is agreement on the matches generated. To
accomplish this, one colleague with content analysis experience reviewed a sample of
three interview transcripts. Krippendorff’s alpha (Krippendorff, 2004) was the basis of
the test for inter-coder reliability. In this case, the observed disagreement divided by the
expected disagreement produced a calculation of desired agreement, with 84% agreement
meeting the minimum goal of 80% acknowledged as appropriate for exploratory studies
(p. 242). Confirming the coding reliability increased the confidence in applying the
results.
5.8 Summary
In this study, I used a mixed methods research design to investigate the relationships
between structural properties of inter-organizational networks and network effectiveness
among organizations members of a community of interest in humanitarian information
management and exchange. Data collection included surveys, individual semi-structured
interviews, and database search. The process of triangulation encompassed in this
research design provided the means to corroborate findings and extend the results beyond
this research setting. The next chapter presents the research results.
90
6 ANALYSIS
6.1 Introduction
In this chapter I present the findings of my investigations. The chapter contains four main
sections. In the first section, I analyze the qualitative data collected through interviews. In
section two, I present my three criteria for assessing effectiveness. As discussed in the
research model, they are my dependent variables. The third section presents my findings
related to the network level of analysis. In this third section, I present my analysis of the
relationships between the structural properties (density, clique, clique overlap and
multiplexity) and network effectiveness. Finally in the fourth and last section, I present
my findings related to organizational level effectiveness.
6.2 Qualitative Data Analysis
As discussed in the method section, in my research, I collected data through multiple
sources. In this section, I analyze the qualitative data gathered through interviews. I
conducted a total of nineteen semi-structured interviews among members of the Global
Symposium. I coded the data using deductive and inductive methods. The five deductive
code categories that I used were guided by my research questions and the literature. They
included (i) network benefit, (ii) network effectiveness, (iii) collaboration factors, (iv)
barriers to collaboration and (v) measures of network effectiveness. Three inductive code
categories emerged from the data including (i) from advice to collaboration, (ii) network
scope and (iii) network audience.
6.2.1 Deductive codes
6.2.1.1 Network benefit
One of the objectives of the interview was to assess the benefit of the Global Symposium
to individual organizations as well as to the community as a whole. It is important to
91
recall that membership to the Global Symposium network is voluntary. If members do
not benefit from being part of the network, they will more likely cease to participate in
the activities of the network, and if the feeling is widespread, the network will cease to
function. Recognizing the benefits that members receive from network membership is
therefore a crucial tool to assess how well the network is functioning. Based on the data
that I collected, numerous benefits were perceived to be associated with the Global
Symposium inter-organizational network.
Benefit to individual organization
I asked to the interview participants, what benefits their organization had gained from
being a member of the Global Symposium. About eighty five percent (84.21%) of the
interviewees answered this question. Some of the most commonly cited benefits of the
network to individual organization members included: increased access to humanitarian
information; expertise and financial resources; solidarity and support; and increased
networking. Another important perceived benefit reported was increased credibility of the
Global Symposium members. One subject expressed this benefit in the following term:
Subject#18: The greatest as I said, was meeting with various people from all over, networking and then of course it was informal relations but strangely enough you could think that this networking will be closer with the working group which I was part, but it was not. At the time I made those networks it was at the closing session in fact.
Another subject said:
Subject#1: There are people and entities we met that we are now discussing with and sharing information with sharing ideas with and you know just keep in touch at an informal level. I think that is very good for us.
Benefit to the whole community
The second aspect of the question on benefits was related to the contributions of the
Global Symposium to the community. I asked the interview participants about how the
Global Symposium benefited the humanitarian community as a whole. Approximately
fifty eight percent (57.89%) of the interviewees answered this question. Among those that
responded, the vast majority expressed highly positive opinions on the Global
Symposium’s contributions to the community. Especially, they believed the Global
92
Symposium benefited the community in two major aspects including (i) promoting the
use of humanitarian information management principles and dissemination of best
practices; (ii) fostering collaboration on humanitarian information management projects. I
provide below, some illustrative quotes from interview data.
Subject#8: I think the development of the principles was useful. I think that just the working groups to address certain issues was useful although I would be the first to admit there has been not any real organized follow up. But I think documenting the information management principles and the actual document itself that came out of the symposium I think was useful in terms of the issues of humanitarian information management. Subject#17: It was a very very important networking opportunity, and you know in some respect it was very cutting edge. The only thing I find very disappointing was the lack of invitation to some key players and so, the one in particular. Subject#5: I think one of the great benefits is actually making like minded people and you realize that we are all confronting the same problems, so I think that firstly is one big positive aspect. I think the second, is that there was a lot of networking going on that was actually quite crucial.
However, for some interviewees, the benefit of the Global Symposium was limited to
certain linguistic regions. They said the English speaking members of the community
were those that benefited the most.
The impact of the Global Symposium for this region [Spanish] is very very low, I would say or almost invisible. I am not sure if we would benefit from the Global Symposium in that sense. No unfortunately not. Subject#16
Across the interviews, the discussion on the benefits of the Global Symposium to
individual organizations on the one hand and to the whole community on the other hand
was done with almost the same intensity (Figure 12). The former represented
approximately fifty eight percent (57.81%) of occurrences of benefits discussed in all the
interviews combined while the later represented about forty two eight percent (42.19%).
93
Figure 12: Network Benefit Code’s Coverage
Figure 13: Aggregated Benefit Cross Network
Aggregated data per network (Figure 13) shows that cross networks, the Global
Symposium was perceived to benefit more to individual organizations than to the
community as a whole. In the network of NGOs the proportion is almost seventy percent
94
(70.00%) and thirty percent (30.00%). In the two other networks, the proportion is
approximately sixty percent (60.00%) and forty percent (40.00%).
6.2.1.2 Network effectiveness
Another objective of the interview was to assess the perception of the participants about
the effectiveness of the Global Symposium as a network of organizations engaged in
humanitarian information management and exchange. Though I often used open
questions during the interviews, in this case, I recalled some of the main objectives of the
Global Symposium to the interviewees. I then asked them to comment on whether or not
the Global Symposium was effective in the pursuit of these objectives. I coded the data
collected using four categories. They included (i) resource availability, (ii) internal
processing characteristics, (iii) goal achievement, and (iv) multiple constituencies’
satisfaction. These categories came from the literature on organizational and network
effectiveness (Parson 1964; Yuchtman and Seashore, 1967; Price, 1971; Cameron and
In this code category, effectiveness was defined in terms of the ability of the organization
/ network to acquire resources necessary to it survival. The greater the ability of the
organization / network to acquire needed resources, the greater its effectiveness. Only
about thirty two percent (31.58%) of the interviewees discussed the effectiveness of the
Global Symposium in term of resource availability. The majority of these subjects had a
negative opinion about the ability of the Global Symposium to make resources more
available to its members. For example, Subject#2 talked about the “unrealistic”
objectives of the Global Symposium.
Subject#2: How could you possibly do that? I mean these objectives are ummm let us be realistic. How are we going to help organizations get more resources? Subject#5: I think a lot of the organization including ours, have simply not got the time or context to seek the necessary funding and resources.
95
Internal processing characteristics
In this category, effectiveness was conceptualized as the absence of internal strain and a
smooth internal functioning of organizations / networks. Approximately fifty eight
percent (57.89%) of the interviewees discussed the effectiveness of the Global
Symposium with regards to its internal processing characteristics. Once more, the
majority of these subjects expressed a negative opinion. The two majors grievances most
frequently reported included the relatively big size of the Global Symposium and the lack
of a clear and concise definition of the objectives of the event. I provide below, some
illustrative quotes from interview data.
Subject#1: you know you can only coordinate it if people, institutes want to be coordinated. And to do that you need a certain trust, it has to be a two way things, you cannot just come waving the coordination flag and expect everyone to lineup nicely. Subject#12: we spent so long figuring out what we were supposed to be talking about that we never got to the details. Subject#14: Well I think that we need to look at what is the purpose? What are you trying to achieve? and I would certainly not go in any kind of precooked formula which will reflects the earlier symposiums, on come up with a list of those and answer these questions. Subject#2: any type of meetings and workshops where you are gathering various organizations and numerous people I think it’s important to clarify terms and terminologies and I am not sure if this happened there. Subject#2: It’s always very difficult globally to bring people together. So I would say, first try to do it within a country or a region instead of trying to do it globally.
Another negative view expressed on the internal processing characteristics of the Global
Symposium was related to the lack of follow-up activities. A good number of interview
participants noted that the fact that there was no rigorous planning of follow-up activities
significantly compromised the effectiveness of the Global Symposium.
Subject#1: I don’t see any specific follow up or activities. Subject#11: It would have been good to have even virtually, not necessary another event, but if there was some follow up, it would have been good.
96
Subject#9: But I think we need follow up. Follow up in particular at the regional level; follow up at the country level. How they helped? And what are the achievements at the regional level.
Goal achievement
In the goal achievement category, effectiveness was defined as the extent to which
organization / network’s goal was achieved. The greater the degree to which an
organization / network achieves its goals the greater its effectiveness. During the
interviews, I recalled to the participants some of the main goals of the Global
Symposium. They included: (i) promote the use of humanitarian information
management principles; (ii) disseminate best practices of humanitarian information
management; (iii) improve the community’s preparedness in humanitarian information
management; (iv) help organizations/agencies acquire resources; (v) improve the level of
professionalization in the field of humanitarian information management (vi) foster
collaboration on humanitarian information management projects (vii) facilitate sharing
of expertise among organizations/agencies; (viii) promote humanitarian information
sharing; (ix) strengthen relationships between organizations/agencies; (x) increase
awareness of humanitarian information systems and (xi) improve humanitarian
information quality.
Approximately eight five percent (84.21%) of the interviewees discussed the
effectiveness of the Global Symposium with regards to achieving its goals. For most of
these respondents, the Global Symposium was effective in the pursuit of its goals. More
specifically, as illustrated in the following quotes from the interview data, the Global
Symposium was reported to be very effective in strengthening relationships between
organizations/agencies, in promoting the use of humanitarian information management
principles and in promoting humanitarian information sharing.
Subject#11: So I would say the Global Symposium was one of other events which promoted interactions among different partners. Subject#15: I think the event was mostly successful in coming to agreement among the various actors on certain standards for use of information in humanitarian response. So I
97
think the report was useful and the input that all the groups provided into that report was useful. Subject#19: the global symposium has actually been very important to us, as group of organizations that has been working on this effort. Because originally, this working group wasn’t on the agenda, but we requested that it be put on the agenda, to recognize the various information that affect the population. Subject#2: I mean I am talking specifically about one the positive things that came out with which is the ability to meet others in the field and in particular in bringing together the private and NGOs the UN communities which was good. Subject#8: I would say that there have been more information related projects and initiatives in the last two years and so I mean I think it encouraged information related projects.
However, for some interviewees, the impact of the Global Symposium was limited to a
single event (e.g Subject#15:). According to these participants, this was mainly due to
the lack of follow-up activities.
