EMPIRICAL ANALYSIS OF HYBRID OPEN SOURCE SOFTWARE MODEL:
DETERMINATION OF EFFICIENCY AND GOVERNANCE
Mohamed A. Sidahmed
B.S. (honors), University of Khartoum, 1998
M.S., Cleveland State University, 2001
A thesis submitted to the
University of Colorado Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Computer Science and Information Systems
2009 by Mohamed A. Sidahmed
All rights reserved.
This thesis for the Doctor of Philosophy
Mohamed A. Sidahmed
has been approved
7~ H' 2oof
Sidahmed, Mohamed A. (Ph.D., Computer Science and Information Systems)
Empirical Analysis of Hybrid Open Source Software Model: Determination of Efficiency and
Thesis directed by Professor James H. Gerlach
Over the past decade, open source software (OSS) projects underwent significant changes
and restructuring, symbolizing venture from the original approach of producing open
source code by volunteer contributors. The latest trend, which is fueled by commercial
organizations, systems integrators, and IT vendors' interest in OSS, involves the joint
collaboration between open and proprietary concerns. This study investigates the
implications of a collaborative hybrid OSS development model from the perspective of the
OSS project. By extending transaction cost economics and interorganizational cooperation
frameworks to the OSS domain, the research investigates factors affecting network
governance structure of this hybrid alliance. Theory predicts that efficient forms of
collaborative transaction leads to perceived satisfaction with an alliance. The study
examines factors that contribute to satisfactory institutional governance and underlying
principles that influence OSS projects to engage in hybrid relationships with commercial
partners. Moreover, the research identifies the necessary dimensions of interorganizational
cooperation and safeguards, which minimize projects vulnerability to detrimental behavior
by commercial partners. Data was collected across OSS projects to analyze the impact of
commercial partners involvement on project efficiency and governance. Results
demonstrate that OSS projects seek to establish arrangements that give rise to atypical
structure for efficient management of the development process. Results reveal that the
formation of institutional establishment based on streamlined information flow, flexible non-
formal relationship, collective collaboration responsibility, and moderation of influence
effect give rise to a favorable form for governing transactions. The study also found that
trust and branding play significant roles in reinforcing a thriving governance structure.
Findings support the notion that hybrid projects achieve gains in product distinctiveness,
yet collaboration is fragile to commercial parties behavior of seeking self-interest. The
emerging structure yields a perceived meritorious outcome for the OSS project; strongly
suggesting that transaction cost efficiencies are realized. Practical implications for the
study include identifying significant factors that contribute to OSS project efficiency and
optimal governance, in addition to, establishing brand identity as a mutual benefit that
binds the alliance. Hybrid OSS project success is largely dependent upon how well these
factors are managed. The research fills a gap in empirical analysis of both OSS hybrid
development model and OSS business model research.
This abstract accurately represents the content of the candidates thesis. I recommend its
James H. Gerlach
I dedicate this thesis to my parents and dearly loved grandparents, who taught me the
essences of life. Their teaching always enlightens my path. Also, I dedicate this
dissertation to my wife, thank you for the tremendous support and patience throughout the
I would like to express my gratitude to several people, who provided me with tremendous
help and support either directly or indirectly. I owe a great credit to my advisor and mentor
James Gerlach. His support extends beyond research, as one of his valuable life and a
career counsels was opportunities are made!
When it comes to research, Jims mentoring approach stimulates talented ideas and
always challenges one to go an extra step. I also, want to thank my committee members,
professors, Cliff Young, Ron Ramirez, Woody Eckard, and Tom Altman for their generous
time and efforts. It has been a privilege to have such diverse and supportive cohort of
scholars. Their collective intelligence and constructive feedback helped improving the
quality of this research.
My deepest gratitude to professor Adam Fadlalla. Adam is a life-long friend, mentor, and
an advisor. He is one of the remarkable teachers, who had big impact on my academic
career. His influence taught me that achievement, to a large extent, is a function of
articulated goals, persistence, and perseverance.
I want to extend my appreciation to professor Peter Bryant for his teaching, generosity and
encouragement. I will always miss our discussion sessions and extraordinary class
experience. My thanks go to professor Dawn Gregg for her collaboration and career
advice. It has always been a pleasure working together. Last but not least, I would like to
extend my appreciation to the many people, I did not explicitly mentioned, and without their
support this would not be achievable. I deeply appreciate it and always grateful.
TABLE OF CONTENTS
1. Introduction ...............................................................1
1.1 Study Framework.............................................................1
1.2 Statement of the Problem....................................................4
1.3 Thesis Statement............................................................6
1.4 Motivation of the Study.....................................................6
1.5 Thesis Organization.........................................................8
2. Open Source Software Overview...............................................9
2.1 Background .................................................................9
2.2 Motivation to Contribute...................................................11
2.3 Open Source Licenses.......................................................14
2.4 Projects ..................................................................16
2.5 Project Governance.........................................................18
2.6 Governance Challenges......................................................25
2.6.1 Commercialization of Open Source...........................................26
2.6.2 Institutions Involvement..................................................26
3. Theoretical Framework.......................................................29
3.1 Transaction Cost Economics Theory...........................................29
3.1.1 Governance Branch of TCE...................................................31
3.1.2 Semi Specific Integration..................................................33
3.1.3 Asset Specificity..........................................................34
3.1.4 Uncertainty ...............................................................35
3.1.5 Opportunism ...............................................................37
3.2 Interorganizational Cooperation Theory......................................38
3.3 Application of Theoretical Framework to Open Source.........................40
3.3.1 Branding ..................................................................43
3.3.2 Trust 44
3.3.3 Product Distinctiveness....................................................45
3.4 Competing Theories ...........................................................45
3.4.1 Contingency Theory..........................................................47
3.4.2 Resource-based View of the Firm............................................47
3.4.3 Agency Theory..............................................................48
3.4.4 Resource Dependency Theory.................................................49
3.4.5 Structuration Theory.......................................................49
3.4.6 Noncooperative Game Theory.................................................50
4. Hybrid Open Source Software Project Efficiency..............................53
4.1 Research Hypotheses.........................................................56
4.2 Pilot Study ................................................................60
4.3 Data and Method.............................................................61
4.3.1 Study Sample ........................................................61
4.3.2 Non-response Bias...................................................65
4.4 Measures ............................................................65
4.4.1 Dependent Variable...................................................65
4.4.2 Independent Variables...............................................66
4.5 Data Analysis and Results............................................69
4.5.1 Test of Research Model and Hypotheses................................69
4.5.2 Measurement Model...................................................70
4.5.3 Structural Model....................................................82
4.6 Discussion ..........................................................90
4.7 Conclusion ..........................................................93
5. Governance Mechanisms of Hybrid Open Source Projects.................94
5.1 Research Hypotheses..................................................97
5.2 Data and Method.....................................................109
5.2.1 Study Sample........................................................109
5.2.2 Non-response Bias .................................................112
5.3 Measures ...........................................................112
5.3.1 Dependent Variable..................................................112
5.3.2 Independent Variables..............................................113
5.3.3 Mediating Variables................................................114
5.4 Data Analysis and Results...........................................117
5.4.1 Test of Research Model..............................................117
5.4.2 Measurement Model..................................................122
5.4.3 Second Order Measurement Model.....................................124
5.5 Hypotheses Testing............................................................132
5.5.1 Structural Model..............................................................132
5.6 Discussion ...................................................................143
5.7 Conclusion ...................................................................145
5.8 Implications .................................................................145
6. Hybrid Open Source Software Projects Case Study...............................148
6.1 Xen Hypervisor, Open Source Virtualization Platform...........................148
6.2 Pentaho Open Source Business Intelligence Suite...............................153
6.3 MySQL Open Source Relational Database Management System.......................157
6.4 xTuple Open Source ERP Software...............................................158
6.5 OpenOffice Productivity Suite.................................................159
6.6 Manifestation of Governance in Hybrid OSS Projects............................160
7. Conclusion ...................................................................165
7.1 Contributions................................................................ 165
7.2 Implications of Research......................................................166
7.2.1 Theoretical Implications....................................................166
7.2.2 Practice Implications........................................................167
7.3 Limitations ..................................................................168
7.4 Future Roadmap................................................................169
A. Survey Instrument and Measured Items..........................................170
B. Human Subjects Certificate of Exemption.......................................177
C. Informed Consent.............................................................179
D. Software Contributor Assignment Agreement....................................181
4.1 Efficiency theoretical research model.......................................55
4.2 Efficiency measurement model................................................75
4.3 Efficiency structural model.................................................83
5.1 Hybrid OSS governance theoretical model.....................................99
5.2 Governance model second-order confirmatory factor analysis.................125
5.3 Hybrid OSS governance measurement model....................................129
5.4 Hybrid OSS governance structural model.....................................133
6.1 Relationship arrangement between Pentaho company and the community........155
2.1 Open source software non-volunteer contributors...............................23
3.1 Summary of alternative theories.................................................52
4.1 Efficiency model constructs reliability and measures............................67
4.2 Efficiency measurement model factor score regression............................70
4.3 Efficiency measured variables mean, standard deviation, and correlation.........72
4.4 Efficiency measurement model fit indices........................................77
4.5 Efficiency measurement model squared multiple correlations......................79
4.6 Efficiency measurement model standardized regression weights....................80
4.7 Efficiency measurement model estimates..........................................81
4.8 Efficiency structural model estimates...........................................85
4.9 Efficiency structural model squared multiple correlations.......................86
4.10 Efficiency structural model standardized regression............................87
4.11 Efficiency structural model fit indices.........................................88
4.12 Efficiency model hypotheses testing summary....................................92
5.1 Governance model constructs reliability and measures..........................115
5.2 Governance model means, standard deviations, and correlations of observed
5.3 Hybrid governance model factor score weights..................................122
5.4 Governance model second-order CFA standardized regression weights......127
5.5 Governance model second-order CFA squared multiple correlations........128
5.6 Hybrid OSS governance measurement model goodness of fit indices........130
5.7 Hybrid governance structural model goodness of fit indices.............135
5.8 Governance structural model regression weights.........................137
5.9 Governance structural model squared multiple correlations..............139
5.10 Governance structural model standardized regression....................141
5.11 Results summary of governance model hypotheses testing.................147
6.1 Efficiency model factors functional aspects............................161
6.2 Governance model factors functional aspects............................163
1.1 Study Framework
Accelerated technological advancements and continuous demands for new solutions to
fulfill users and businesses' needs impose extra challenges on the traditional model of
software and applications development. The constrained closed software development
approach, often ascribed as rigid structure, is criticized for failing to exploit open innovation
artifacts beyond institutional boundaries. In general the proprietary scheme is increasingly
imposing supplementary licensing structure and additional fees to match overhead
structuring costs and associated development expenditures.