Subject#15: I think they disseminated [information on humanitarian information management best practices] but mostly to the conference participants. I don’t know how much this was disseminated beyond the conference.
Constituencies’ satisfaction
For constituencies’ satisfaction, effectiveness was defined the ability of network to satisfy
key multiple stakeholders. In the case of the Global Symposium, some of the main
stakeholders include the individual organizations, the governments, the United Nations,
the victims of humanitarian disasters and the international community. My study included
only members of the Global Symposium. They could be grouped into the following four
categories: NGOs, governmental organizations, private organizations and United Nations
organizations. I did not for example interview any victim of the humanitarian disaster.
Approximately eight five percent (84.21%) of the interviewees discussed the
effectiveness of the Global Symposium with regards to constituencies’ satisfaction.
Almost all of them expressed a mixed feeling about the effectiveness of the Global
Symposium in satisfying its multiple stakeholders.
Subject#8: I think my organization has used the symposium to a high degree. I would say the broader community has lessen so. I mean at a scale of 1 to 10, I think that the HIU
98
has used and benefited from the symposium probably of a 7 or 8, and the entire humanitarian community at a 4 or 5. Subject#15: Well I think the event was mostly successful in coming to agreement among the various actors on certain standards for use of information in humanitarian response. So I think the report was useful and the input that all the groups provided into that report was useful.
Figure 14: Network Effectiveness Code’s Coverage
Across the interviews, the intensity of the discussion on network effectiveness varied
significantly depending on the code category (Figure 15).
99
Figure 15: Network Effectiveness Code’s Loudness
Exploring my data cross networks (Figure 16), I found almost a similar pattern in the
ranking of the loudness of the different categories of network effectiveness. In all the
networks, the category “goal achieved” was ranked first followed respectively by
“internal processing”, “constituencies’ satisfaction” and then “resource availability”. The
pattern of the percentage of discussion of the different categories was also similar. They
were over fifty percent (50.00%) for the “goal achieved” category, approximately twenty
two percent (22.00%) for the “internal processing” category; approximately fifteen
percent (15.00%) for the “constituencies’ satisfaction” category; lest than ten percent
(10.00%) for the “resource availability” category.
While conducting the interviews, I was also interested in assessing the participants’
opinion about factors for collaboration among members of the Global Symposium. I
present below the major factors that emerged from the interviews. A total of seven factors
were found including mandate/goals, skills, trust/reputation, funding, size, geographical
proximity/language and processes. The identification of these factors was guided by the
literature.
Mandate /Goals
One of the most frequently cited factors for inter-organizational collaboration on
humanitarian information management and exchange was related to the similarity in the
mandate and goals of the organizations wishing to work together. About eighty five
101
(84.21%) of the interviewees discussed the similarity of mandate / goals as being an
important motive for their organization to engage into collaboration with another
organization. I provide below, some illustrative quotes from the interview, expressing this
point of view:
Subject#9: We have to have a common work plan in order to work together. Subject#17: What it takes for us to collaborate is just a kind of share objective.
The issue of mandate and goals as driving factor for inter-organizational collaboration
was not only cited by the greatest percentage of participants, this factor was also among
those that were the most intensively discussed. It represented approximately twenty nine
percent (28.38%) of occurrences of collaboration factors discussed in all the interviews
combined (Figure 19).
Figure 17. Factor’s Coverage
Skills
The skills set of the potential collaboration partners was the second most reported factor
for inter-organizational collaboration among the members of the Global Symposium.
Approximately seven four percent (73.68%) of the participants reported that in deciding
102
to engage into collaboration, their organization would highly consider the type of skills of
the potential partners (Figure 17). The issue of skills as factor for collaboration was
discussed in two different perspectives. The majority of participants, who mentioned this
issue, discussed it in term of availability of complementarity set of skills in the potential
partner organizations. Their organization would collaborate with another organization if
the later possessed a set of skills that they lacked. For example, participant number two
(Subject#2) said:
Subject#2: Both have to be able to bring to the table their competitive advantage. You can’t have two organizations that do the same thing. So you need different skills set from any of the organizations.
Other participants discussed the issue of skills in term of high quality and competency. In
deciding to engage into collaboration their organizations would consider the quality and
competency of the skills available to the potential partner organizations.
Subject#7: We think about the quality of what that agency does and the quality of what that agency is known to do. Subject#17: We are trying to be a service provider to those organizations. So I guess we are trying to provide a competency. But we also have interest in the ability of these other organizations to develop new competencies.
Similarly to the issue of mandate and goals, the discussion of skills as factor of
collaboration was also very intense. This factor also represented approximately thirty
percent (29.73%) of occurrences of collaboration factors in all the interviews combined
(Figure 19).
103
Figure 18. Factor’s Loudness
Trust / Reputation
The third most discussed factor for inter-organizational collaboration among the members
of the Global Symposium was trust and reputation. Approximately forty eight percent
(47.37%) of the interview participants mentioned that they will more likely not get into
collaborative activities with an organization that they do not trust or an organization that
has a poor reputation (Figure 17). Below, I illustrate this point of view with quotes from
subject number five and subject number eight.
Subject#5: I thing on the one hand getting the quality information which is credible and I think the emphasis has to be on the word credible because there is no good having information which is bad because people see through that very very quickly and you can lose your credibility very quickly. Subject#8: we are looking for partners that have a good reputation that provide value added to what we can provide.
This third most reported factor of collaboration was also very intensely discussed. It
represented roughly seventeen percent (16.22%) of occurrences of collaboration factors
in all the interviews combined (Figure 18).
Geographical Proximity / Language
Geographical proximity was the fourth most reported factor for inter-organizational
collaboration among the members of the Global Symposium. Approximately thirty seven
104
percent (36.84%) of the interview participants mentioned proximity and especially the
language as an important factor that drives collaboration (Figure 17).
Subject#16: I think we are gaining popularity among humanitarian here because of the Spanish.[….] Spanish is a really an important and imperative thing if you want to enter here. English is a secondary one and really a secondary one. So that is why I think we are gaining popularity among humanitarian here because of the Spanish. Subject#18: if people are approaching us, I would say that first that because we are French and people that approach us usually are also French NGOs.
The intensity of discussion around the geographical proximity and language as factor for
collaboration was relatively low as to compare with others factors. This factor accounted
only for roughly eleven percent (10.81%) of occurrences of collaboration factors in all
the interviews combined (Figure 18).
Size
The fifth most discussed factor for inter-organizational collaboration among the members
of the Global Symposium was the size of the potential partners. Approximately sixteen
percent (15.79%) of the interview participants reported that they consider the size of the
organizations that approach them to seek for collaboration (Figure 17). Participant
number one for example reported that:
Subject#1: We tend to work I guess it is natural, we tend to work better with the smaller entity that seem to be more flexible, more users oriented than big entities, be they national entity or the private companies or of course UN entities.
The issue of size was also among the factors the least intensively discussed. It accounted
only for roughly seven percent (6.76%) of occurrences of collaboration factors in all the
interviews combined (Figure 18).
Funding
Funding was ranked sixth most reported factor for inter-organizational collaboration
among the members of the Global Symposium. Approximately sixteen percent (15.79%)
of the interview participants reported that they look at the funding possibilities available
at the potential partners before deciding to engage into collaboration (Figure 17).
Subject#8: we all provide any funding for you know we contribute to our work, and they contribute to their work. So you know they need to be sort of self-sufficient.
Surprisingly, the issue of funding as factor of collaboration was among the factors the
least intensively discussed. This factor represented only approximately eight percent
105
(5.41%) of occurrences of collaboration factors in all the interviews combined (Figure
18).
Processes
The seventh and last reported factor for inter-organizational collaboration among the
members of the Global Symposium was the processes. Nearly eleven percent (10.53%) of
the interview participants reported that they engage into collaboration with partners that
follow clearly predefined processes (Figure 17). This was especially important for
collaborating with donors organizations.
Subject#14: We have very well defined process for people, partners contacting us.
The issue of processes the least intensively discussed. It accounted for less than three
percent (2.7%) of occurrences of collaboration factors in all the interviews combined
(Figure 18).
Borrowing from the framework developed by Ngamassi et al., (2011) to analyze factors
that hinder inter-organization coordination and collaboration among humanitarian
organizations, the seven factors identified in this study could be grouped into the
following three categories: organizational, structural and behavioral (Figure 19). The
organizational category would include factors related to the mandate/ goals and the
processes. The factors in the structural category would be skills, funding, size and
geographical proximity. The last category, behavioral, would include factor related to
trust and reputation. This categorization allows to have another perspective of the
influential drivers of inter-organizational collaboration in the humanitarian relief field.
Figure 19 below, depicts the aggregated loudness of factor per category.
106
Figure 19: Loudness of Collaboration Factors Grouped per Category
It is important to note that nearly half of the reasons to collaborate are structural. This is
important because these structural factors are the ones most likely to be supported and
affected by the use of information technologies. The use of this framework helps to
highlight the fact information technologies have important potential to influence inter-
organizational collaboration relationships among humanitarian organizations.
Analyzing cross networks (Figure 20), the loudness of the different collaboration factors
that emerged from the interviews I made the following two observations. First, there was
a similar pattern in the ranking of the different factors of collaboration based on the
number of their occurrences. The structural factors were the most reported cross networks
followed respectively by organizational and lastly the behavioral factors. Second, I found
a wide discrepancy in the loudness of the different factors, cross networks. For example,
in the network of the United Nations agencies the discussion around the structural factors
represented approximately sixty seven percent (66.67) of the discussion related on
collaboration factor in that network. This proportion was fifty (50%) for the network of
Governmental organizations and just about forty two percent (41.67) for the network of
non-governmental organizations. Another important discrepancy was observed on the
collaboration factors grouped in the behavioral category. Discussion around this category
107
represented less than ten percent (9.52%) in the network of the United Nations agencies
and approximately twenty percent in the two other networks.
Figure 20: Loudness of Collaboration Factors Cross Network
Information Technologies
During the interviews, I also asked participants to give their opinion specifically on the
implications of information technologies on inter-organizational collaboration among
members of the Global Symposium. Approximately half (47.37%) of the interviewees
shared their opinion on this issue. I registered a wide range of diverse point of views.
Some participants, roughly thirty one percent (30.77%) of those that answered the
question, had a very positive opinion about the implications and especially the catalytic
role that information technologies play in fostering humanitarian inter-organizational
collaboration. The vast majority (69.23%) however, expressed mixed feelings.
For the participants that had positive opinions, information technologies served as an
important catalyst for inter-organizational collaboration in the Global Symposium
community. They argued that if without information technologies, effective simple
108
communication is difficult, collaboration would be even harder. Participant number five
for example reported that:
Subject#5: I think that information technology is extremely important because we basically need to communicate to all these different communities in as many different ways as possible.
They also believed that the use of information technologies is instrumental in quickly
gather analyze and disseminate humanitarian information leading to effective disaster
response. Below, I illustrate this point of view with quotes from three participants,
number six, seven and eleven.
Subject#6: You cannot do it without information technology. Gathering information, managing information, analyzing information, distributing information, really you cannot do all this without information technology. So I think the question is kind of obvious. Subject#7: Information technology essentially supports what we do. It helps in sharing information, mainly transporting information around, maintaining our communication. Subject#11: I think the information technology is key of cause, because without proper systems in place, you will not be able to do that.
The participants who expressed mixed feelings about the role of information technologies
as catalyst in inter-organizational collaboration believed that taken alone, information
technologies would not lead to better / more collaboration. They gave a number of
reasons that could be grouped into two main categories. The first category of reasons was
related to the information technologies infrastructure. Participants argued that more often,
organizations in the field do not have the necessary technology tools either because they
were destroyed by the disaster or because they did not even exist in the first place. They
also talked about the discrepancy in term of infrastructure between organizations based in
developed countries and those in the developing countries. They argued that people in
developed countries often enjoy latest technologies but the realty in developing countries,
scenes of most humanitarian disasters is quite different. Participant number twelve for
example reported that:
Subject#12: when you get out on the fields you see that the most basic important tool is paper map and a pencil. And I think we have got to really recognize that fact. […]You know we do this information technology that we love where they follow the latest systems and the fastest processor and stuff like that and we really like to paddle
109
ourselves on the back on what we are able to do here in Washington DC. And then you get out on the fields and everyone is using paper maps and a pencil.
Finally they talked about the fast pace of change in technology which makes it difficult
for organizations to have and especially keep the technical staff that possesses adequate
knowledge to make use of these new technologies.
Subject#6: as the technology changes, it is hard to find the people that have skills that are up to date.
The second category of reasons concerned the management of information. Participants
believed that without proper standard for humanitarian information exchange, the
technology will be of no effective use.
Subject#5: I think yes, continue to explore all the new technologies that are available but at the same time realize that in the end what it really comes down to is quality information and information that is based on facts and that’s credible but people actually belief in. So I think we should not be allowed to be measured by technology if the content is not there. Subject#14: One is developing some basic standards, and some basic platforms for information exchange.
They also believed that the humanitarian field needs better processes and well trained
staff in order to make good use of the technology.
Subject#8: I think there are certain organizations who think that technology can solve all the problems, so they don’t have a proper appreciation and understanding of the information management challenges and obstacles, but at the same time there is probably some information, people who are very skeptical about technology and do not sort of realize the value that it has.
6.2.1.4 Collaboration barriers
The interview included questions about barriers to inter-organizational collaboration. I
asked the interview participants to identify the major barriers to inter-organizational
collaboration among members of the Global Symposium on humanitarian related
projects. The lack of leadership, extensive bureaucracy, the lack humanitarian
information standard, the lack sharing spirit the lack of skills and the lack of resources are
some of the most frequently reported barriers to inter-organizational collaboration among
members of the Global Symposium. I coded these barriers using three categories
including (i) structural, (ii) behavioral and (iii) mandated. These categories were once
110
more borrowed from the framework developed by Ngamassi et al., (2011) to analyze
factors that hinder inter-organization coordination and collaboration among humanitarian
organizations. As discussed by Ngamassi et al., (2011), this higher order analytical
structure that organizes these collaboration barriers into three larger categories is
appropriate to both the context of humanitarian relief as well as to collaboration among
organizations engaged in humanitarian information management and exchanged. I
present below, these barriers in order of their intensity as discussed by interview
participants.
Structural
Collaboration barriers in the structural category included barriers such as extensive
bureaucracy, lack of humanitarian information standard, problems of communications /
language, size of organization, lack of tools (IT/IM) for collaboration, geographical
distance, lack of technical skills, lack of resources and lack of leadership. This category
represented the most frequently reported barriers to collaboration. All of the participants
to the interview identified at least one collaboration barrier that fell into the structural
category. In other to refine my investigation and pay more attention to IT and IM related
barriers, I distinguished the following three subcategories of structural barriers.
Information Management (IM) related
Approximately eighty five percent (84.21%) of the interview participants talked about
challenges to inter-organizational collaboration related to information management.
Issues such as information quality, information standards and information security were
frequently reported.
Subject#5: I thing on the one hand getting the quality information which is credible and I think the emphasis has to be on the word credible because there is no good having information which is bad because people see through that very very quickly and you can lose your credibility very quickly. Subject#14: For instance, information security, you know, that is becoming more and more of a concern. It used to be that you could pretty much share information freely, but now it is not more the case. [….] Subject#7: there are things like not sharing security information because you think it is so important to you.
111
Another major barrier to inter-organizational collaboration related to information
management that emerged from the data concerns the language.
Subject#16: Spanish is a really an important and imperative thing if you want to enter here. English is a secondary one and really a secondary one. So that is why I think we are gaining popularity among humanitarian here because of the Spanish. But unfortunately we are not gaining worldwide reach or even in headquarters because we do not have too many Spanish readers so they do not see this importance. Subject#18: I would say that first that because we are French and people that approach us usually are also French NGOs […] as I have mentioned in your second survey who have establish relationship with the Groupe URD and that is because we are French.
Information management related barriers to inter-organizational collaboration were also
the most intensively discussed. It represented approximately forty percent (40%) of
occurrences of collaboration barriers discussed in all the interviews combined.
Information Technology (IT) related
Approximately seventy four percent (73.68%) of the interview participants discussed
challenges to inter-organizational collaboration related to information technology. They
talked for example about some technology tools that are not wide spread and are used
only specific organizations.
Subject#2: And in the same way these organizations are all doing information management and a lot of these organizations have these tools which they only know about within that organization. And so we talk about the community but in reality there isn’t much of a community.
They also talked about lack of IT skills
Subject#8: I think there are certain organizations who think that technology can solve all the problems, so they don’t have a proper appreciation and understanding of the information management challenges and obstacles, but at the same time there is probably some information, people who are very skeptical about technology and do not realize the value that it has.
112
Barriers to inter-organizational collaboration related to information technology
represented approximately thirty three (33%) of occurrences of collaboration barriers
discussed in all the interviews combined.
Other structural barriers
I grouped in this category all other structural barriers that were identified and that were
neither IT nor IM related. Approximately fifty eight percent (57.89%) of the interview
participants discussed challenges to inter-organizational collaboration that fell into this
category. The most frequently reported barrier in this category was the lack of
humanitarian dedicated staff and also the competition for funding.
Subject#18: In fact the main challenge is human resources that are dedicated and that have time to do the work. Because why? Most of the people that were there were note really specialist or were not fully dedicated to the job of information management. […] you cannot ask someone to share information if it is not his job. You cannot ask someone to produce a map with the right standard with the right quality if it is not his job. Subject#5: Unfortunately what is happening now is that there are too many organizations running after the money, running after they think what others want, and not running after the real needs. Subject#16: I would say that the most difficult part ummmm. I use a word here “humanware”. [….] In my opinion, and considering my experience in Africa, in Asia this is where most of the information management systems are struggling to survive or to go ahead, to move forward to achieve their objectives. Subject#11: Very often we have a situation that the information in available and everything but who is able then to present it, to analyze it, to prioritize it, and all of these, that is for me the role in information management, or information management for doing that, I think for me one of the key issues.
These other structural barriers represented approximately twenty seven percent (27%) of
occurrences of collaboration barriers discussed in all the interviews combined (Figure
21).
113
Figure 21: Break Down of Structural Barriers
Behavioral
Inter-organizational collaboration barriers in the behavioral category included barriers
such as lack of sharing spirit, lack of trust, lack of incentives. They were the second most
frequently reported barriers to collaboration. Approximately forty seven percent
(47.37%) of the interview participants identified at least one collaboration barrier that fell
into the behavioral category (Figure 21). I present below some illustrative quotes drawn
from the interview data.
Subject#12: There is a big problem with information sharing. But that you know that’s the problem of the world. I do not know if that is a very specific problem with these organizations. Subject#13: I think the main challenge here is that the idea of sharing formation has always been said in many areas. It is usually always said yeah it is good to share but you do not sometime see concrete platforms or formalities on how to share this information. It is not formalize. It is always thought as an objective but never formalize.
Behavioral barriers represented approximately sixteen percent (16.10%) of occurrences
of collaboration barriers discussed in all the interviews combined.
114
Mandated
Collaboration barriers in the mandated category included barriers such as conflict of goal,
conflict of interest and lack of leadership. This category of barriers was the third and last
most frequently reported barriers to collaboration. Approximately twenty one percent
(21.05%) of the interview participants identified at least one collaboration barrier that fell
into the mandated category (Figure 21). I present below some illustrative quotes drawn
from the interview data.
Subject#9: I mean the different administrative work of organization is very difficult and it varies from organization to organization. There is no commonality among organizations. Subject#5: I think the problem is getting the decision making of all the organizations to actually understand what the issues are and have to understand that they have the responsibility. Subject#9: I think there is the issue of contingency plan. Contingency plan is very important, contingency fund. Because by the time funding happens it might be already too late. I know that in some countries they already have it at the governmental level but where we work where the government is very weak or even nonexistent you do not have contingency plan. So it depends on international actors to provide that type contingency fund that can be utilized during emergency.
Inter-organization collaboration barriers the mandated category represented
approximately five percent (5.08%) of occurrences of collaboration barriers discussed in
all the interviews combined.
115
Figure 22: Loudness of Barriers to Collaboration Grouped per Category
Summing up, the most significant barriers to inter-organizational collaboration among
organizations in the humanitarian relief field are structural (Figure 22 above).
Exploring the intensity of the different collaboration barriers cross networks, I made the
following observation (Figure 23): Behavioral barriers to inter-organizational
collaboration were less discussed in the network of non-governmental organizations
(approximately ten percent – 10.26%) than in the two other networks where this
percentage was almost double (21.88% for the network of Governmental organizations
and 17.39% for the network of the United Nations agencies).
116
Figure 23: Loudness of Barriers to Collaboration Cross Network
6.2.1.5 Measures of effectiveness
Effectiveness is a multidimensional concept that is especially challenging to measure in
humanitarian assistance and disaster relief which often involve a large variety of
stakeholders with diverse goals and for which outputs are not easily operationalized. One
other objective of my interview was to get the opinion of the member of the Global
Symposium of what would make an appropriate metric for measuring network
effectiveness in their community. About sixty nine percent (68.42%) of the interviewees
answered this question. Guided by the literature, I coded the data in the following four
categories, range of activities, level of coordination, level of collaboration and
availability of resource including funding. The intensity of discussion was almost evenly
distributed cross these categories. I provide below some illustrative quotes for each of the
categories.
117
Range of activities
Subject#1: I think if we were to be an effective symposium, one thing would be add new activities that could be implemented within a reasonable timeframe. [….]Research activity, research projects; or even operation activities that could be implemented within a reasonable short time following direct results you know, or direct resources or direct resolution from the symposium. Subject#18: I think for me, I would look at the number of implemented projects.