The open source software (OSS) development model emerged as a contender to closed
proprietary software development controlled by commercial vendors. The OSS approach is
characterized as a transparent decentralized volunteer-based collaboration for developing
software products that fulfill both authors personal needs and made available for free or a
nominal fee to a wide user base. The growth of OSS as a viable low-cost alternative to
proprietary software total cost of ownership (TOO) and licensing fees, appealed to a broad
capacity of commercial organizations, systems integrators, and IT vendors. Yet, full
adoption of open source software by individuals and particularly within the enterprise is
undermined by several risk factors. According to Dixon (2007), some of these impediments
Unclear or imprecise roadmap
Lack of formal support and services
Projects rate of change
Lack of endorsement by independent software vendors
Software license types
Recognizing the fact that risk-related elements sometimes supersede cost attributes and in
seeking to overcome some of the above obstacles, a growing number of volunteer
community-based projects attracted attention and direct involvement of large corporations
that recognized the value of either supporting the open source development approach, or
embracing OSS business models.
A well-known example of collaboration between commercial organization and open source
community started more than a decade ago by Netscape Communications endeavors to
create network externalities via open source code of their flagship browser, Netscape
Communicator, and initiated efforts for building a sound community around the project. A
renowned pattern of established collaboration between OSS and commercial partners is
revealed by the level of investment in OSS projects. Currently, several different-size
organizations and giant IT vendors such as Intel, Google, Oracle, and Sun Microsystems
are participating and endorsing multiple open source projects. A variety of support
measures include: contributing physical assets and manpower resources, releasing
patents and proprietary code, and announcing platform certification and endorsement of
open source products. In addition, monetary donations are also among common forms of
endorsement. The significant investment and collaboration between the two parties mark
an emerging pattern of software development model that leverages organizations
economies of scale and communitys pool of talent and resources. Research indicated that
joint collaboration between open source software project community and commercial
partners establishes synergy that contributes positively to OSS ecosystems (Capek, Frank,
Gerdt, & Shields, 2005).
The open source software community is a heterogeneous group of individuals, which
include members with diverse backgrounds and domains of expertise involved in the
process of creating and maintaining OSS production model. The dimension of cooperation
and contribution of commercial organizations and IT vendors in open source projects is
regarded as an effort to augment value and resource utilization (Clemons & Row, 1992).
Also, the commercial partners involvement is regarded as an attempt to support
commoditized OSS products that complement organizations proprietary products (Deek &
McHugh, 2008). An example of explicit coordination and contribution made by these
organizations to one of the prominent open source software projects, Linux operating
system kernel, shows that commercial and proprietary vendors strategy recognized the
significance of open source software development approach.
In opposition to what is commonly believed, the majority of contributions to successful and
well-established open source software projects are made by commercial partners and IT
vendors paid contributors. The case studies chapter of the dissertation (Chapter 6)
discloses details about volunteer vs. commercial companies paid employees contribution
A recent study investigating major contributors to the Linux kernel project found that more
than twenty nine organizations and IT vendors are making significant contributions to the
development of the open source project (Kroah-Hartman, Corbet, & McPherson, 2008).
This revealing information evidently substantiate the fact that open source paradigm is
undertaking remarkable amendment to the conventional pure volunteer-based model. The
OSS community is welcoming collaboration with IT vendors and commercial partners, who
contribute considerable amount of code development and perform mundane maintenance,
support, and documentation tasks through paid employees. This study investigates
outcomes of established relationships between OSS projects and commercial partners.
Specifically, the research assesses cooperation efficiency and optimal ways for governing
the relationship between an open source software project and commercial partners.
1.2 Statement of the Problem
The study embarks upon the application of transaction cost economics to open source
software (OSS) development organizations. Latest trend in OSS witnessed significant
involvement of commercial IT vendors and proprietary software development shops by
embracing the bazaar1. Until recently, commercial vendors considered open source
software as a low-profile hobbyist activity with insignificant impact on their product market
share. As OSS continues to develop into a mainstream IT and become part of the
enterprise domain mix, commercial IT vendors shift strategy to accept the movement.
The value of collaboration and IT vendors sponsorship of OSS projects requires rigorous
research to determine the prospect of such alliance. As parties of the alliance hold
dissimilar and sometimes contradictory goals and missions, it becomes important to take
into consideration special characteristics of non conventional transaction established
between the two parties. Open source software project perspective has been adopted in
this study to examine the outcome of the hybrid partnership. The study extends TCE
vertical integration decision, by adopting the principal-agent metaphor to OSS project and
commercial associate partnership. The partnership ascribes OSS project as the client
(principal) making the decision whether to adhere to the typical pure volunteer contribution
mode (vertical integration), or establishing a hybrid structure, through alliance with
commercial partners (agents). OSS project has the option of completely relying on the
open source standard mode mechanisms, fueled by community volunteers for developing
the product, or accepting the involvement of sponsoring and/or collaborating commercial
organizations. Economic rationale predicts that open source software projects will continue
to carry out those talent-oriented tasks that appeal to volunteers and leverage commercial
partners competences for other activities. The strategic alliance between the parties
positions itself to take advantage of opportunities arising in a sustained relationship. Such
values include advancement of product and service offerings and utilization of talent to
accommodate relationship revolution (Subramani, 2004).
The OSS community develops software as a public good. Unlike other proprietary vendors,
OSS volunteers are not concerned about disclosing innovation in product design or
algorithm development. Quite the opposite, the latest move by proprietary software
vendors calls for protecting claimed intellectual property even before identifying the
potential role in the final commercial product. This contradicting nature between the parties
of the relationship is unique in a sense. Although they share some incentives of traditional
inter-organization alliance, yet OSS project community and commercial partners represent
divergent forms of structures that are involved in the OSS hybrid model of collaboration.
Open source software development is regarded as a mode of governance structure that
differs from the classical profit maximization firm model. Development in TCE suggests that
hybrid models for alliance organizing are more likely to take place when each party realizes
the added value of the relationship. The study attempts to empirically validate the
assumption that forming a relationship between open source software project and
commercial partners will result in favorable outcomes for the project. The study embraces
the open source software project point of view to assess the causal structure of the hybrid
model efficiency and governance.
1.3 Thesis Statement
Structural changes in the open source software governance model and increased
involvement of commercial companies and proprietary IT vendors represent a departure
from the classical pure open source software model to acceptance of a hybrid
commercialization approach for organizing. The open source software project should
embrace commercial organizations collaboration and sponsorship, if the governed alliance
yields increased efficiency and recognized brand entity.
1.4 Motivation of the Study
Open source software development is a dynamic fuzzy organization structure that proved
to be effective in producing high quality software products by volunteer communities. Yet,
OSS development is experiencing mutation as a consequence of increasingly significant
involvement of commercial IT vendors and other proprietary software development
companies. It could be argued that OSS will gain momentum due to the combined efforts
and support from both volunteers and commercial partners, but at the cost of transforming
its bazaar process for development. Therefore, analyzing the consequences of this
adaptation on the project becomes one of the goals of the study.
Open sources extraordinary method of organization should give rise to a new collaborative
structure that engages competing resources. Given recognition of the special form of
collaboration between open source community and commercial partners, it is expected that
a new form of interorganizational cooperation would emerge to account for two-party
differences since governance is recognized as the fundamental function of control and
administration that takes place when a group of people come together to legally
incorporate under the laws of a state for a nonprofit organizational purpose (Gies, Ott, &
Shafritz, 1990, p. 178). In addition, it is likely that the role of transaction cost will presume
another level of importance and allow for alternative interpretations within the open source
context. From a transaction cost point of view, various forms of interorganizational
relationship, such as joint ventures or network structures, are considered alternative forms
of governance and departures from the generic organizational hierarchy (Barringer &
Harrison, 2000). Some examples of typical businesses relationships include: marketing
distribution partnerships, sales partnerships, and R&D partnerships.
This research is unique in a sense that emphasizes organizational interdependency that
involves the establishment of joint cooperative activities between open source project
community, a casual non-profit volunteer-based organization, and hierarchical-oriented
commercial partners (e.g. OSS companies, IT vendors). Moreover, as participants in the
hybrid alliance maintain potentially conflicting goals and missions, it is imperative to take
into consideration special characteristics of the transaction.
This research adopts a multidimensional view for examining the hybrid form of OSS
development. It applies the premise of transaction cost economics and interorganizational
collaboration theory of adopting cost-minimizing governance structure for open source
project community and commercial (for-profit) organization partner(s), designed for
software development transaction. Attempts to understand the implications of such
relationships is conducted on the basis that both parties agreed to form an out-of-band
association in the form of cooperative adaptation (Williamson, 2002), which is a departure
from the classical form of alliance, in order to achieve mutual benefits and cost savings.
While it is expected that both parties have relative dependency on the other, nevertheless
adoption of safeguard mechanisms and low perception of opportunistic behavior likely
enable launching successful institutional establishments. The research objective is to
identify a meritocratic governance structure for managing the hybrid partnership and
interorganizational cooperation. The study adopts open source software project level of
analysis and community perspective to examine cooperation patterns and antecedents of
efficiency and bilateral governance structure of hybrid OSS projects. The research
presumes novel vision to help understand open source software phenomenon. It
contributes to the OSS literature and leads the way for future research directions.
Moreover, results of the study are of interest for practice, by offering insights on factors
contributing to developing successful alliance between the open source software
community and commercial partners.
1.5 Thesis Organization
The dissertation is organized into seven chapters. Following the introductory chapter, the
remainder of the thesis structure is presented next. Chapter two provides an overview of
open source software in general and relevant issues to the focus of the study. Significant
background work in OSS is also presented in this chapter. Chapter three presents the
theoretical foundation and the underlying guiding theories adopted for the study.
Comparative theory assessment is incorporated in this chapter; a summary of alternative
competing theories and their relevancy for the study is offered. The fourth chapter
introduces the first empirical study. The efficiency of the hybrid OSS model is examined in
this chapter. Chapter five includes the second analytical model that analyzes dimensions of
governance and other influential factors for achieving meritorious relationship outcome.
The sixth chapter investigates practical implementations of the hybrid open source
software model across several projects. Finally, chapter seven presents the summary of
research findings, study limitations, and future research directions.
2. Open Source Software Overview
Traditionally, free or open source software (OSS) development is a distributed coordinated
process and highly successful innovative mode of producing free software by large
volunteer-base contributors (von Hippel & von Krogh, 2003; Osterloh & Rota, 2007). The
volunteer contributors are typically working without promise of direct monetary
reimbursement or indirect reward compensation. Therefore, open source software is
habitually ascribed as a voluntary nonprofit community organized around a vast number of
software projects supported by the community. Free Software Foundation1 (FSF)
distinguishes free software as being a matter of liberty, not price. The metaphor of "free
speech" rather than "free beer is commonly used to help understand the concept.
The OSS project is recognized as the coordination entity that receives significant support
from unpaid developers and adheres to the communitys values and norms of organizing.