Level of coordination
Subject#17: I think it would definitely be knowledge exchange. I think at this point, I would want to be able to read, learn about things that I didn’t know anything about or features about things that I have just heard about, and then and those things that I would assume would be facilitate and lead to collaboration. […] So the first step of that would be sufficient information sharing. An then one of things that I would expect to see is the consolidation of this and a lot less competition and a lot more synergy. So one of the eventual outcomes is that we will just become a lot more collaborative and a lot more of these systems talk to each other and a lot more links to one another, and we don’t do as much replication of efforts. Subject#4: organization inter-operate although they are competing for donors whom they appeals, I would seek to analyze the attitude of all those organizations, to see to what extent they compete in the market and to what extent they understand what synergy mean.
Level of collaboration
Subject#10: I think you need to look at the level of coordination and funding. How much of funding have organizations successfully secured to work in this area? The extent to which there are working with other partners or coordinating. Subject#17: I think it would definitely be knowledge exchange. I think at this point, I would want to be able to read, learn about things that I didn’t know anything about or features about things that I have just heard about, and then and those things that I would assume would be facilitate and lead to collaboration. […] So the first step of that would be sufficient information sharing. An then one of things that I would expect to see is the consolidation of this and a lot less competition and a lot more synergy. So one of the eventual outcomes is that we will just become a lot more collaborative and a lot more of these systems talk to each other and a lot more links to one another, and we don’t do as much replication of efforts. Subject#4: organization inter-operate although they are competing for donors whom they appeals, I would seek to analyze the attitude of all those organizations, to see to what extent they compete in the market and to what extent they understand what synergy mean.
118
Funding and other resources
Subject#10: I think you need to look at the level of coordination and funding. How much of funding have organizations successfully secured to work in this area? The extent to which there are working with other partners or coordinating. Subject#11:it is probably easier to use the money which has been given because that will at least express a certain level of satisfaction of what we are doing. Because we are funded by volunteering contribution from donors. So at least I would say that if we get a lot of money for one of the other projects that at least indicate the level of satisfaction from our stakeholders. So maybe that is a better one. Subject#3: I think another way which is being relatively successful is the way you have the agencies to have high number of technical people in them. Especially when they come from a professional background where you can do humanitarian practices across agencies where people really know how to improve the competency or work.
6.2.2 Inductive codes
The inductive coding process of my interview data yielded three set of codes that I
believe would help to shed more light in inter-organizational collaboration in the Global
Symposium community and consequently to better understand the effectiveness of this
community in providing disaster assistance. These three code categories included (i)
from advice to collaboration, (ii) the scope of the Global Symposium community and (iii)
the audience / stakeholders.
6.2.2.1 From advice to project collaboration
The first inductive code was related to the connection between advice and project
collaboration relationships. My data highlighted the fact that in the Global Symposium
community, organizations that are linked through advice relationship would in a long run
collaborate in humanitarian relief project.
Subject#1: yes I would say so, I mean it is not humongous impact, but it is an important element as well. There are people and entities we met that we are now discussing with and sharing information with sharing ideas with and you know just keep in touch at an informal level. I think that is very good for us. Also because the UN being UN some
119
institutions are more difficult to approach officially or let say institutionally. While if you have this more ad hock loose network where you could exchange without it become very formal that is very useful. Subject#8: I mean you know, I have been in this community for a long time and I have a lot of informal contacts with people asking for advice or asking for you know and you it sort of this informal, this sort of one on one type of thing.
6.2.2.2 Network Scope
The second inductive code that emerged from the data concerned the scope of the Global
Symposium. Some interview participants reported that the large size of the Global
Symposium would more likely negatively impact its effectiveness. Other participants also
highlighted the fact that the “ambitious objectives” of the Global Symposium would
undermine its effectiveness.
Subject#2: I would say, first try to do it within a country or a region instead of trying to do it globally. Because then you have a smaller community and those community are much more important, the regional community or the national community. Subject#2: Tried to strive for much less ambitious objectives and discuss some of the core issues within each of those sub sectors if you will or the sectors. Subject#5: there should be smaller groups that held very very specifically with mixed of media communication people and these organizations perhaps have small groups that meets for one day but in a highly intensive manner, and really look at the issues maybe to review what happened in the last symposium, but reviewed this in a very very pragmatic manner and a very outspoken critical manner as well. Subject#11: May be a smaller group, because it was rather a large event, so maybe if you could do it regionally, let say one in Latin America, and another one in Africa or central Africa, west Africa, maybe that would be more effective, because you would have fewer participants. On the other hand it would not necessary always get the global perspective, so then you would not have everybody in one place western countries, eastern countries, developing, developed countries and whatever. May be a mixed you know a regional approach is not too bad. It can also be as an advantage. Subject#12: I think it needs to be longer basically. […] We spent so long figuring out what we were supposed to be talking about that we never got to the details. And I remember I think we were only in that room for like you know a day or less than a day or something like that. […] And these are big difficult issues and I think you need to spend a lot longer on them rather of assuming that you are going to come up with answers in a few hours especially when you have so many different organizations at the table.
120
Subject#12: I think it would be a very very very good idea for just you knows the UN organizations themselves to get together. And not try and bring on all those non-governmental NGO types. Because you know the UN needs to kind of figure out its own game and how it wants to interact with the rest of us. Subject#16: In this region, it seems that this type of global agreement or these international events are not having strong impact this is my perception at this point.
6.2.2.3 Network Audience
The third and last inductive code that emerged from the data was related to the audience
of the Global Symposium. Interview participants would have liked to have a more
diversified audience especially people from the field of disaster reduction.
Subject#1: For the other external partners or external participants, it would be nice to see their contributions are really recognized, but for them it would be less important but for us it would be a major boost because it would be an overall recognition of our work and what we have been doing since 2001, but there were some forces in OCHA that for personal reasons did not want to see that. Subject#5: One of the things that we were suggesting, is that when you have a symposium like this, it doesn’t help just to have media or information people talking among themselves. We need to mainstream. We need to bring in the heads, in fact the very senior people within these organizations; whether it is the UN agencies such as FAO or UNHCR. They should not be communication people, they should be operation people. Subject#5: I think that what need to be done is that the major organizations and the senior operation people needs to be invited and included, and also may even see in them. They should say, look unless you are serious about his, then you shouldn’t be in the business. But if you are serious about helping refugees, displaced people, you name it, then you have another responsibility to communicate with these people. Subject#10: As I said these thing needs to be mainstream at all levels of the disaster response rather than be treated as a separate area of expertise or data collection. Subject#17: I thought the symposium would have been extremely relevant to national platforms for disaster prevention or disaster reduction what even they are calling them. And I know that ISDR was part of the planning process, but I think that, as far as I know, the part of ISDR that was most internally involved in the planning process was the Information Management Division and they did not involve the rest of ISDR and so nobody realized that this was a tremendous opportunity to bring some very specifically qualified people from national disaster reduction platforms. And you know there is probably a hundred of those. And you know rather than… usually it’s important to
121
participate in things and we never get to actually get the people together who do the information systems work. That would have been an incredible addition to the audience. Subject#12: I think it would be a very very very good idea for just you know the UN organizations themselves to get together. And not try and bring on all those non-governmental NGO types. Because you know the UN needs to kind of figure out its own game and how it wants to interact with the rest of us.
6.3 Effectiveness Measures
As I stated in the previous chapter, the measurement of effectiveness has always been a
nagging and unsolved problem for inter-organizational network researchers. There is no
consensus on the criteria of measuring effectiveness among researchers. Prior research
has used wide varieties of measures (see Table 4). These measures include the perception
of solving problems, decreased service duplication, improved coordination (Provan &
Milward,1995); service quality (Grusky, 1995); and perceived benefit to various
stakeholders of the network (Weech-Maldonado et al., 2003). In my research, I use three
different measures of network effectiveness including one subjective (perceived network
effectiveness) and two objectives measures (number of funded projects and number of
funding partners). The number of funded projects measures effectiveness in term of level
of activities in humanitarian assistance while the number of funding partners measures
effectiveness in term of level of collaboration.
6.3.1 Perceived Network Effectiveness
I used as one criteria of effectiveness, the perception of the members of the Global
Symposium community about the effectiveness of their humanitarian information
management and exchange network. As discussed earlier, subjective measures of
effectiveness have been widely used in previous research. I also chose this measure in
order to take into account the context of my study. My survey instrument included a five
point Likert scales question that asked respondents about their perception of the
effectiveness of the network on the following items: (i) dissemination of best practices
and humanitarian information principles; (ii) accessibility to resources; (iii) community
development; and (iv) knowledge and information exchange. These items were drawn
122
from my prior research and also from the literature. Network effectiveness ranged from 1
(strongly disagree) to 5 (strongly agree). Cronbach’s alpha coefficient for this four-item
scale was .83.
For each of the three networks investigated, I computed the mean score of the responses
on each of the items. The results are presented in Table 7 below. I also conducted an
independent-samples t test to evaluate the differences among the three networks on these
items. I found that there was not a statistically significant difference.
Perceived Effectiveness
Network
Governmental
Organizations
(n= 12)
Non-Governmental
Organizations
(n= 17)
United Nations
Agencies
(n= 11)
Dissemination of best practices
and information principles
M
(SD)
2.14
0.98
2.27
0.64
2.45
1.00
Accessibility to resources M
(SD)
3.08
1.44
2.82
1.29
2.18
0.60
Community development M
(SD)
2.30
1.09
2.59
0.80
2.06
0.49
Knowledge and information
exchange
M
(SD)
2.17
0.98
2.40
0.87
1.93
0.69
Mean score 2.42 2.52 2.16
Table 7: Perceived Network Effectiveness Index Table
My first general observation was that the overall patterns of results within each of the
three networks (governmental organizations/agencies – GO; non-governmental
organizations – NGO; and United Nations agencies – UNA) were similar. The main
contrasting difference observed concerned UNA. This network registered the lowest
score on the following three items: accessibility to resources (UNA score: 2.18; mean
score: 2.70), community development (UNA score: 2.06; mean score: 2.35), and
knowledge and information exchange (UNA score: 1.93; mean score: 2.21). Conversely,
perceptions of respondents of UNA revealed this network is more effective in the
dissemination of best practices and humanitarian information management principles.
UNA displayed the highest score on this (UNA score: 2.45; mean score: 2.27). When
considering all the different survey items on which network effectiveness was measured,
123
accessibility to resources came first in two out of the three networks including GO (AR
score: 3.08; mean score: 2.42) and NGO (AR score: 2.82; mean score: 2.25).
In order to reflect the overall network level of perceived effectiveness, I developed an
index by averaged the four factor scores generated for each network to produce a single
mean item score for that network. Based on this mean score, NGO displayed the highest
level of network effectiveness follow respectively by GO, and UNA in the last position.
6.3.2 Level of Activities and Level of Collaboration
The two objective measures for assessing effectiveness that I used were respectively the
number of funded projects and the number of funding partners in humanitarian relief. I
used the number of funded projects as proxi measure for the level of activities while the
number of funding partners was the measure of the level of collaboration. These indices
are important performance factors in humanitarian disaster assistance. Both measures are
related to the concept of social capital. Social capital refers broadly to characteristics of
social structure that function as a resource for individuals and groups. Putman (1993)
defines social capital as the “features of social organization, such as trust, norms and
networks that can improve the efficiency of the society by facilitating coordinated
actions” (P. 167). Social capital can be interpreted as combining a structural component
consisting of involvement in voluntary associations and a cultural component consisting
of norms, values and trust. In my study, I used this interpretation of social capital. The
structural component of social capital was measured by the level of collaboration while
the cultural component was measured by the level of activities. As I discussed earlier, I
considered that greater level of activities and or greater level of collaboration was
associated with higher level of effectiveness. One outcome of voluntary interaction
among members in a community is the development of social trust that facilitates
collective social action toward achieving common social goals. The level of collaboration
in a community is therefore a function of interaction among members via their social
networks. As the level of collaboration increases, so does the effectiveness of the
community in achieving its goals. Thus communities with vibrant communication
networks are likely to display higher level of effectiveness.
124
Moreover, I considered the opinion of research participants to choose these two measures
for my study. For instance, during interviews, I asked participants about what in their
opinion would be an appropriate measure that could be used to assess network
effectiveness in their community. More than half of the interviewees (53%) answered
this question. Not surprisingly, I registered a diversified range of responses.
Summarizing, approximately nine (9) different criteria for assessing network
effectiveness emerged from the interviews. They included (i) the range of activities
provided by the network, (ii) the level of preparedness in of the Global Symposium
community in responding to humanitarian disaster especially with regards to information
management and exchange, (iii) the level of coordination in the network, (iv) the level of
collaboration among members (v) the availability and access to funding (vi) the timely
response to crises especially with regards to information sharing (vii) the level of use of
best practices (viii) the availability and access to resources especially technology tools
and technical staff, and finally (ix) the level of attendance to Global Symposium events.
I present in the table below (Table 8) some illustrative quotes from the interview data
concerning these effectiveness measures.
Effectiveness measure
Illustrative quotes
Range of activities
I think if we were to be an effective symposium, one thing would be add new activities that could be implemented within a reasonable timeframe. [….]Research activity, research projects; or even operation activities that could be implemented within a reasonable short time following direct results you know, or direct resources or direct resolution from the symposium. (Subject#1)
I think for me, I would look at the number of implemented projects (Subject#18)
Level of Coordination
I think it would definitely be knowledge exchange. I think at this point, I would want to be able to read, learn about things that I didn’t know anything about or features about things that I have just heard about, and then and those things that I would assume would be facilitate and lead to collaboration. […] So the first step of that would be sufficient information sharing. An then one of things that I would expect to see is the consolidation of this and a lot less competition and a lot more synergy. So one of the eventual outcomes is that we will just become a lot more collaborative and a lot more of these systems talk to each other and a lot more links to one another, and we don’t do as much replication of efforts. (Subject#17)
organization inter-operate although they are competing for donors whom they appeals, I would seek to analyze the attitude of all those organizations, to see to what extent they compete in the market and to what extent they understand what synergy mean. (Subject#4)
125
Effectiveness measure
Illustrative quotes
Level of collaboration
I think you need to look at the level of coordination and funding. How much of funding have organizations successfully secured to work in this area? The extent to which there are working with other partners or coordinating. ( Subject#10)
I think it would definitely be knowledge exchange. I think at this point, I would want to be able to read, learn about things that I didn’t know anything about or features about things that I have just heard about, and then and those things that I would assume would be facilitate and lead to collaboration. […] So the first step of that would be sufficient information sharing. An then one of things that I would expect to see is the consolidation of this and a lot less competition and a lot more synergy. So one of the eventual outcomes is that we will just become a lot more collaborative and a lot more of these systems talk to each other and a lot more links to one another, and we don’t do as much replication of efforts. (Subject#17)
organization inter-operate although they are competing for donors whom they appeals, I would seek to analyze the attitude of all those organizations, to see to what extent they compete in the market and to what extent they understand what synergy mean. Subject#4
Funding and other resources
I think you need to look at the level of coordination and funding. How much of funding have organizations successfully secured to work in this area? The extent to which there are working with other partners or coordinating. (Subject#10)
[…]it is probably easier to use the money which has been given because that will at least express a certain level of satisfaction of what we are doing. Because we are funded by volunteering contribution from donors. So at least I would say that if we get a lot of money for one of the other projects that at least indicate the level of satisfaction from our stakeholders. So maybe that is a better one. (Subject#11) I think another way which is being relatively successful is the way you have the agencies to have high number of technical people in them. Especially when they come from a professional background where you can do humanitarian practices across agencies where people really know how to improve the competency or work.(Subject#3)
Table 8: Choosing Effectiveness Measures: Illustrative Quotes from the Interview
It is by analyzing and trying to synthesize this wide range of diverse opinions on
effectiveness measures and by considering the findings of my previous research in this
community (Ngamassi et al, 2010) that I chose to use the number of funding partners and
the numbers of funding projects. As mentioned earlier, I collected the data related to the
number of funded projects and funding partners from the ReleifWeb Financial Tracking
Service, a UNOCHA web based database which records all reported international
humanitarian financial assistance. The the ReleifWeb Financial Tracking Service was
126
implemented and launched in 1999. In the humanitarian relief literature, data from the
ReleifWeb database has been used in a number of academic work and reports to donors
Though it would be risky to generalize about research results from a sample of only three
networks in a single area of humanitarian information exchange, my study contributes to
the literature on inter-organizational humanitarian networks in a number of ways. Firstly,
building on Provan & Sebastian (1998), my study further highlights the need to consider
network analyses in smaller substructures than what has been done previously. Large
scale integration across an entire network of organizations is difficult to achieve and is
probably not a very efficient way of organizing (Provan & Sebastian, 1998). For
instance, in the field of humanitarian relief field, disaster response often involves
heterogeneous organizations, both for-profit and nonprofit, with a wide range of different
characteristics. In this field, achieving effective inter-organizational collaboration is more
138
challenging especially with regards to information management and exchange (Ngamassi
et al, 2011; Maitland et al., 2009). As the findings of my research suggest, it is more
appropriate to assess network effectiveness in smaller substructures such as subnets or
cliques. These findings derived from quantitative analysis corroborated with results from
qualitative data. For instance, some interview participants reported that the large size of
the Global Symposium would more likely negatively impact its effectiveness. I provide
below, illustrative quotes from Subjects #2, 5 and 11.
Subject#2: I would say, first try to do it within a country or a region instead of trying to do it globally. Because then you have a smaller community and those community are much more important, the regional community or the national community. Subject#5: there should be smaller groups that held very very specifically with mixed of media communication people and these organizations perhaps have small groups that meets for one day but in a highly intensive manner, and really look at the issues maybe to review what happened in the last symposium, but reviewed this in a very very pragmatic manner and a very outspoken critical manner as well. Subject#11: May be a smaller group, because it was rather a large event, so maybe if you could do it regionally, let say one in Latin America, and another one in Africa or central Africa, west Africa, maybe that would be more effective, because you would have fewer participants.
Secondly, my research extends Provan & Sebastian’s model in the humanitarian relief
field. My research offers some evidence that similarly to the public health service
delivery sector, network effectiveness can be explained by intensive integration and
network cliques in the humanitarian relief field. My data supported the idea that
differences in effectiveness across networks could be better understood by focusing on
cliques and the overlap among cliques of multiple relationships among humanitarian
organizations. My study would help to do the clique analysis or to search for closely
connected and cohesive subgroups. Additionally, my work can help to design efficient
inter-organizational network structures in the humanitarian relief sector. For example, by
increasing the level of clique overlap (one dimensional or multidimensional) in inter-
organizational humanitarian networks, network designers should expect a higher level of
inter-organizational collaboration.
139
Thirdly, by empirically testing Provan & Sebastian (1998) conceptual framework for
assessing network effectiveness, my study contributes to further research in inter-
organizational collaboration in the humanitarian relief field. During my investigations, I
realized the importance of understanding the different type of relationships that exist
among humanitarian organizations. I found out that the relationships were significantly
complex especially when considering their motives. As mentioned earlier, disaster
response often involves heterogeneous organizations with a wide range of different goal
and need which render collaboration very challenging. In my study for example, when
asked about their reasons for getting into a relationship, my study subjects provided a
wide range of different reasons. Network designers need to examine more closely the
nature of relationships in which humanitarian organizations are engaged and the self-
reinforcing dynamic of overlapping groups.
Fourthly, my research also highlights the need to explore network effectiveness using a
set of different measures. The majority of existing work on network effectiveness,
including that of Provan and Sebastian (1998) was conducted using one measure. As
mentioned earlier, Provan & Sebastian used client outcomes, a subjective measure, to
assess network effectiveness. Moreover, in most cases, the effectiveness measure was not
selected with input from the various network members. In my study, I used input from
network members to determine the three measures of effectiveness. Using a set of three
different measures for network effectiveness allowed me to find consistent ranking
pattern for each of the six network structural characteristics studied. Moreover, my
findings suggest that the subjective and objective forms of network effectiveness are
better explained by different network structural attributes. Whereas subjective network
effectiveness is better explained by the number of cliques and clique membership,
objective network effectiveness is better explained by the multifaceted nature of inter-
organizational relationships as measured by clique overlap and multiplexity. My study
serves as an example of effectiveness being measured with multiple criteria. In a nut
shell, my work extends in the humanitarian relief field, Provan & Sebastian (1998)’s
model of inter-organizational network effectiveness.
140
Lastly, my research also has implications for network theories. For many organization
theorists, the study of both inter-organizational and intra-organizational networks has
primarily been an exercise in analysis and methods (Salancik, 1995). Building upon
Provan & Sebastian (1998), my study further develops an alternative method for network
analysis and contributes to building network theories by examining and explaining how
network structural properties including network density, cliques and overlapping cliques,
might promote the interests of network members and that of the community as a whole.
6.5 Ego-Net Characteristics and Effectiveness
I used the multiple linear regression method to investigate effectiveness at organizational
level. The independent variables, nine in total, were grouped into the following three
categories: organization, ego-network and network. These independents variables are
described in Table 18 below. I also used two interaction variables with the purpose to
assess the combined impact of technology and network characteristics on organizational
effectiveness. I conducted the regression analysis on two different measures of
organizational effectiveness, the dependent variable. The first measure was the level of
activities, measured as the number of funded projects and the second was the level of
collaboration, measured as the number of funding partners. In order to examine
separately the influence of each category of the independent variables on the dependent
variable, I developed four models as described below.