The shared norms value system is fundamental in strengthening the community internal
relations and protecting against 'outsiders intrusion (Soderberg, 2007, p. 28). Unlike firm
or market coordination mechanisms, the OSS production process orchestrates composite
entity contribution efforts in an unconventional mode to produce functional products. The
small group of core developers is reportedly accountable for about 80% of source code
contributed (Crowston & Howison, 2005). Moving further away from the center of the OSS
community, a larger group of seasonal developers with supporting roles to the core
commonly exists. A much bigger set of users represent a third layer surrounding the
The produced source code is freely available for download, modification, and redistribution,
under a precise licensing scheme. One of the most widely used licensing schemes is GNU
General Public License (GPL). Contrary to proprietary software, where development is
executed in a closed-organization setting, OSS society adopts a globally distributed and
transparent process of developing software products that span across languages, cultures,
and geographic regions. The free and open model of collaboration is ascribed as a self-
organizing society that fosters rapid knowledge creation and innovation diffusion (von
Hippel & von Krogh, 2003; Lee & Cole, 2003).
The loosely structured open source model enables large scale collaboration and requires
an agile, incremental development approach. The internal classical network organization
structure and hierarchy is based on reputation, contacts, and demonstrated technical skills.
Yet, the project's dynamic hierarchy and unrestricted shift of power is not based on
economic, legal, or architectural dependencies (Soderberg, 2007).
The OSS paradigm positions itself as a revolutionary organization structure (Lerner &
Tirole, 2002). This form of organization is distinct from the classical hierarchical structure
with a chain of command and decision authority. Situated as a lean formation, open source
organization is configured to reflect communitys impulse and stimulate improvement and
restructuring for efficiency. The community is a superset of all developers, users, and other 1
supporters of OSS, who carry out fundamental activities of product design, implementation,
defects reporting, and bug fixing.
Deeply rooted in Hackerdom (AIMarzouq, Zheng, Rong, & Grover, 2005), the primary
concern of open source hackers (real programmers) culture is the assurance of public
access to the source code. Members demand lifting all restrictions on obtaining, modifying,
and freely redistributing source code and binaries. Adherence to this unwritten agreement
is what establishes membership in the open source software community.
Intrigued by the freedom to express themselves (such as the freedom of speech) through
the code, OSS developers elect for themselves what they want to work on and what fits
their interests and capabilities. As collaborative activities continue, more developers join
the efforts and eventually a social network structure emerges as a result.
Presently, there are two main campaigns, the Free Software Foundation lobby group and
the Open Source Initiative2, that steer OSS. The two merely differ on principles and agree
in terms of practicalities. They resemble political parties with different views and stands on
certain issues. Yet, the relationship between the two camps is more synergetic rather than
being acrimonious to each other. Both contest a common opponent represented by
proprietary software. The two camps reflect philosophical differences and endeavor to
maintain their separate identities. For the Free Software camp, this reflects their claim to
their original establishment of the community.
2.2 Motivation to Contribute
OSS developers motivation to contribute to open source software has been a rich topic for
research. Several studies investigate factors driving voluntary participants to contribute
valuable resources such as time and effort without being directly compensated (Lakhani &
von Hippel, 2003; Shah, 2006; Hertel, Niedner, & Herrmann, 2003). Essentially most of the
developers possess skillful human capital (Hars & Qu, 2002), and could be economically
rewarded in a free market. The OSS human capital encompasses knowledge, capabilities
and skills acquired through participation and learning by doing.
Moreover, other studies embark upon determining factors driving sustained participation
and contribution to open source (Fang & Neufeld, 2009). Results showed that some of the
key reasons for contribution include an OSS gift driven culture (Bergquist & Ljungberg,
2001) that values altruism and reciprocity (Wu, Gerlach, & Young, 2007), satisfying
personal needs for the software (Lerner & Tirole, 2002; Hertel et al., 2003), self enjoyment
(Roberts, Hann, & Slaughter, 2006), career development (Lerner & Tirole, 2002; Wu et al.,
2007), enhanced reputation (Franke & Hippel, 2003), and seeking direct compensation as
a result of their involvement (Krishnamurthy, 2006; Roberts et al., 2006). While former
factors reflect intrinsic drivers that motivate developers to contribute to OSS, there are
other external factors that might have an influence on developers choice to contribute to a
particular project. A recent study found that the status of a project and prior ties with its
members might influence developers decisions to contribute to that project (Hahn, Moon,
& Zhang, 2008).
Expressed in economic terms, OSS developers motivation to contribute to OSS could be
measured as the sum of immediate and delayed payoffs (AIMarzouq et al., 2005). While 2
satisfaction from using the software represents immediate payoffs, prospective career
opportunities and community recognition symbolize delayed future payoffs.
Dissimilar to the exchange economy society where allocation of scarce commodities (i.e.
goods and services) takes place in a decentralized fashion through trade and voluntary
contribution, gift culture, such as OSS, is characterized by abundance not scarcity
(Raymond, 2001). To achieve the goal of attaining reputable social status in a gift culture,
OSS contributors are judged by what they give to the community rather than by what they
control. In a sense, an open source software developer denotes their prominence in the
community by handing over solitary control over developed code and sharing with the
Some studies argue that the success of a project is largely a function of that projects
ability to attract and retain skillful contributors. The recruitment initiative and project
attributes have to match prospective contributors motivations (Stewart, Ammeter, &
Maruping, 2006) and also assure that the project creates distinguished value
(Raghunathan, Prasad, Mishra, & Hsihui, 2005). Some developers are primarily driven by
the enjoyment and satisfaction of solving complex problems and crafting well designed
operational software. Others stress the importance of being recognized in the community
as outstanding individuals capable of delivering successful results. This notion of
recognition compensation as substitution for monetary compensation is acknowledged by
an open source developer as:
You may not work to get reputation, but the reputation is a real payment
with consequences if you do the job well" (Raymond, 2001).
This testimony confirms that open source developers' renewed interest and commitment to
open source is determined, in part, by the level of recognition received for contributed work
and peers' high regard.
2.3 Open Source Licenses
Some critics of free/open source software regard OSS as being antagonistic towards
intellectual property rights. The fact that OSS developers and contributors in general
acknowledge and preserve individuals intellectual property is reflected by the diverse open
source software licensing schemes. The licensing scheme, in part, serves as a governing
mechanism (Bonaccorsi & Rossi, 2003). This governance apparatus functions as
safeguard measures against opportunistic and ill-behaviors that might inflict a community
or transgress against the communitys norms and values.
OSS software source code is primarily hosted on the public domain for free use.
Accordingly, the free and open source group invented the concept of copyleft as an
alternative to proprietary software copy rights, which prevent users from having access,
modifying, or redistributing rights over source code. The free software adheres to four
levels of freedom (zero-3). These levels are established by the Free Software Foundation
as a measure of software openness. Essentially, the higher levels of freedom are
dependent on satisfying the basic level of having access to the code (level zero). As a
general rule, users should be able to have the following privileges (FSF, 2009):
1. Run the program for any purpose (freedom level zero).
2. Study program design and functionality, and adapt it to personal needs (freedom
3. Redistribute copies of the program (freedom level 2).
4. Improve the program and subsequently release improvements (and modified
versions in general) to the public (freedom level 3).
The two major entities that provide approved free/OSS software licenses are Open Source
Initiative (OSI) and Free Software Foundation (FSF). FSF licenses are classified in terms
of whether the license qualifies as a free software license, whether it is a copyleft license,
or whether it is compatible with the principal GNU GPL. Finally a license is assessed
whether it causes any particular practical problems that could jeopardize its use.
The main purpose of setting up OSS intellectual property protection schemes is to
stimulate ongoing innovation and creative thinking of members developing software while
protecting an individuals contribution from being misused in a way against OSS mission or
original contributors intent. The establishment of the licensing system is an integral part of
the OSS social structure, which seeks to empower users and a wider field of developers by
assuring free access to the code and granting rights to utilize it. In addition, the licenses
apply necessary check points to discourage attempts of restricting access to other users
and developers (Weber, 2004).
Although there are practical differences between various free/open source licenses, the
differences primarily reflect the original software developers motivations and level of
restrictiveness associated with code usage and derivative works. While some licenses are
relatively lenient in terms of usage, others are considered viral and restrictive. For
instance, GNU GPL is considered more restrictive than Berkeley Software Distribution
(BSD). GPL requires that all derivatives of the original code to be also free/open. In
addition, GNU GPL demands that code distributed under this license be included only with
other software distributed under copyleft licenses.
A project choice to adopt a specific type of license defines direction of the project and
potential opportunities for commercialization initiatives involving commercial partners
supporting the project or building auxiliary services around core projects.
Furthermore, the OSS licensing scheme behaves as a safeguard mechanism for protecting
projects from detrimental pursuits that violate open source mission and objectives. Some
studies argue that choice of license type and organization sponsorship of a project will
have an influence on users' interest in the project (Stewart et al., 2006). However, it may
subsequently limit potential adoption due to marginalization of users need to incorporate
the code with other software.
Open source software projects are virtual communities that operate under social structures
(Butler, 2001). An open source software project is recognized as the entity that receives
significant support from voluntary contributions represented in time, effort, and/or money. A
project is also a mechanism for protecting and maximizing developers reputation
incentives (Raymond, 2001). Reputation guarding is realized by protecting the integrity of
unit of work against negatively regarded activities by the OSS community such as forking
Research found that administrative and infrastructure open source software projects with
advanced technical user-base have more momentum and enjoy larger developer
contributions (Soderberg, 2007). These applications serve the needs of users/developers,
whose participation in the project is driven by fulfilling personal needs.
One example of administrative and infrastructure software adoption and diffusion among
various user levels is the Linux open source operating system compared to Microsoft
Windows proprietary software. End users familiarity with the graphical user interface and
relatively minimal acquaintance with shell scripting and command-line instructions limit
mass adoption of a more robust operating system for client machines. To fulfill this
limitation, a group within OSS community is pushing to extend the power and advantage of
Linux beyond enterprise data centers to the desktop/notebook space. Ubuntu3 is an
example of a growing OSS project that vows to fill this gap by simplifying end user
deployment and usage of the Linux operating system.
The above examples reflects the OSS community dynamism and continuous strive to
adjust itself to fulfill and take advantage of emerging opportunities. In response, new
project structures emerge, existing ones undergo restructuring, and projects fork and
mergers occur. Some projects are successful at achieving their goals and mission. Others
fail to take off and draw critical mass to carry on substantial operations. A project cannot
survive if founders are unable to recruit and retain talented and competent developers and
contributors. A project also needs to have a large enough community user-base for
product testing and improvement.
OSS projects follow an agile incremental development method. The concept of release
early and release often, where developers work on frequent iterations of software released
by members of the community, is one of the principles that define open source software
development process. Perpetual beta (PP) or continuous underdevelopment proved to be
an optimal software development life cycle (SDLC) methodology. Many commercial
software development organizations adopted open source software SDLC approach for
their product offerings. Some of the direct benefits of this process are lowering cost of
development and exploiting the power of community for debugging and functionality
2.5 Project Governance
Open source projects vary widely in terms of attributes and management styles. Project
life-cycle and application domain are some of the primary factors in determining
appropriate administration configuration. While some projects employ structured and well-
defined processes, others opt for more freedom and creativity in defining and managing
project tasks (Shah, 2006). Currently, there is no taxonomy for open source software
project governance. As a result, painting an absolute picture of the OSS management
process is a far reaching goal. Overall, it could be argued that OSS movement is primarily
cooperatively instituted. Yet, research shows that not all OSS projects were successful in
realizing a governance structure that drives tidiness and efficiency (Scacchi, 2002).