141
Variable level Variable Definition
Organization
Size
Size of the organization
Service Range of services provided by the organization
Com_Med Varieties of Communication media (e.g. Internet, Website)
Coll_SM Varieties of Collaboration social software (e.g. wiki, shared
db)
Cty_SM Varieties of Community social software (e.g. Facebook)
Ego-network
Centrality Degree centrality of the organization in the network
Bridge Structural hole value of the organization in the network
cliques Number of distinct cliques to which organization is a member
Network
Density Density of the network to which organization is a member
Interaction
Com_Med x Density Interaction between communication media and density
Com_Med x Centrality Interaction between communication media and centrality
Table 18: Organizational Effectiveness Variables
6.5.1 Models Building
As a first step in the model building process, I examined my data to check for consistency
and eventual errors of data due to data manipulation. This examination led to the
identification of one outlier in the data. Before proceeding to the next step of the analysis
I removed the outlier.
The next step in the model building process was to compute some basic statistics and to
check the correlation between the variables. Table 19 reports the descriptive statistics and
correlations between the variables. One preliminary observation was that all the
correlations between the independent variables and the dependent variables were
positive. This was an indication that organizations that have higher number on these
variables would tend to display higher level of effectiveness.
Hypothesis HO#2 concerned bridging structural hole in a network. My proposition was
that organizations will enhance their effectiveness by bridging structural holes both in
terms of level activities as well as the level of collaboration. When using the level of
activities as effectiveness measure, I found that bridging structural hole was an important
predictor of effectiveness only in Model IVb (β = -0.207; p < .05). This variable showed
no significance in in Model IVa. Bridging structural hole was found to be an important
predictor of effectiveness when using the level of collaboration as effectiveness measure.
I obtained the following statistics: Model IVa (β = -0.352; p < .005); Model IVb (β = -
0.433; p < .005). But contrary to my proposition, my findings rather showed a negative
relationship between bridging structural hole and effectiveness. This result is not an
isolated case. For instance, while several previous studies have shown that organizations
improve their performance as a result of bridging structural holes (e.g., Hargadon &
153
Sutton, 1997; Finlay & Coverdill, 2000), other studies have shown negative performance
effects of bridging structural holes (e.g., Ahuja, 2000; Dyer & Nobeoka, 2000). This
result may be due to the high heterogeneity of the humanitarian organizations concerned
in this study.
Hypothesis HO#3
My Hypothesis HO#3 was related to the importance of the density of the network in
predicting network effectiveness. I proposed that organization will benefit from high-
density networks to enhance their effectiveness both in terms of level activities as well as
the level of collaboration. When using the level of activities as effectiveness measure, I
found that network density was an important determinant of effectiveness. Both Model
IVa (β = 0.196; p < .05) and Model IVb (β = -0.403; p < .05) yielded statistically
significant evidence that supported this hypothesis. The negative sign on the β coefficient
in Model IVb results from adding the interaction term (communication media X network
density) in the model. This hypothesis was somewhat supported in Model IVa (β =
0.120; p < .1) when using the level of collaboration as effectiveness measure. The main
effect of network density was not statistically significant in Model IVb.
Hypothesis HO#4
My proposition in Hypothesis HO#4 was that the effectiveness of an organization will
increase both in terms of level activities as well as the level of collaboration, with the
number of the distinct cliques to which it belongs. This hypothesis was not supported by
my findings. In both measures of effectiveness the number of cliques was found to be an
important predictor of effectiveness, but contrary to my proposition, the final models
rather showed a negative relationship. When using the level of activities as effectiveness
measure, I obtained the following statistics: Model IVa (β = -0.414; p < .05); Model IVb
(β = -0.303; p = .119). When using the level of collaboration as effectiveness measure,
the statistics were as follow: Model IVa (β = -1.335; p < .005); Model IVb (β = -0.1307;
p < .005).
154
Hypothesis HO#5
I argued in Hypothesis HO#5 that the size of an organization will be positively associated
with its effectiveness both in terms of level activities as well as the level of collaboration.
I found no support to this hypothesis. When using the level of activities as effectiveness
measure, I obtained the following statistics: Model IVa (β = 0.089; p = .395); Model IVb
(β = 0.059; p = .579). When using the level of collaboration as effectiveness measure,
the statistics were the following: Model IVa (β = 0.101; p = .249); Model IVb (β =
0.091; p = .291).
Hypothesis HO#6
In Hypothesis HO#6, I proposed that the range of service provided by an organization
will be positively associated with its effectiveness both in terms of level activities as well
as the level of collaboration. This hypothesis was also not supported in any of the models.
The range of service provided by an organization was not found to be an important
deterrent of organizational effectiveness. When using the level of activities as
effectiveness measure, I obtained the following statistics: Model IVa (β = 0.011; p =
.913); Model IVb (β = 0.063; p = .540). When using the level of collaboration as
effectiveness measure, the statistics were the following: Model IVa (β = -0.063; p =
.442); Model IVb (β = -0.040; p = .622). This result may suggest that in inter-
organizational network for humanitarian information management, organizational
effectiveness measured both in terms of level activities as well as the level of
collaboration is driven by other factors regardless of the number of services an
organization provides.
Hypothesis HO#7
Hypothesis HO#7 was related to my proposition that the greater the variety of
communication media available in an organization the higher its effectiveness both in
terms of level activities as well as the level of collaboration. When using the level of
activities as effectiveness measure, Model I yielded statistically significant evidence that
supported this hypothesis (β = 0.418; p < .05). Communication media was also found to
significantly contribute to explain effectiveness in one of the full models (Model IVb).
155
When using the level of collaboration as effectiveness measure, none of the models
produced statistically significant evidence and thus failed to support the hypothesis.
Hypothesis HO#8
I proposed in hypothesis HO#8 that the greater the variety of collaboration social
software available in an organization the higher its effectiveness both in terms of level
activities as well as the level of collaboration. When using the level of activities as
effectiveness measure, none of the models produced statistically significant evidence and
thus failed to support the hypothesis. This hypothesis was supported when using the level
of collaboration as effectiveness measure. The full models showed that the range of
collaboration social software available in an organization was an important predictor for
effectiveness. I obtained the following statistics: Model IVa, (β = -0.274; p < .05);
Model IVb, (β = -0.277; p < .05). But contrary to my proposition, the final model rather
showed a negative relationship between the range of collaboration social software and
effectiveness.
Hypothesis HO#9
I hypothesize in HO#9 that the greater the variety of community social software available
in an organization the higher its effectiveness both in terms of level activities as well as
the level of collaboration. When using the level of activities as effectiveness measure,
similarly to the previous hypothesis, none of the models produced statistically significant
evidence and thus failed to support the hypothesis. When using the level of collaboration
as effectiveness measure, I found support to the hypothesis. The full models yielded
statistically significant evidence that the range of community social software was an
important predictor of organizational effectiveness (Model IVa, β = 0.271; p < .05; and
Model IVb, β = 0.237; p < .05).
156
6.5.2.2 Information Technology Interaction Effects
Hypothesis HO#10
In hypothesis HO10, I proposed that organizations that possess a wide variety of
communication media will benefit more from high network degree centrality to enhance
their effectiveness than those that do not. I found support to this hypothesis for both
measures of organizational effectiveness. When using the level of activities as
effectiveness measure, I obtained the following statistics: Model IV (β = 1.736; p <
.005). When using the level of collaboration as effectiveness measure, the statistics were
the following: Model IV (β = 0.667; p < .05).
Hypothesis HO#11
In hypothesis HO11, I proposed that Organizations that possess wide varieties of
communication media will benefit more from high network density to enhance their
effectiveness than those that do not. I found support to this hypothesis for both measures
of organizational effectiveness. When using the level of activities as effectiveness
measure, I obtained the following statistics: Model IV (β = 1.013; p < .005). When
using the level of collaboration as effectiveness measure, the statistics were the
following: Model IV (β = 0.465; p < .05).
In Table 22 below, I summarize the results of the hypotheses testing for both measures of
effectiveness. For each hypothesis, I indicate (with ‘S’) whether the independent variable
in the hypothesis was found to be an important predictor of effectiveness. I also indicate
if the hypothesis was support (with ‘SS’).
157
Number Hypothesis Significant & Supported
Level of Activities
Level of Collaboration
HO# 1 Greater centrality increases organization effectiveness SS SS
HO# 2 Greater bridging of structural holes increases effectiveness S
HO# 3 Organization effectiveness increases with the density of the network it which it belongs
SS SS
HO# 4 Organization effectiveness increases with the number of distinct cliques to which it belongs
S S
HO# 5 The size of an organization is positively associated with its effectiveness
HO# 6 The range of service provided by an organization is positively associated with its effectiveness
HO# 7 The greater the varieties of communication media available in an organization, the higher its effectiveness
S
HO# 8 The greater the varieties of collaboration social software available in an organization, the higher its effectiveness
S
HO# 9 The greater the varieties of community social software available in an organization, the higher its effectiveness
SS
HO# 10 Organizations that possess wide varieties of communication media will benefit more from high network degree centrality to enhance their effectiveness than those that do not.
SS SS
HO# 11 Organizations that possess wide varieties of communication media will benefit more from high network density to enhance their effectiveness than those that do not.
SS SS
Table 22: Summary of Hypotheses Testing at Organizational Level of Analysis
6.5.3 Discussion
I begin this discussion section by restating that previous studies that used the theoretical
lenses of Resource Based View primarily focused on characteristics internal to the
organizations to predict effectiveness and performance. Most of these studies were
conducted in the for-profit sector, conceptualizing organizations as atomistic profit-
seeking entities (Arya & Lin, 2007). Subsequent research on organization performance
and effectiveness highlighted the importance to view organizations as embedded in a web
of inter-organizational relationships which may serve both as resources themselves and as
Portes, 1998; Gulati et al., 2000; Shipilov, 2006). On the other hand, most studies that
apply social network approach to explore effectiveness focused on the characteristics of
158
network structure, without paying much attention to the attributes of the organizations
that comprise the network. In my study, I draw on both the RBV and the social network
theories to investigate how organizations’ attributes combined with network structural
characteristics influence organizational effectiveness. I especially explore the influence of
information technology on effectiveness. I discuss below the findings of my
investigations with regards to (i) the measures of organizational effectiveness (ii) the
relationship between organization internal characteristics and organizational
effectiveness (ii) the relationship between ego-network characteristics and
organizational effectiveness (iii) the relationship between network structural
characteristics and organizational effectiveness and (iv) the Catalytic role of Information
Technology on organizational effectiveness.
Measure of organizational effectiveness
The findings from my investigations suggest that in networks of organizations engaged in
humanitarian information management, organizational effectiveness would be better
assess using the level of collaboration. When using the level of collaboration as
dependent variable in a regression model, the linear combination of the independent
variables explained almost 76% of the variances. This proportion was less than 64%
when effectiveness was measured as the level of activities. Figures 36 and 37 below
depict these variations. Figure 36 presents the variation for the case where the interaction
term is combination of communication media and degree centrality; while for Figure 37
the interaction term is communication media and network density.