An open source project typically has a non-standard process of inception. The prevalent
mode is community established projects, initiated by one or more individuals independent
of their employment context (e.g. Linux, GNOME, etc) (West & O'Mahony, 2005).
Passionate volunteers exchange knowledge and ideas, which leads to the shaping and
creation of a working product (prototype). Project roadmap and feature prioritization are
usually decided via voting.
In relatively less popular instances, an open source project could be established by
commercial companies. Such companies opt for one of two means. They typically either
decide to open source part or whole of their products by releasing the source code, or
seek to establish an open source community to work on a project. One of the prominent
examples of a company initiated project is Mozzila4 that was established as a result of
Netscape releasing its browser source code (Mockus, Fielding, & Herbsleb, 2002). A
recent initiative by Google aims to establish a strong community by taking the Android5
project open source.
Normally, project initiators/owners are in charge of resolving some of the conflicts that may
arise as the project continue to develop. Despite the fact that some developers might have
more influence on the direction of the project; either because of seniority or better design
approach; such practice of having the whole community involved in the decision making
process insures everyone is respected and their opinion gets heard.
In general, three main conflict dilemmas require constant and careful resolution to insure a
healthy project community that abides by the OSS values. Research found that timeliness
and helpfulness of communication to be crucial in OSS work (Stewart & Gosain, 2006).
Disagreements could emerge as a result of identifying an individual or a group in command
of a) arbitration and final decision on the project, b) efforts to reduce duplication, and c)
demand accountability for work performed on the project (Raymond, 2001).
The process of documenting and defining work processes comes into mainstream as the
community grows and tasks become more complex and interdependent. The project
attracts external collaboration and sponsorship through communicating project features,
5 Android open source mobile platform: http://source.android.com/
functionality on mailing lists, newsgroups or online news services (Evers, 2000). Interested
participants recognize potential mutual benefits and examine project results and
milestones. Organizations planning to establish an alliance with an open source project
usually provide suggestions and recommendations for improvement to the community.
Favorable signals of community acceptance of commercial organization suggestions and
feedback sets the stage for forging a partnership between open source project and
Various relationship structures exist between open source projects and commercial
partners (Dahlander & Magnusson, 2005). The symbiotic approach, where both parties
gain from the alliance, might be the most effective at influencing the project members and
achieving collaborative governance mechanisms. The symbiotic relationship between open
source projects such as Linux and large commercial partners such as IBM enables the
project to achieve the expanded reach and efficiencies (Etemad, Wright, & Dana, 2001).
One of the advantages of such vendor collaboration with Linux for example, is developing
the project into enterprise-level software that meets high demands of scalability and
However, the symbiotic approach introduces managerial challenges related to decision
rights and control between the different parties. A consensual shared authority is needed
prior to parties attempt of establishing successful governance structure. Some of the
operational means of enabling governance require resolving ambiguity about control and
ownership, aligning different interests, creating and maintaining a positive reputation, and
investing in channels for proactive interactions.
Violation of open source formal and informal social structure for managing projects
methodical process or failure to direct individual efforts towards a common goal could
result in creating a project clone or a process of forking (O'Mahony & Ferraro, 2007). A
project fork is a situation where a particular aspect or feature of an established project is
used as the basis for creating a new project. The process implies that the newly formed
group is dissatisfied with the direction of the original project and they vow to take it to a
It is considered a relatively simple process to start a new open source software project fork
that typically has a different governance structure and common goals. As an anti-
authoritarian system, freedom of acquisition and modification of source code is conferred
by open source licenses. New project initiators are not required to secure permission from
the original developers, nor do they have an obligation to collaborate with the developers of
the base project. In practice, it is quite often the case that the developers of a fork will have
frequent interactions with the original projects developers and administrators, or the two
projects will develop a common working relationship. Yet, the level of communication and
collaboration is entirely voluntary. The forking of Foswiki project from TWiki in October
2008 is an example of community dissatisfaction with the hostile governance model
imposed by the commercialization of the open source project. It is unlikely both projects will
survive and dominate across market share in the long run.
Another form of structuring an open source project is by joining the work forces of two or
more projects in a merger-like transaction. The new structure is set to overcome scalability
issues and decision making complexity. The merger of open source web development
frameworks projects, Rails6 and Merb7, announced late December 2008 (Niccolai, 2008),
into a single project reflects a restructuring of individual projects management and
processes to accommodate newly formed project needs. The merger is considered a
healthy step to carry the project forward, since rivalry development leads to community
segregation and duplication of effort. It has been argued that changes in decision-making
processes and managerial procedures of the newly created project will benefit both parties
by ensuring architectural decisions that meet the needs of both parties. The decision to join
forces of the two projects position Rails, the merged project, to receive additional
contribution from commercial supporters.
Perens (2005) taxonomy of contributors to the open source software development
demonstrates several groups with diverse agendas and venture interest in open source
software. Contributors pay-offs expectations and course taken to get involved in OSS is
beyond the scope of this research. However, occasionally, commercial companies adopt a
secretive strategy (or at least a less publicized strategy) when determined to get involved
in open source initiative. Some of these companies objective is merely to counter
proprietary competitors market dominance (Kogut & Turcanu, 2000) or to tap into open
source think tanks for new ideas and innovations (Lerner & Tirole, 2005).
In part, this study explores viability and organizing form, when opposing incentive groups
come together to engage in OSS development transaction. Table 2.1 presents diverse
segments of non-volunteer contributors to OSS. These groups have different needs and
motivations. Also each stakeholder category satisfies a different role or function in the open
Table 2.1 Open Source Software Non-volunteer Contributors
Contributor Motivation Major Role Example
Software Packagers Revenue generating through brand quality service Integration, certification Fedora Core, Novell, Red Hat, Ubuntu
Flagship open source software companies Revenue stream from dual licensing mode, proprietary add-ons, training, and support Funds ongoing operation of underlying OSS project, certification MySQL AB, Pentaho, Sendmail Inc.
Commercial hardware and middleware vendors Open source software as an enabler of hardware or solutions sales, cost-efficiency Sponsor Hewlett Packard, IBM, Intel
Service Businesses Develop custom business solutions, take advantage of loopholes in OSS licenses by limit to internal use only Integrators of multiple OSS projects Financial industry
End-user businesses and their contractors Utilize OSS in daily operations Contribute own employees/contractors to work on OSS projects Amazon, eBay, Google, Yahoo
Government Not favoring particular vendor, avoid potential switching costs OSS development as a public benefit European Union (EU)
Table 2.1 (Cont.)
Contributor Motivation Major Role Example
Academic and scientific researchers Platform for publications, cost efficiency Free labor contribution by students and grants Bell Labs, University of California Berkeley
Software Foundations Accelerate the development and usage of free and open source software Directorship driven organization, provide support for the project community Apache Foundation, Linux Foundation, Mozilla Foundation
2.6 Governance Challenges
Open source software is predominantly a community-oriented development. All design,
implementation, and maintenance activities are carried out by a community of volunteers,
whose actions determine the prospects and directions of their project. This approach is
enormously different from proprietary software development. The emergence of OSS as a
viable product and successful method of organizing appealed to organizations and
In a different way, the tremendous success achieved by open source and emergence of
recognized brands such as Apache, Sendmail, MySQL and Linux associated with OSS
projects poses a threat to proprietary closed-code vendors. Proprietary software vendors
expressed their deep concern over OSS momentum and felt compelled to react in some
way. Microsofts reaction to open source momentum is expresses by the CEO statement:
Noncommercial software products in general and Linux in particular
present a competitive challenge for us and our entire industry and they
require our concentrated focus and attention (Topdog08.com, 2003).
Nowadays, several organizations actively pursue collective intelligence and innovative
production in partnership with project communities. Recent research identified finding the
optimal form of governance for this newly formed alliance between OSS project and
commercial partners to be a major obstacle (Sadowski, Sadowski-Rasters, & Duysters,
2008). The study also found that bazaar governance proved to be inefficient as OSS
project grows in terms of technical and structural complexity. As a result, an alternative
mode of governance emerged to account for new tasks structures and control mechanism.
Various levels of quasi-hierarchical governance forms are being shaped across various
projects to different degrees. Projects such as Linux and Apache demonstrate adapted
forms of hybrid bazaar governance.
2.6.1 Commercialization of Open Source
The phenomenal progress of the OSS bazaar style development fueled interest in adopting
the method in commercial and revenue-driven environments. In some cases, OSS
outperforms competing proprietary software in terms of market share dominance. One of
the key measures of successful open source project is level of adoption and user base.
Organizations and commercial vendors have been tracking open source projects that have
the potential to fill some gaps in commercial software product offerings or venture into new
technology innovation. The primary focus of these companies is to identify projects that are
able to present 'plausible promise (Raymond, 2001). Acknowledging the fact no open
source project has a complete set of features; emerging business models are established
to fill gaps in OSS software ecology. Several commercial open source companies are
competing with proprietary software providers introducing themselves as direct competitors
providing value-added services and business support (Fitzgerald, 2006).
Analytical investigation of commercial open source revealed that several OSS companies
dominated proprietary software in terms of market share and achieved higher user base,
when proprietary software vendors have relatively low strength of network effects (Sen,
2.6.2 Institutions Involvement
The passion about open source software and the intensity of growth has extended beyond
the community level. A number of commercial organizations and IT vendors made a
decision to seize the opportunity of OSS thrust. Today, organizations make an attempt to
reap both direct and indirect benefit gains. Building auxiliary services and ecosystem
around OSS core products are examples of additional software related services exploited
by these organizations. Furthermore, several organizations are exploring the potentials of
leveraging an OSS-like environment to garner some of the organizational and structural
benefits enjoyed by OSS (Shamna, Sugumaran, & Rajagopalan, 2002).
Some studies looked at the entry strategies adopted by such organizations to engage in
OSS (Bonaccorsi, Giannangeli, & Rossi, 2006), while others attempt to understand the
mechanism of reconciling differences between OSS community and rent-seeker institutions
(Franck & Jungwirth, 2003). Results show that commercial organizations and IT vendors
do not adopt a free ride strategy mentality of not contributing back to the OSS community.
Quite the opposite most of these organizations regard their participation in open source
software projects as an opportunity for gaining alternative operational approach and
acquiring novel competitive methodologies.
Organizations involvement in open source projects could take different forms and various
levels of involvement. Nevertheless, the top engagement patterns include coordination,
collaboration, and provision of code (Bonaccorsi, Lorenzi, Merito, & Rossi, 2007). Project
coordination is a widely common practice by organizations and IT vendors to facilitate
project activities and provide numerous support activities. One example of coordination is
organizations offering hosting services for OSS projects. On the other hand, collaboration
represents actual involvement in several phases of software development activities. For
instance design, development, testing, and bug fixes are part of collaboration activities
between project community and commercial partner. Ultimately, provision of
code/protocols includes organizations and commercial vendors specification of code or
protocols such as network communication protocols that define handshake and fall-over
Apart from organizations involvement type, OSS project partnership with commercial
partners could serve as a signal for evaluating project that give confidence and invigorate
users interest (Stewart et al., 2006). Since most users lack technical expertise for
thoroughly evaluating the project and examining its full features, presence of commercial
partners sponsors and supporters act as a surrogate for OSS project usefulness and
quality. Overall, the quality of project and its success is dependent, to a large extent, on
level of organization, communication, and control mechanisms. These factors and other
dimensions of governance are examined in more details in the next chapters.
3. Theoretical Framework
3.1 Transaction Cost Economics Theory
Transaction cost economics (TCE) is an interdisciplinary field that borrows from
economics, organization theory, and contract law (Williamson, 1979). Its part of micro-level
institutional economics that deals with the establishment of governance and institutional
arrangements. Neoclassical institutional theory identifies institutional arrangement as:
Arrangement between economic units that governs the ways in which
these units can cooperate and/or compete. It... [can] provide a structure
within which its members can cooperate ... or [it can] provide a mechanism
that can effect a change in laws or property rights (Davis & Smorodin,
The Nature of the Firm (Coase, 1937) seminal article is considered a path-breaking
contribution (Hardt, 2006) that laid the foundation for originating the theory by pointing out
controversial resource allocation problems planned through market mechanism versus
those planned within firm. In an attempt to justify existence of the firm, Coase argued that
the cost associated with usage of the market could be avoided or minimized through the
establishment of a substitute form of organizing (firm). In addition, Coasian theorem
admits a trade-off between the cost of organizing transactions within the firm and those
incurred in open markets. Critics of these assumptions contend that Coase failed to
acknowledge the focal characteristic of the firm as a governing structure of the production-
distribution process (Fourie, 1989). The differentiation factor between the markets role and
the firms is that the first involves connecting production and consumption entities, while
the later launches production of goods and services (Fourie, 1989).
Expanding Coases proposal, extension work on transaction cost is founded upon three
major elements: asset specificity, opportunism, and bounded rationality. Seminal empirical
work on the theory asserts that opportunism and bounded rationality are the key behavioral
assumptions on which TCE relies (Williamson, 1985). Asset specificity has been ascribed
as the degree to which durable and human assets are tailored for a specific economic
exchange, or quasi-rents that lose value when redeployed in a different transaction.
Adopting classical behavioral notion of the bounded rationality construct (Simon, 1957,
1987), Williamson embraced the concept based on two aspects. The first factor deals with
individuals limited capacity for dealing with information complexities. The second factor
has to do with economic actors incomplete access to information about the world. A
phenomenon recognized by Simon as intendedly rational, but only limitedly so. The
bounded rationality concept stems from the fact that humans have limited capacity, in
terms of skills, knowledge, and foresight. The organization model is a feasible alternative
for overcoming human limitations.
The principal thesis of transaction cost analysis is that since transactions with different cost
attributes can have different costs consequences, transactions with certain characteristics
can be organized relatively more efficiently, incurring lower transaction costs, by the
organization than by market. Some studies argue that transactions should be organized so
as to economize on bounded rationality while simultaneously safeguarding them against
the hazards of opportunism (Williamson, 1993). The theory explains why some
products/services are produced internally within a firm (vertically integrated) with a
hierarchical governance structure while others are produced and purchased on an external
market. This traditional view of optimal degree of outsourcing and choice of governance
structure is regarded in terms of market or hierarchy alternative (Williamson, 1975).
Over the last several decades, transaction cost economics has been instrumental in
explaining different forms of organizing. Compared to the neoclassic economic view,
transaction cost economists consider transaction as the fundamental unit of analysis.
Precisely, a transaction is considered any form of exchange involving good or service
transfer across separate economic actors barriers. Examples of broad discipline
application of the theory include outsourcing, strategic alliance, and contractual agreement
(Monteverde & Teece, 1982; Walker & Weber, 1987; Masten, 1993). The theory implies
that poor management of transaction costs would lead to an institutions decline and failure
(Rindfleisch & Heide, 1997). A summary of the underlying elements of TCE8 is presented
in the next subsections.
3.1.1 Governance Branch of TCE
Transaction cost operationalization emphasized relative efficiency of alternative
governance structures (Geyskens, Steenkamp, & Kumar, 2006). As one of TCE
emphasized concepts, applications of the theory largely focus on governance structures.
Operationalization work on the theory hypothesized that establishment of various forms of
governance is an indication of distinct explanation of various arrangement types
(Williamson, 1991b). Yet, the governance is largely a function of the institutional
8 Due to the extensive literature of TCE across various domains, selective review related to
governance in collaborative setting is presented here. We direct attention to cited sources
for further details.
environment within which transactions are taking place (Williamson, 1993). According to
TCE theory, Williamson argues that the key driving hypothesis is to:
Align transactions, which differ in their attributes, with governance
structures, which differ in their costs and competencies, in a discriminating
(mainly, transaction cost economizing) way (Williamson, 1991b).
Furthermore, the theory predicts that transactions are embedded in governance structures
that minimize vulnerability (Williamson, 1998). Gies, Ott, and Shafritz (1990, p. 178)
recognized governance as the function of control and administration, which takes place
when, a group of people come together to legally incorporate under the laws of a state for
a nonprofit organizational purpose. TCE theorists argue that in interfirm cooperation
settings, there might be adverse consequences to collaboration due to potential
untrustworthiness and self-interest behavior of argents involved in a transaction
(Williamson, 1979). Hence, having collaborating parties retain the incentive for maintaining
the relationship, by the establishment of governance mechanisms, also acknowledged as
safeguards, in order to reduce transaction costs incurred by opportunism and environment
uncertainty becomes an essential goal of the theory.
TCE differentiate between three general forms of governance mechanisms, namely:
vertical (hierarchical) highly-specific governance (Williamson, 1979; Barney, 1999),
intermediate or hybrid (semi-specific) governance (Williamson, 1979,1985), and market
(non-transaction-specific) governance (Williamson, 1979; Judge & Dooly, 2006). This study
is motivated by the second type of governance; explicitly semi-specific or intermediate
structures for governing alliances between an open source project and profit-oriented9
9 Could be any group, not just for-profit organizations
partners. In an attempt to address the governance decision, some earlier studies
maintained that the governance question is simply a factor of early selection and setting
some form of socialization efforts, or a combination of both (Heide & John, 1990). Equally,
level of integration of operational decisions between two economic entities is regarded as a
key aspect of governance structure (Clemons & Row, 1992). Williamson (1991a)
recognized exchange agreement and reciprocal trading as forms of hybrid structures.
Other studies recognized this form of organization as value-added partnership (Johnston &
Lawrence, 1988), strategic network (Jarillo, 1988), and strategic alliance (Webster, 1992;
3.1.2 Semi Specific Integration
Vertical integration represents a restrictive form of organization favoring internal exchanges
within firm boundaries. There are two main characteristics that identify vertically integrated
firms; a) employing whole output of a process, as part or all, of one intermediate input into
the process; b) acquiring the entire quantity of one intermediate input of the process from,
all or part of, the output of the process (Perry, 1989). This method implies full control of
resources by organization and eliminates dependency on external assets.
According to one view (Perry, 1989); transactional economies perception is one of the key
determinants of vertical integration. Seemingly, the outcome of vertical integration is
reducing the requirements of intermediate exchange inputs. TCE theorize that one of the
incentives of firms adoption of integration of a process is to internalize transactional
economies and eventually reduce its transaction costs. From a transaction cost point of
view, various forms of interorganizational relationship, such as joint ventures or network
structures, are considered alternative forms of governance and departures from the
generic organizational hierarchy (make) or market (buy) decision (Barringer & Harrison
3.1.3 Asset Specificity
Asset specificity is a characteristic of an investments transferability from one transaction or
setting to an alternative one. Recognized as a notion of sunk cost, TCE maintain that
switching specific assets from one setting to a different one will result in lowering the value
of these assets. Therefore, partners associated with a transaction that involves
appropriable quasi-rents are more likely to remain in partnership and work together to
attain mutual satisfactory benefits. Essentially, the theory upholds the precondition
assumption that asset specificity implications come into play only under circumstances of
incomplete contracts (Williamson, 1975,1979; Klein, Crawford, & Alchian, 1978).
Correspondingly, TCE research differentiates between two types of asset specificity:
intangible and tangible. Intangible factors are identified by the uniqueness of skills,
functions and business knowledge required for completing a particular transaction
(Williamson, 1985; Young-Ybarra & Wiersema, 1999; Subramani, 2004). On the other
hand, tangible assets include any form of physical or monetary contributions. In summary,
the six major types of assets specificity recognized in the literature can be represented as
(Williamson, 1989): 1
1. Site specificity, which is most common in manufacturing facilities that benefit from
close proximity and geographical location to complete the transaction.
2. Physical asset specificity, which represents special purpose equipment and other
tangible capital investments dedicated to a particular transaction.
3. Human asset specificity, referring to specialized skills and knowledge capital
developed by engaging in a specific transaction.
4. Dedicated assets, in the form of discrete investments devoted for a single
5. Temporal specificity, occurs when synchronous and timely response is required for
6. Brand name capital, that provides useful information to users and customer about
product quality and value.
Its important to point out that other forms of asset specificity might take place during the
course of a transaction. For example Polanyi (1998) identified personal knowledge as an
important type of idiosyncratic assets. This implies that the more an asset is customized for
a specific transaction, the more idiosyncratic that asset becomes.
As a consequence, the theory claims that asset specificity leads to diverse forms of
governance structure, formed in response to protect such investments (Williamson, 1989).
This assumption becomes an integral notion in explaining why organizations adopt
different forms of governance.
Human bounded rationality and limited capacity to encompass all variables of decision
making has an effect on degree of uncertainty present in a transaction. TCE acknowledge
that hazards are eminent due to behavioral uncertainty appearing as a result of a
combined effect of incomplete contracting and asset specificity. The theory highlights the
critical effect of uncertainty in the presence of substantial asset specificity investment
(Williamson, 1985). Therefore, determination of most relevant aspects of a transaction is
achieved with reasonable amount of uncertainty. In cases of high environmental variability,
the writing of complete contracts might be difficult and contractual gaps may demand and
escalate adaptations of the contract as the transaction evolves.
Empirical studies have shown that certain forms of uncertainty have direct influence on the
choice of governance structure and subsequently transaction costs (Heide & John, 1990;
Masten, Meehan, & Snyder, 1991; Zaheer & Venkatraman, 1995). A multi-dimensional
uncertainty construct identified in the literature includes: technological uncertainty, which
involves technical level of future product change (Walker & Weber, 1984; Balakrishnan &
Wernerfelt, 1986), behavioral uncertainty involving parties joined in a condition of bilateral
dependency (Anderson, 1985; Williamson, 1989; Heide & John, 1990) and environment
uncertainty originating from external factors associated with surroundings (Koopmans,
1991; Walker & Weber, 1984). The theory also involves a secondary type of uncertainty
factor, demonstrated by lack of communication between decision makers that restrict
access to concurrent decisions and plans made by others (Koopmans, 1991). While some
studies claim the latest type of uncertainty to be nonstrategic (Williamson, 1989, p. 143),
this study argues that uncertainty due to lack of or non-established channels of
communication becomes important in transactions linking parties of corporate and
Some studies also argue that presence of environmental uncertainty induce organizations
to build coalition in the form of vertical and horizontal interorganizational relationships
(Galaskiewicz, 1985). By expanding boundaries, interorganizational relations are better
positioned to sense-and-respond to environment uncertainty.