159
Figure 36: Variations in the Effectiveness Measures (Model VIa)
Figure 37: Variations in the Effectiveness Measures (Model VIb)
Organization internal characteristics and organizational effectiveness
My research also showed the importance of considering the characteristics internal to
organizations when explaining effectiveness. As discussed in the analysis section, when
using only the organizational internal resources as independent variables to predict
effectiveness, the regression model showed that the linear combination of these variables
was significantly related to effectiveness. Taken alone, organizations internal
characteristics explained only approximately 5% of the variances in organizational
effectiveness when using the level of collaboration but this percentage was much higher
160
(over 13%) when using the level of activities. This finding was consistent with previous
studies that used the Resource Based View to assess organizational effectiveness (e.g.
Zaheer & Bell, 2005; Arya & Lin, 2007).
Moreover, the findings of my investigation suggest that among humanitarian
organizations engaged in information management and exchange, information technology
would be one of the most important internal characteristics that would more accurately
predict effectiveness. Previous research has shown an increase in the adoption and use of
information technology in general, for disaster relief among humanitarian organizations
(Comfort, 1993; Quarantelli, 1997). For these organizations, information technology
plays a vital role in disasters relief. Research has also shown that the use information
technology may have a positive impact on inter-organizational collaboration and
coordination (Malone & Crowston, 1994). Studies have also highlighted the importance
of the use of social software in humanitarian disaster relief and crisis management (Palen
et al., 2007a; Palen et al., 2007b; Palen et al., 2007c; Sutton et al., 2008; Vieweg et al.,
2008; Hughes et al., 2008; Lui et al., 2008). Although most of these studies investigated
the use of social software at the individual user level of analysis, the positive impact of
these tools for disaster relief could easily be extrapolated at other levels of analysis
including the organizational level and the network level.
In my study, all the three information technology related variables were in some ways
found to significantly contribute to explain organizational effectiveness. However, not all
these information technology related variables were found to be positively related to
organizational effectiveness as I hypothesized. For instance, while wide varieties of
community social software were found to be positively associated with effectiveness, my
findings rather suggest a negative relationship between collaboration social software (e.g.
wiki, shared database) and effectiveness.
These contrasting results obtained from my statistical analysis concerning the importance
of information technology to humanitarian organizations were somewhat similar to those
obtained from the qualitative data gathered through interviews. As mentioned earlier,
161
interview participants were asked to give their opinion specifically on the implications of
information technologies on inter-organizational collaboration among members of the
Global Symposium and the contribution of these technologies in helping to meet the
organizational goal. Approximately seventy percent (68.42%) of the interviewees shared
their opinion on this issue. I registered a wide range of diverse point of views. Some
participants, roughly thirty one percent (30.77%) of those that answered the question, had
a very positive opinion about the implications and especially the catalytic role that
information technologies play in fostering humanitarian inter-organizational
collaboration. The vast majority (69.23%) however, expressed mixed feelings.
For the participants that had positive opinions, information technologies served as an
important catalyst for inter-organizational collaboration in the Global Symposium
community. They argued that if without information technologies effective simple
communication is difficult, collaboration would be even harder. Participant number five
for example reported that:
Subject#5: I think that information technology is extremely important because we basically need to communicate to all these different communities in as many different ways as possible.
They also believed that the use of information technologies is instrumental in quickly
gather analyze and disseminate humanitarian information leading to effective disaster
response. Below, we illustrate this point of view with quotes from three participants,
number six, seven and eleven.
Subject#6: You cannot do it without information technology. Gathering information, managing information, analyzing information, distributing information, really you cannot do all this without information technology. So I think the question is kind of obvious. Subject#7: Information technology essentially supports what we do. It helps in sharing information, mainly transporting information around, maintaining our communication. Subject#11: I think the information technology is key of cause, because without proper systems in place, you will not be able to do that.
Most of these participants who had a very positive opinion about the important
implications of information technologies in fostering collaboration among humanitarian
162
organizations also believed that information technologies were instrumental for their
organizations in meeting their goals.
The participants who expressed mixed feelings about the role of information technologies
in inter-organizational collaboration gave a number of reasons that could be grouped into
two main categories. The first category of reasons was related to the information
technologies infrastructure. Participants argued that more often, organizations in the field
do not have the necessary technology tools either because they were destroyed by the
disaster or because they did not even exist in the first place. They also talked about the
discrepancy in term of infrastructure between organizations based in developed countries
and those in the developing countries. They argued that people in developed countries
often enjoy latest technologies but the realty in developing countries, scenes of most
humanitarian disasters is quite different. Participant number twelve for example reported
that:
Subject#12: when you get out on the fields you see that the most basic important tool is paper map and a pencil. And I think we have got to really recognize that fact. […]You know we do this information technology that we love where they follow the latest systems and the fastest processor and stuff like that and we really like to paddle ourselves on the back on what we are able to do here in Washington DC. And then you get out on the fields and everyone is using paper maps and a pencil.
Finally they talked about the fast pace of change in technology which makes it difficult
for organizations to have and especially keep the technical staff that possesses adequate
knowledge to make use of these new technologies.
Subject#6: as the technology changes, it is hard to find the people that have skills that are up to date.
The second category of reasons concerned the management of information. Participants
believed that without proper standard for humanitarian information exchange, the
technology will be of no effective use.
Subject#5: I think yes, continue to explore all the new technologies that are available but at the same time realize that in the end what it really comes down to is quality information and information that is based on facts and that’s credible but people actually belief in. So I think we should not be allowed to be measured by technology if the content is not there.
163
Subject#14: One is developing some basic standards, and some basic platforms for information exchange.
They also believed that the humanitarian field needs better processes and well trained
staff in order to make good use of the technology.
Subject#8: I think there are certain organizations who think that technology can solve all the problems, so they don’t have a proper appreciation and understanding of the information management challenges and obstacles, but at the same time there is probably some information, people who are very skeptical about technology and do not sort of realize the value that it has.
Ego-network characteristics and organizational effectiveness
My investigations highlighted the significant impact that ego-network characteristics
have on organizational effectiveness. I explored how humanitarian organizations would
benefit from better ego-network characteristics (e.g. degree centrality, and bridging
structural holes, numbers of cliques). When using the level of collaboration as
effectiveness measure, ego-network related variables accounted for approximately 67%
of the explanatory power of the full regression model. This proportion was about 32%
when using the level of activities as effectiveness measure. These findings corroborated
with the view of organizations as embeddedneed in a web of relationships that provide
opportunities and values (Gulati, 1999). Network characteristics can be understood as
external resources embedded in organizations’ networks. According to Gulati et al.,
(2000), the embeddedness of organizations in networks holds significant implications for
organization performance.
The degree centrality was found to be the most important predictor of effectiveness. This
variable was consistently found cross models and cross effectiveness measures to be
significantly and positively related to organizational effectiveness. As mentioned earlier,
this finding is consistent with most previous research that explored the influence of
network position and especially the degree centrality on outcome such as performance
with those of Provan & Sebastian my findings suggest that at the network level of
analysis, an inter-organizational network in the field of humanitarian relief is more
effective when it is more integrated at the subnet level (clique) and displays higher level
of multiplexity. My study however makes one significant addition to Provan & Sebastian
model. Unlike Provan & Sebastian, in my study, I used three different measures of
network effectiveness (one subjective and two objectives). Using these effectiveness
measures allowed me to find consistent ranking pattern for each of the six network
structural characteristics used in my work. It is important to note that Provan &
Sebastian’s study which forms the foundation of my study, matched two out of the six
network structural characteristics. This study found a match in ranking only among
multiplexity and identical clique overlap and effectiveness.
Moreover, my findings suggest that the subjective and objective forms of network
effectiveness are better explained by different network structural attributes. Whereas
subjective network effectiveness is better explained by the number of cliques and clique
membership, objective network effectiveness is better explained by the multifaceted
nature of inter-organizational relationships as measured by clique overlap and
multiplexity. These findings highlight the importance of multiple criteria for assessing
network effectiveness. Finally, comparing the three measures of effectiveness that I used
in my study, my findings suggest that the level of activities is the best. This measure
matched three out of the six network structural characteristics investigated.
At the network level, the findings of my investigations could be summarized as follow:
Finding #1: In inter-organizational humanitarian information management networks,
network effectiveness will be better explained by network structural characteristics when
assessed at subnet levels.
Finding #2: In inter-organizational humanitarian information management networks, the
level of effectiveness will likely be higher in networks that are more dense and cohesive
at subnet levels.
170
Finding #3: Given the high heterogeneity and differentiation among humanitarian
organizations, network effectiveness in humanitarian information management, will more
accurately be explained by multiplexity and clique overlap.
RQ#2: Relationship between organization internal characteristics, network structural
properties and organizational effectiveness.
In addition to investigating effectiveness at the network level, in this work, I also studied
organizational effectiveness. I took several steps to explore organizational level
effectiveness. I used two measures of effectiveness including the level of activities
measured as the number of funded project and the level of collaboration measured as the
number of funding partners. For each of these effectiveness measures, I built a set of
four consecutive multiple linear regression models. In the baseline model, I modeled
effectiveness as a function of the variables of the organization category. I then gradually
added variables from ego-network category (Model II), network category (Model III) and
two inter-action terms (Model IVa and Model IVb). Overall, my findings suggest that in
humanitarian inter-organizational networks, organizational effectiveness can be
accurately explained by a linear combination of organizational internal attributes and
network structural properties.
RQ#2a: Relationship between organization internal characteristics and especially
information technology and organizational effectiveness.
My research also showed the importance of considering the characteristics internal to
organizations when explaining effectiveness. Taken alone, organizational internal
characteristics accounted for over 13% of the variances in organizational effectiveness
when I used the level of activities as effectiveness measure. The regression model
showed that the linear combination of organizational internal characteristics was
significantly related to effectiveness for both effectiveness measures. Moreover, my
findings suggested that among humanitarian organizations engaged in information
management and exchange, information technology would be one of the most important
determinants of effectiveness. In my study, all the three information technology related
variables that I used, were found to significantly contribute to explain organizational
171
effectiveness. However, not all these variables were found to be positively related to
organizational effectiveness as I hypothesized. For instance, while the availability of a
wide variety of community social software was found to be positively associated with
effectiveness, my findings rather suggest a negative relationship between collaboration
social software (e.g. wiki, shared database) and effectiveness. These contrasting results
obtained from my statistical analysis concerning the importance of information
technology to humanitarian organizations are somewhat similar to those obtained from
the qualitative data gathered through interviews.
RQ#2b: Relationship between ego-network characteristics and organizational
effectiveness.