Acknowledged as one of transaction cost economics behavioral assumptions, opportunism
is characterized as the human trait of seeking self-interest with guile (Williamson, 1989, p.
139). These moral hazards take the form of intentional efforts to mislead, distort, and
misrepresent factual situations to achieve personal gain and reward at the expense of the
other party or relationship in general. The potential of a partner or human agent to default
on the other or exhibit opportunistic behavior will have negative implications reflected in
incurring higher transaction costs.
The theory also argues that threats of significant opportunism will lead players in a
transaction to opt for a governance of collaboration that safeguard against these threats.
Safeguarding against potential opportunistic behavior might involve transaction costs in the
form of negotiating costs, bonding costs, monitoring costs, and enforcement costs (Hill,
1990). Correspondingly, research reveals that opportunism is deterred by choosing
between two main protection modes: equity-based and non-equity-based safeguard
mechanisms (Oxley, 1997). These mechanisms largely came in response to prior studies
that pointed to the negative effect opportunism had on partnership performance and
outcome (Anderson, 1988; Hill, 1990; Miranda & Kim, 2006).
Strategic alliance research showed that opportunism is a particularly important problem,
especially in alliances that engages members from different organizations (Judge &
Dooley, 2006). In addition, several studies revealed that while it is not necessary all agents
have or exhibit the same level of opportunistic behavior, any perception of opportunism
between the parties of an alliance would negatively impact performance (Williamson 1979;
Ring & van de Ven, 1994; Saxton, 1997).
3.2 Interorganizational Cooperation Theory
While transaction cost theory represents a general framework for examining governance, it
does not fully address some of the important aspects of alternative forms of governance.
To the extent that TCE is an adequate theoretical lens for this study, other aspects of
interorganizational cooperation and the role of trust in joint collaboration between open
source and for-profit organizations must be taking into account. Directing the investigation
within definitive boundaries, by concentrating on key activities characterized by the
institutional model yields more informative explanation of the hybrid relationship between
OSS project and commercial partners. Employing the interorganizational cooperation
premise, within the domain of transaction cost economics, the study introduces key
elements of the partnership governance. Although transaction cost economics is a useful
tool to investigate various aspects of strategic alliance (Gulati, 1995), this research posit
that TCE and articulated dimensions of interorganizational cooperation behavior, together,
provide a richer foundation for explaining governance issues of open source projects and
commercial companies collaborative transactions.
Emphasizing the impact of interdependence between the parties involved in a transaction,
interorganizational cooperation theory argues that parties tend to cooperate when there are
shared assets and dependency on each other (Williamson 1985, 1991b; Osborn &
Hagedoorn, 1997). Early research investigating determinates of interorganizational
cooperation found that cooperation is more likely to take place in circumstances where
organizational domains are not sensitive issues (Schermerhorn, 1975). In most cases, the
parties are not engaged in competing activities. Also, collaboration is likely to occur where
mutual goals among parties are obtainable.
Although studies differ in their justifications for why organizations want to enter into this
form of alliance (Lefton & Rosengren, 1966; Aiken & Hage, 1968; Doz & Hamel, 1998;
Barringer & Harrison, 2000), some of the common reasons include: increased efficiency
(Clemons, Reddi, & Row, 1993), knowledge sharing, internal organizational diversity,
lateral and longitudinal dimensions of organizational commitment to clients, attainment of
objectives that can only be achieved through cooperation, help firms create value by
combining resources, increasing speed to market, and gaining access to foreign markets.
However, one of the main benefits of cooperation is the potential reduction of transaction
cost for interorganizational alliances (Jarillo, 1988).
There are also some negative implications associated with interorganizational cooperation.
For example, entities participating in interorganizational cooperation venues might suffer a
loss of decision-making autonomy, experience loss or damage to their identity and image,
or they might over burden their limited organizational resource (Schermerhorn, 1975).
Furthermore, all aforementioned forms of disadvantage potentially incorporate extra costs
attributable to interorganizational cooperation.
Due to the diverse and multifaceted nature of interorganizational relationship formation, the
scope of this study lays emphasis on the TCE paradigm, as one of the major theory-
perspectives for explaining collaboration. Next section embarks upon transaction costs
institutional arrangements and established cooperative activities in the context of
governance of open source and commercial organization alliance.
3.3 Application of Theoretical Framework to Open Source
Following from the previous discussion, TCE is considered the study of 'alternative
institutional governance' (Klein, 2008). As affirmed by the theory, TCE assert that
economizing is the core problem of economic organizations. Moreover, the principle of
defining various forms of governance structures, or safeguard and control mechanisms, is
to promote transactions egalitarianism. This study argues that open source software
development is considered a special economic arrangement. Although it appears to lack
monetary incentive drivers, still other forms of organization apply including labor and
resource allocation in order to produce public goods and services. Open source software
development provision as public good stems from the fundamental voluntary contribution
notion of the public goods theory (Johnson, 2002). In particular, OSS development regimes
create software products available for both original volunteer contributors, as well as, the
A strong open source projects community is a predecessor for success. Projects cannot
survive and take-off without a resilient community. The significance of building a vibrant
community becomes more critical for proprietary spinoff projects seeking to initiate a new
open source project. The importance of dedicated and enthusiastic leadership is
recognized as one of the major factors for ensuring focused vision and fueled innovation.
In contrast to commercial software development process management, OSS project
leadership is less authoritative and more about recruiting and vetting good talent for the
project and avoiding interference.
In a hybrid collaborative setting, the partnership between OSS and the alliance
organization could be a result of the company being the initiator of the project. For example
Hewlett-Packard open sourced proprietary code of the Spectrum Object Model (SOM)
linker and created some of the governance structure for the open sourced project (Feller,
2005, p. 68). Alternatively, the partner proprietary organization joins an existing open
source development effort. Regardless of original partner leadership status, the new
leadership typically has to earn credibility (Fogel, 2005). One of the objectives of
collaborative software development between open source projects and commercial
partners is the arrangement of institutional form based on principal values and an interest
in minimizing cost. This new configuration would lead each party to perceive themselves at
an advantage point by continuing the alliance than by ending it. As a consequence of
establishing governance mechanisms for managing the partnership, parties would reduce
transaction costs incurred by opportunism and other environment uncertainty (Williamson,
1985). Although no formal contacts are defined to oversee cooperation between the two
parties, yet OSS-commercial vendor affiliation is governed through embedded unofficial
'self-enforcing' agreements (Telser, 1980), sponsorship and safeguards.
Extending transaction cost analysis to non-profit organizations, the OSS domain requires
careful treatment of the predicted variable. Nevertheless, the hybrid model attempts to
define an optimal governance structure that yields meritorious results. The special control
form is established to oversee exchange agreement and joint operation as a semi-specific
structure. From a commercial perspective the partnership between an OSS project and
business entity is considered a special form of value-added partnership.
Given recognition of the special form of collaboration between open source and
commercial organizations, its expected that a new form of interorganizational cooperation
would emerge to account for two-party differences. In addition, it is likely that the role of
transaction cost will presume another level of importance and allow for alternative
interpretations within the open source context. For the purpose of the research, the study
focus attention towards the type of organizational interdependency that involves the
establishment of joint, cooperative activities between open source software project
communities and revenue-driven organizations (i.e. OSS companies, IT vendors, etc...).
This study uses a multi-construct scale for assessing governance arrangements for open
source project and commercial vendor collaboration, pertaining to information exchange,
level of flexibility, influence restraint, shared responsibility, and overall satisfaction with the
relationship. In conformity with established view of alliance and domains of cooperation
(Heide & John, 1990; Heide & Miner, 1992), the study hypothesize that the degree of
cooperation between the two parties involved in a hybrid mode of open source software
development, as opposed to pure open source or totally proprietary development, is a
function of their boundary transparency and readiness to rationalize commitment to each
other. Such practice is a first step towards achieving a justifiable cost- suppressant
governance structure along with underlying building blocks. Research revealed that
satisfaction with the alliance and experienced meritorious results leads to sustained
cooperation between the members of an alliance (Williams, 2005).
Primarily, open source software development is a diverse economic activity, wherein
players are conscious about potential hazards and dynamically adjust processes and work
norms for enhanced efficiencies. As a result, software development as a transaction is
embedded in governance structures that minimize vulnerability (Williamson, 1998) and
create a perpetual environment for future growth and scalability. The importance of
establishing bilateral communication channels is recognized in classical organizational
theory to have both supplementary and complementary roles to existing channels in any
organization structure (Hrebiniak & Joyce, 1984). In the absence of formal firm boundaries,
open source projects undertake transparency and a free flow of knowledge and information
on project mailing lists and discussion boards as mechanisms for mitigating uncertainty
and potential conflicts.
As commercial vendors show active presence in the project, amplified levels of
communication and information sharing becomes more eminent for resolving dependency
and assuring complete engagement. Collaboration research shows that the alliance form of
governance is influenced by the level of interdependence and requirement for information
sharing (Gulati & Singh, 1998). The higher the level of interdependence and transaction
complexity, the greater the amount of information sharing needed between partners.
Branding is claimed to be driven by human needs to differentiate between different
products and services. Brand name is considered a special form of specific assets
investment by parties engaged in a transaction. It creates an expression of the core values
of the organization (Holland, 2006). Although brand capital and brand management are
well investigated aspects in the commercial sector, it started making its way into research
within nonprofit contexts. Although most of brand research for nonprofit is conceptual
based or case study type, it constitutes a valuable addition and distinction from marketing
and profit-based brand equity. Saxton (1994) highlighted the importance of and process
involved in creating strong charity brand. In addition, Ewing and Napoli (2005) found that
brand image has a significant role in shaping stakeholders attitudes and actions, and
stimulating their commitment. Another study found that human resources management,
brand communication and brand leadership are pre-required factors for provoking brand
commitment (Burmann & Zeplin, 2005). From a transaction cost standpoint, the collective
efforts put by the parties could determine the value of brand capital for the partnership.
Such value will diminish by parties exhibition of opportunistic behavior (Klein, Crawford, &
Trust as a multi-dimensional construct across personal and organizational levels plays a
key role in interorganizational relations and relational governance. In particular, trust is
proven to have positive impact on negotiation costs, managing cross-organizational
conflicts, and ultimately overall performance (Zaheer, McEvily, & Perrone, 1998).
Emphasizing role of 'societal culture on economic organizations, Williamson (1993) argues
that group culture, such as open source software culture, serves as a check for
opportunism in informal organizations. An informal organization, where moral hazards such
as lying and deception are inhibited is more likely to experience efficiency enforcing social
contract and maintaining personal integrity.