Exploring the relationship between ego-network characteristics and organizational
effectiveness I also got some interesting results. My findings suggested that ego-network
characteristics have a significant impact on organizational effectiveness. Taken alone,
ego-network variables accounted for approximately 67% of the variances in
organizational effectiveness when I used the level of collaboration as effectiveness
measure. This proportion was about 32% when using the level of activities as
effectiveness measure. Among the ego-network variables, the degree centrality was
found to be the most important predictor of effectiveness. This variable was consistently
found cross models and cross effectiveness measures to be significantly and positively
related to organizational effectiveness. Bridging network structural hole was another ego-
network category variable found to be an important determinant of organizational
effectiveness. However, contrary to my proposition that organizations will enhance their
effectiveness by bridging structural my findings rather suggested a negative relationship
between bridging structural hole and effectiveness. This result may be due to the high
heterogeneity of the organizations that I investigated. In the humanitarian relief field and
especially in humanitarian information management and exchange, maintaining non
redundant may be very costly to humanitarian organizations. The number of cliques
another ego-network category variable explored was also found to be significantly and
negatively related to organizational effectiveness measured both as the level of
collaboration as well as the level of collaboration. This finding was one of the most
172
surprising of my investigations. In my knowledge, no previous study had explored the
relationship between the number of cliques and organizational effectiveness. However,
given some previous research on effectiveness conducted at network level (e.g. Provan &
Sebastian, 1998) and theoretical reasoning, this was an unexpected result. One possible
reason for this finding may the fact that there may have been a high level of overlap in
the cliques. Using distinct cliques may have probably yielded more meaningful results.
RQ#2c: Relationship between network characteristics and organizational effectiveness.
Investigating the relationship between network structural characteristics and
organizational effectiveness was one of the peculiarities of this study. None of the two
previous similar studies (Zaheer & Bell, 2005; Arya & Lin, 2007) had examined the
impact of network category variable on organizational effectiveness. The density of the
network, the only network category variable that was in the study was found to be an
important predictor of effectiveness. Taken alone, network density accounted for
approximately 3.2% of the variances in organizational effectiveness when I used the level
of activities as effectiveness measure. For both measures of effectiveness, my findings
suggested that high-density networks benefited more to organizations than low density
networks.
RQ#2d: Impact of inter-action between information technology and network structural
characteristics on organizational effectiveness.
Examining the impact on organizational effectiveness of the inter-action between
information technology and network structural characteristics in humanitarian
information management networks was another important peculiarity of my research.
My findings suggested that organizations that possess a wide variety of communication
media (e.g. internet - available to the majority of staff-, website – regularly updated-,
blogs, etc…) will benefit more from high network degree centrality to enhance their
effectiveness than those that do not. These organizations will benefit more from high
network density than those that do not possess these technologies.
173
At the organizational level of analysis, the findings of my investigations could be
summarized as follows:
Finding #1: In inter-organizational networks for humanitarian information
management, centrally located organizations will more likely display higher level
of effectiveness than those situated at the peripheral.
Finding #2: Organizations that are member of dense and cohesive humanitarian
information management networks will more likely display higher level of
effectiveness than loosely connected networks.
Finding #3: In humanitarian information management networks, organizations
that possess wide varieties of communication media will benefit more from high
network degree centrality to enhance their effectiveness than those that do not.
Finding #4: Other things being equal, in humanitarian information management
networks, organizations that possess wide varieties of communication media will
benefit more from high network density to enhance their effectiveness than those
that do not.
Summing up, my investigations confirmed the proposition that organizational
effectiveness is affected by different organizational and network attributes in
humanitarian information management networks. More broadly, my findings on the one
hand pointed to a need in inter-organizational social network studies to go beyond a
structuralist view and take into consideration the characteristics of individual
organizations, as predicted by the RBV, in assessing effectiveness. On the other hand, my
study highlighted the fact that organizational level network studies that tend to overlook
resources internal to organizations may be missing a large source of variance in
effectiveness. Finally my study highlighted the important role of communication media
in organizational effectiveness among humanitarian organizations.
174
7.4 Contributions
Network Effectiveness
My research extents to the humanitarian relief field, Milward & Provan’s (1998)
framework for evaluating public-sector organizational networks. My research contributes
to the literature on network effectiveness in a number of ways. First, my findings
confirmed some of the results of previous research and especially those of Provan &
Sebastian (1998) which showed that most effective networks are those that are integrated
at clique level. Specifically, my findings confirmed the importance of network structural
characteristics such as integration and cohesion to network effectiveness measure.
Moreover, building on Provan & Sebastian (1998), my study further highlighted the need
to consider network effectiveness analyses in smaller substructures instead the whole
network as has usually been the case.
Secondly, my research highlighted the need to explore network effectiveness using a set
of different measures. The majority of existing work on network effectiveness, including
that of Provan and Sebastian (1998) was conducted using one measure. Moreover, in
most cases, the effectiveness measure was not selected with input from the various
network members. In my study, I used input from network members to determine the
three measures of effectiveness. Using a set of three different measures for network
effectiveness allowed me to find consistent ranking pattern for each of the six network
structural characteristics. Moreover, my findings suggested that the subjective and
objective forms of network effectiveness are better explained by different network
structural attributes. Whereas subjective network effectiveness is better explained by the
number of cliques and clique membership, objective network effectiveness is better
explained by the multifaceted nature of inter-organizational relationships as measured by
clique overlap and multiplexity. My study serves as an example of effectiveness being
measured with multiple criteria. In a nut shell, my work builds on various models of
effectiveness already present within the literature on inter-organizational effectiveness to
provide a multidimensional model for evaluation effectiveness in the nonprofit
humanitarian field.
175
Lastly, my research also has implications for social network theories. For many
organization theorists, the study of both inter-organizational and intra-organizational
networks has primarily been an exercise in analysis and methods (Salancik, 1995).
Building upon Provan & Sebastian (1998), my study further develops an alternative
method for network analysis and contributes to building network theories by examining
and explaining how network structural properties including network density, cliques and
overlapping cliques, might promote the interests of network members and that of the
community as a whole.
Characteristics of successful inter-organizational networks
My investigations have helped to identify the following four main characteristics that
seem to be common among organization members of successful networks. These
characteristics include (i) their ability to share, (ii) their ability to contribute, (iii) their
commitment to networking and (iv) the level of their embeddedness through multiplex
ties in the network.
1. Sharing spirit: Organization members of a network must “dare to share” (ICCO
2004). They need to be open, willing and able to learn from each other. In my research,
the lack of sharing spirit was consistently reported as one of the biggest problems that
undermines network effectiveness.
Subject#13: I think the main challenge here is that the idea of sharing formation has always been said in many areas. It is usually always said yeah it is good to share but you do not sometime see concrete platforms or formalities on how to share this information. It is not formalize. It is always thought as an objective but never formalize.
Network members must feel confident enough about what they do and the information
they possess that they are willing to share with others. There must therefore exist an
atmosphere of openness among members and potential members which allows them to
admit mistakes and to learn from them. Networks cannot flourish without this trust. A
network can help to develop sharing spirit among its members by creating an open
environment in which people are willing to analyze and learn from both their successes
176
and their mistakes. Networks and partnerships are more likely to become effective when
they are founded by members that share a history of working together, that know each
other and have relationships characterized by mutual trust. This suggests that networks
may have a longer incubation and startup period before they can reach the stage of
maximum effectiveness.
2. Capacity to contribute: Organization members of a network must have the capacity
to contribute especially in terms of skills, access and time/money available. In my study,
the ability of a potential network partner to contribute was reported to be one of the main
collaboration factors.
Subject#2: Both [organizations] have to be able to bring to the table their competitive advantage. You can’t have two organizations that do the same thing. So you need different skills set from any of the organizations. Subject#7: We think about the quality of what that agency does and the quality of what that agency is known to do.
In order to foster inter-organizational coordination/collaboration there must be space for
learning, reflection and interaction. Also, it is paramount that senior leaders of
organization provide support to network by emphasizing the importance of networking.
They must also encourage the involvement of staff in the activities of the network.
Moreover, all network members must have equal access to any technology that the
network uses so that certain groups are not marginalized.
3. Commitment: Organization members of a network must be committed to the
networking activities. They must consider the priorities of the network their own. They
must also be motivated by self-interest because networking is a potential added-value to
their daily work. Commitment will be strong if members see the network as adding value
to their work, and if the priorities of the network match their own. According to ICCO
(2004) incentive grants are of little value in enticing members. I agree with this author in
his contention that funding should not be the reason that a NGO joins a network. For
instance, he suggests that a golden rule for success may be to let a network start from its
own resources with the idea that initial self-reliance builds commitment (ICCO 2004).
The author also mentions, however, that this does not mean networks do not need funding
177
for the activities they would like to undertake. Networks need funding for example to
help support a facilitator, coordinator, or staff of some sort that is able to spend the time
required to nurture relationships and in order to keep the group together. It is important
that careful attention is given to these aspects when funding is initially proposed.
Another important condition is that the initiators of networks are enough committed to
overcome the organizational and establishment phase, which takes a lot of effort, while
often working for not immediately seen results with little money.
4. Multiplexity: Organization members of a network must strive to keep multiple type of
connection with other members. Multiplexity can be measured at the individual network
member level and at the level of the whole network. A high degree of multiplexity of a
member indicates high embeddedness of the member in a network and signifies less
liability to disruption of single relationships. A member with a large number of multiplex
relations is expected to have a high potential of mobilizing different resources and
information through these relations. On the other hand, such a member is subject to a
high level of social control. At the network level, the degree of multiplexity specifies the
overlap between the different relation-specific networks. For evaluating network
effectiveness, multiplexity can be a particularly useful measure (Provan & Milward
2001). Effective networks might have a majority of network members connected through
two or more different types of relationships. In this case, multiplexity will be high,
reflecting commitments among network members to one another through multiple
activities.
Organizational Effectiveness
Concerning the literature on organizational effectiveness, my study illustrates the
importance of both internal organizational characteristics as well characteristics external
to organizations, for effectiveness. My findings confirm the extended Resource Based
View perspective of organizational effectiveness. More specifically, my analysis of the
relationships of the various determinants of effectiveness illustrated that variables from
all the three categories (organization, ego-network and whole network) are found to be
178
important predictors for organizational effectiveness. One of the most significant
contributions of my research to organizational effectiveness literature and especially to
the resource based view perspective concerns including whole network category variables
in assessing organizational effectiveness. None of the two previous similar studies
(Zaheer & Bell, 2005; Arya & Lin, 2007) had examined the impact of network category
variable on organizational effectiveness. By showing that organizations enhance their
effectiveness when they occupy a better network position, my research contributes to
demonstrate the value of including external resources, or the ability of an organization to
exploit a favorable network structural position (Gnyawali & Madhavan, 2001; Gulati,
1999). Another important contribution of my study is that it extends the Resource based
view perspective in the nonprofit sector and especially in the humanitarian relief field.
Most of the previous studies that draw on the RBV examine effectiveness in for-profit
network contexts. By applying the RBV to a collaborative nonprofit context as opposed
to a competitive for-profit context, my research shows that internal and external resources
allow some organizations to enhance their capabilities by collaborating with others.