According to TCE, firms opportunistic behavior (Williamson 1975, 1985) leads to increased
transaction costs due to necessity of safeguard mechanisms. However, trust as a mediator
factor in relational exchange is recognized as a source of competitive advantage (Barney &
Hanson, 1995). Zaheer et al. (1998) defined interorganizational trust as the degree to
which members of one organization maintain a communal trust orientation toward another
organization. Yet, this definition does not equate interorganizational trust to that of an
individual-to-individual trust. An economic view, recognized trust as a quasi-rational
calculation of the probability of subsequent future benevolent actions (Gambetta, 1988).
Alternatively, an organizational view identifies trust as being relational or reciprocal in
nature (Zaheer & Harris, 2005). It is imperative to acknowledge the fact that asymmetric
nature of interorganization trust may involve power and resource dependency that stem
from one partys vulnerability on other parties of the alliance (Pfeffer & Salancik, 1978).
The role of trust within interorganizational collaboration and transaction cost economics
framework embark on choosing an organizational structure that safeguards against
opportunism and lowers transaction costs. Fostering commitment and trust among parties
or participants in a transaction is found to be an antecedent to ensure effective cooperation
(Heide & Miner, 1992).
3.3.3 Product Distinctiveness
Two products are considered differentiated when users or customers perceive the two
products to be close substitute to each other (Eaton & Lipsey, 1989). Open source
software project and commercial organization alliance involves sharing part of their value-
creation activities (Ghosh & Morita, 2007) such as product design, technology development
and evangelism. These procedures represent a subset of activities for promoting project
distinctiveness. The parties engagement in promoting product exceptionality would lead
each to assume ownership of the alliance and ensure a successful and high rewarding
3.4 Competing Theories
This study assessed various other information systems and economic-based theories that
could be adopted for studying unorthodox forms of structure or capable of explaining some
of the research questions. The research concluded that while some of these theories might
have some relevancy for understanding open source and commercial partner collaboration,
TCE proved to be the most relevant theory for conducting the study. Some of the primary
justifications for adopting transaction cost and interorganizational cooperation analysis as
theoretical lens for this research are:
1. TCE regards the organization (firm) as a governance structure rather than a
2. OSS project is an organization with casual, non-hierarchical governance structure.
3. The theory is self-conscious about organizational behavioral assumptions and
incorporate safeguard methods for protecting against vulnerability related to
4. OSS project is a self-organizing, self-enforcing governance structure that utilizes
alternative mechanisms for formal contracts.
5. TCE relies on comparative institutional analysis (OSS project and IT vendor have
different governance structures).
6. TCE is micro-analytic, which fits the study level of analysis (Williamson, 1989).
This section provides an overview of alternative theories used in IS research and their
applicability to research hybrid form of OSS governance. The decision to adopt TCE and
interorganizational cooperation theories as lens to investigate open source project and
commercial partners collaboration is based on being the best suitable framework for the
special nature of the alliance. However some of the questions addressed by this study
might be approached using alternative theories.
3.4.1 Contingency Theory
The contingency theory deals with behavioral issues related to organization structure and
organization managerial approach. Seminal article on the theory argue that there is no
prescribed best way to organize. Moreover, not all forms of organizing are considered
equally effective (Galbraith, 1973). The theory tries to identify factors that affect
organization performance and efficiency. A study investigated the organizational
contingencies and their influence on the selected mode of IT governance found that
multiple interacting contingency forces have reinforcing, conflicting, or dominating effect on
the IT governance mode (Sambamurthy & Zmud, 1999).
Prior IS research adopted contingency theory for exploring the effects of IT outsourcing
strategies on success (Lee, Miranda, & Kim, 2004). Vet these studies admit that
outsourcing choices, make-or-buy decision, are better understood within the light of the
theories of the firm and governance (TCE).
The theory has potential usefulness in identifying contextual variables related to an open
source project that leads to a successful governance structure. Contingency theory study
of organization strategy could be relevant to examining contextual variables that vary
based on OSS project size and type.
3.4.2 Resource-based View of the Firm
The resource-based view (RBV) theory focuses on organizations resources subset that
enables attainment of competitive advantage, as well as, the subset that enables
sustainability of competiveness (Wade & Hulland, 2004). Although, OSS project does not
own the contributing resources, it is fair to assume that a subset of these resources, mainly
leadership and core developers, drive competitive advantage for the project. RBV
relevancy for IS research in general and open source in particular stems from identifying
distinct resources that contribute to overall performance. Application of the theory would be
more relevant for studies investigating nature and characteristics of each resource that
lead them to achieve sustained competitive advantage. A study trying to identify factors
that lead to particular OSS project dominance would look at the projects asset and
attributes in addition to complementary assets that fueled sustained competiveness. The
theory is not situated to investigate alternative governance structures for hybrid OSS.
Another limitation of the theory application to this study has to do with vagueness of
defining variables that capture competitive advantage for the hybrid model.
3.4.3 Agency Theory
The agency theory contract metaphor deals with the conflicting goals and desires of
cooperation parties, ascribed as principal and agent. The theory posit that the party who
delegates work (principal) has limited or no control on the other party (agent) executing the
task (Eisenhardt, 1989). The latest problem stems from principals vulnerability of agents
self interests and inappropriate behavior. Employing contract as the unit of analysis,
agency theory is motivated by development of efficient contracts that govern the
relationship between principal and agent. However, when an organization opts for vertical
integration of product or service, technically the principal is taking over an end-to-end
process without relying on agents to accomplish task on their behalf.
Lack of contracting in alliance involving open source project and commercial partner
greatly limits explanation of hybrid governance structure using agency theory. The OSS-
commercial partner relationship does not fit theorys discrete roles. Appropriate labeling of
principal/agent parties involved in this type of informal alliance is a challenge. Although the
parties might have competing motives, lack of formal contract held by the theory, provoke
alternative measures for developing efficient structure that safeguard against opportunistic
behavior. Some studies employ both agency theory and transaction cost economics to fully
explain outsourcing risks and managing relationships between parties (Bahli & Rivard,
2003). The focus of this research is to assess optimal forms of governance for OSS project
and commercial partner special alliance. Agency theory might be relevant in addressing
some of the risks associated with conflicting goals between non-profit open source and
profit-oriented commercial partner.
3.4.4 Resource Dependency Theory
Resource dependency theory (RDT) investigates organization success in terms of power
exercising. Power maximization is one of the criteria for determining organizational
success (Pfeffer, 1981). As one of the theories used to investigate interorganizational
relationship patterns, RDT posits that entities lacking resources will establish relationships
with other parties to fulfill needs for external resources. Adjustment of power relation
between organizations engaged in a relationship defines social coalitions, which emerged
from social exchange (Ulrich & Barney, 1984). The theory assumes that increased
dependence of other organizations on a focal actor will lead to the perception that focal
organization is the more influential in the relationship.
Environment uncertainty is one of the factors that determine an organizations ability to
obtain scarce resources from other parties. To reduce uncertainty, organizations attempt to
exercise control over such resources or exert control over other organizations that own
these resources. The theory is more relevant to explore social network power structure. It
could also be applied to investigate commercial partner influence and control on open
source software project when they engage in an alliance or a social exchange transaction.
3.4.5 Structuration Theory
Structuration analysis is primarily positioned as a social organization theory. It deals with
the establishment of social structure at an abstract level. This social structure as a
resource is obtained by human agents interactions (Giddens, 1984). Structuration is
defined as conditions of governing the continuity or transformation of structures.
Connections between actors define the system (Jones & Karsten, 2003). The theory
argues that the rational model of powers is form of dialectic of control with negotiable
power and influence.
The theory is more relevant in investigating some of the non-technology aspects of open
source. The social structure, knowledge management and why participants chose to
maintain a certain form of structure are some of the viable applications for the theory. In
addition, there has been some attempts in IS to employ structuration theory in studying
dynamics of globally distributed virtual teams and their effectiveness (Maznevski &
3.4.6 Noncooperative Game Theory
As a multi-person decision making, noncooperative game theory is a modeling framework
for analyzing stakeholders diverse optimal decisions, noncooperative, influenced by their
background and believes (Fudenberg & Tirole, 1989). The theory assumes that opponent
players are rational and able to assess other partys actions. Research indicated that
games serve as satisfactory models for simulating situations involving conflicts of interest
(Ritzberger, 2002). The three main pillars for designing a game include:
1. interacting parties, participating in the game
2. rules of the game or game form, define specific actions, roles, and behaviors
Since most social interactions incorporate all of the aforementioned constituents, the game
could simulate organizations interactions and OSS agents roles. The theory acknowledges
imperfect players monitoring and potential hazards of ill behavior. Noncooperative game
theory could be used to simulate open source and commercial partner conflicting motives
and decision making process. Having players with dissimilar cultures and backgrounds, the
theory could model interactions to determine the equilibrium point that yields optimal
Table 3.1 presents a summary of identified theories and potential relevance for addressing
particular questions related to the study of open source and commercial partner alliance.
Table 3.1 Summary of Alternative Theories
Theory Potential relevancy for the study
Contingency Theory Optimal form of governance is a function of OSS project and commercial vendor partners cultural and structural differences. Success of a governance structure will vary based on contextual variables.
Resource-based Complementary assets contributed by OSS project and
View of the Firm commercial organization partner to develop product distinctiveness and successful long-term relationship.
Agency Theory Could be applicable for investigating the mechanisms used by an OSS project (principal) to safeguard against IT vendors (agent)opportunism, but not why want to collaborate
Resource OSS project lack of resources could be a driving motivator for
Dependence Theory seeking IT vendor partnership. The theory does not explain what's the optimal form of governance for the alliance
Structuration More appropriate for investigating individual OSS contributors
Theory behavior to shape project casual structure and work norms.
Noncooperative Game Theory Alignment of conflicting motives and interests in an alliance
4. Hybrid Open Source Software Project Efficiency
Applying the premise of transaction cost economics to the open source software domain
addresses some of the puzzling questions about the development of OSS and the
emergence of numerous organizing models. An assessment of transaction cost research
revealed a significant amount of TCE investigation aimed at examining the efficient form of
organization for various economic activities (Shelanski & Klein, 1995). Yet, few endeavors
have been made towards addressing the characteristics of special governance structures,
which involve commercial organizations and open source community organization. This
study maintains that OSS development is an economic activity in quest of establishing
efficient organization for managing activities. Examining determinants of hybrid OSS
efficiency is one of the research objectives addressed in this chapter.
Grounded in the foundational work on TCE presented in the previous chapter, this study
introduces an analytical assessment of the hybrid model of open source software
development. Specifically, this part of the study adopts transaction cost fundamentals for
investigating the efficient form of organization for the OSS hybrid relationship. The
efficiency model is designated as a primary step essential to understanding the
phenomenon of OSS project partnership with commercial organizations and proprietary
vendors. The initiative of extending alliance denotes departure from the classic OSS
bazaar-oriented approach primarily driven and governed by volunteer contributors.
This analysis looks at the underlying factors for determining efficiency of the relationship
and drivers of persistent alliance. Satisfactory outcome of the alliance is an indication of
successful establishment of an efficient form of organization for the OSS-commercial
partner relationship and representation of a sustained relationship. This logic is supported
by positive economics view that assumes inefficient governance arrangements are likely to
be discovered and abandoned (Friedman, 1953). Therefore, the chances of survival for an
alliance that is not perceived as yielding meritorious end results are very limited.
A principal goal of hybrid OSS collaborative development between open source software
projects and commercial partners is the institutionalization of efficient forms of organizing
that achieve mutual goals while adhering to the basic notion of minimizing cost (Sidahmed
& Gerlach, 2009). This novel configuration guides a hybrid alliances parties to realize
added-value of a sustained relationship and commands both OSS project community and
commercial partners representatives to coordinate innovative governance structure that
insures long-term proficiency.
As a consequence of establishing efficient governance mechanisms for managing the
relationship, parties would reduce transaction cost incurred by opportunism and other
forms of uncertainty (Williamson, 1985). Also, joint investment in specific assets dedicated
for the two parties transaction will direct both parties to insure process efficiencies in the
long run. Despite the fact that the relationship between OSS project and commercial
partner is administered under informal contracts, still, other social and community
protection mechanisms may take place as precautions. One example of the preventative
measures is institutionalization of a governance structure with embedded self-enforcing
agreements (Telser, 1980). Subsequent sections introduce the hypothesized model and
empirical assessment outcome. The theoretical research model for the study is depicted in
Figure 4.1 Efficiency Theoretical Research Model
4.1 Research Hypotheses
Building on the theoretical framework presented in chapter 3, this chapter develops and
tests a hypothesized efficiency model. Applying fundamental TCE constructs, the empirical
analysis aims to test strategic factors that determine hybrid OSS model efficiency. The
remainder of this chapter is structured as follows: derived hypotheses are introduced next,
a pilot study assessment is presented in Â§ 4.2. Â§ 4.3 discuss the study sample, variable
measure, and analytical procedures. Hypotheses testing results are presented in Â§ 4.5.
Finally the last section presents the conclusion, study limitations, and future research
Asset specificity is considered the most important dimension of a transaction (Williamson,
1981). This notion entails investments transferability from one transaction or setting to an
alternative one. It has been argued that switching specific assets from one setting to a
different one will result in lowering the value of such assets. As a consequence, partners
associated with a transaction that involves special assets dedicated to the transaction are
more likely to remain in partnership and work jointly to attain mutual satisfactory benefits.
TCE posit that asset specificity criticality stems from the fact that investment in such type of
assets creates a bilateral or to some extent, a quasi-bilateral operating environment that
ensures continuity of relationship for an extended period of time.
Specific assets dedicated to the transaction involving open source project and commercial
partner organization, such as knowledge specific skills to the OSS project, forms a cost-
conscious parsimonious assembly. TCE hypothesizes that the relationship between parties
of an alliance would vary depending on the level of specific assets involved in the
transaction. In general, commercial organizations involved in hybrid OSS development
transaction would insure some promise from the OSS project community before investing
in specific types of assets. Pledging commitment from both parties on investment would
ensure desire for continuation of the relationship and achievement of efficient form of
establishment. Accordingly, the first hypothesis related to specific asset investments on
hybrid OSS efficiency is stated as follows:
H1: Investment in specific assets for open source project and commercial partner
alliance will have a positive effect on satisfaction with the partnership.
TCE differentiates between two types of uncertainty: behavioral or internal, and
environmental or external (Williamson, 1985). Behavioral uncertainty is related to partners
behavior, while environmental uncertainty is a function of external factors related to the
operating environment. It includes such contingencies that unfold during the course of
transaction execution. A recent study, found that environmental uncertainty increases the
risk level for organizations (Fink, Edelman, Flatten, & James, 2006). Technological
uncertainty and resource uncertainty are foremost types of environmental uncertainty.
Technological uncertainty represents future technological changes that could impact
product development and users expectations. Resource uncertainty has to do with
resources availability and steady availability of skillful work force. Some studies argue that
firms perceived uncertainty about resource availability is likely to impact their decision to
form relationships (Williamson, 1985).
The environmental, technological, and behavioral uncertainties are perceived as the major
constituents of this construct. Former study argues that environmental uncertainty can
motivate organizations to develop interorganizational relations to mitigate uncertainty
(Galaskiewicz, 1985). The partnership between open source project and commercial
organization could suffer from emergence of unpredictable behavior from one partner, or
experience of negative consequences for either party. As a mitigation strategy, some
recommend a three-dimensional strategy to cope with uncertainty in interorganizational
collaboration. The triangular plan incorporates forestalling, forecasting, and absorption
(Penning, 1981). The OSS-commercial partners alliance is predisposed and inclined to be
exposed to technological and environmental uncertainty. Such rendering is detrimental to
process efficiencies and require members of both parties to exercise more governing
control and establish ties to the environment (Pfeffer & Salancik, 1978). The greater control
and increased formalization of work processes, to some extent, defeats the purpose of
leveraging OSS informal and inexpensive flexible structure. Overall, hybrid OSS model
efficiency is negatively related to uncertainty associated with the development transaction.
Subsequently, the second hypothesis is positioned as:
H2: Increased level of uncertainty will have a negative effect on satisfaction with
Transaction cost economics recognize the negative impact of opportunistic behavior of
parties involved in a transaction on the outcome (Williamson, 1989). Correspondingly,
open source software project and commercial partner's relationship is not immune against
opportunism. As an economic actor's behavior of satisfying self interest with guile,
opportunism involves cheating, deception, and misrepresentation conduct. Acknowledging
the fact that not all people are honest or trustworthy, and they might seek crooked ways to
take advantage or exploit other parties involved in a transaction, some will occasionally act
opportunistically. TCE theory admits the difficulty or unfeasibility to identify in advance
those who are likely to default or deviate from partnership agreement.
In a hybrid open source software transaction domain, opportunism might occur due to the
fact that OSS project has limited control over commercial partners. Also the commercial
partners have personal agendas and interests. Moreover, lack of contract, incomplete
information and impracticality of monitoring partners activities leads to low efficiency and
unsatisfactory relationship (E. Anderson, 1988). Similarly, a study revealed that
opportunism intensity increases by one partys employment of tactics and methods used to
exercise influence (John, 1984). Also bureaucratic and complex structured settings with
low communication levels between partners, compared to loose and informal relations,
increase the likelihood of experiencing negative consequences of opportunism.
Concerning OSS project and commercial organizations alliance, parties could undertake
different forms of self interest seeking such as misrepresenting information, efforts, and/or
activities. Concisely, opportunism among parties will lead to inefficient and unsatisfactory
hybrid OSS development relationship. The third hypothesis is framed as:
H3: Perceived opportunistic behavior among parties involved in hybrid OSS
development alliance will have a negative influence on open source software
projects satisfaction with the relationship.
One of the value-added advantages of building alliance is leveraging each partys
efficiencies to create a product larger than the sum of individual efforts. Some of the
common activities shared between an OSS project and commercial organization alliance
include product design, technology development and evangelism. Such activities represent
mutual efforts to promote project and product distinctiveness. Each party's endeavor is
driven by their voluntary assessment of leveraged competencies that yield remarkable and
efficient outcomes. By sharing their value-creation activities, the parties engagement in
promoting product exceptionality would lead to an efficient alliance and guarantee a
rewarding relationship. The fourth hypothesis of the model is delineated as:
H4: Creation of product distinctiveness will have positive influence on open source
projects satisfaction with the alliance.
4.2 Pilot Study
To insure validity of theoretical model constructs and underlying variables, a pilot study
was conducted during the early phase of the research. A small set of open source software
project managers were contacted and solicited for participation in the pilot study. Each of
these project managers was asked to indicate whether their project has commercial
partner engagement. Based on the individual responses, each was notified whether they
qualify for the study. The inclusion principle was determined based on whether those
project managers experienced a formed relationship with a commercial partner. To
preclude responses based on perception only rather that genuine experience; subsequent
activities of the study were exclusively limited to participated members of open source
software projects, where at least one commercial partner is/was involved.
The study employs pre-validated instruments. No attempts were made to develop new
measures for the constructs. Instead, the pilot was aimed to test the appropriateness of the
customized instrument to suit the context of the OSS realm. Given the novelty of empirical
treatment of TCE to open source software, careful examination and interpretation of the
theory constructs within the OSS domain was a primary intent of the pilot study.
A beta version of an online survey was sent to a small group of open source project
managers to obtain feedback on the instrument face validity, quality, and clarity of items.
Follow up phone interviews were conducted with a subset of the pilot study participants.
Those who indicated willingness to provide feedback on IT vendors involvement in their
projects were the primary contacts. Each interview lasted between 20-45 minutes. Based
on the outcome of the interviews and the pilot study, rewording and ordering of questions
was applied to the final version of the instrument.
4.3 Data and Method
4.3.1 Study Sample
Building on the pilot study results, a large scale data collection task was launched. The
final survey instrument is presented in appendix A. The study targeted sampling included
open source projects listed on three major open source repositories, namely, SourceForge,
Freshmeat, and Savannah. OSS transparent development process is manifested on these
public domain hosting sites. An OSS repository is an integrated workbench for source code
versioning, bug tracking, mailing lists and discussion forums. Information about project
information, activity, and latest release are also provided by an OSS repository for both
developers and users of the software. A comprehensive overview of OSS project could be
developed based on information derived from these hosting websites.
There are several smaller or dedicated hosting sites for open source projects; however, the
three identified virtual collaborative development hosting sites represent assorted
categories of OSS projects with diverse user bases and community support.
The study data collection method is a common practice in experimental research on TCE
(Richman & Macher, 2006). Positivist survey research and semi-structured interviews are
the primary methods for data collection. During the data collection phase, personalized
emails with an online survey link and an invitation letter were sent to project managers and
administrators of different open source projects hosted on SourceForge, Freshmeat, and
Savannah. Ascribed to the key informant methodology, those individuals are assumed to
be the most knowledgeable people about the project and typically they earn such status
through quality contributions and lengthy attachment to the project (Raymond, 2001).
The data collection efforts acknowledge the fact that not all initiated OSS projects takeoff
and there are many dead-beat projects that are still hosted online. To overcome
predicaments caused by solo developer/user projects and inactive or abandoned projects
listed on these sites, a systematic filtering criterion was implemented. Selected projects
Has to have the following attributes: a) project has active status, b) project activity is
ranked at the 80th percentile or higher, c) project has at least five developers working on
the project, and d) project maturity is greater than or equal to three years. Project maturity
is defined as the date of inception on the hosting development site. In addition, each
project had at least one designated project manager responsible for coordinating project
activities. The number of project managers is usually a good indicator of project activity.
Typically, moderately active projects have one or two project managers, and highly active
and more complex projects involve several project managers. Project managers are
primarily responsible for coordinating development activity and resolving conflicts regarding
coding and feature design.
A total of 1437 survey invitations were sent out. Of the 340 who responded to the invitation,
116 respondents indicated belonging to OSS project with commercial partner involvement.
After eliminating incomplete responses, the final dataset has 101 points, each representing
a different open source project. The response rate was 23.7%.
Due to the non-traditional nature of the relationship between open source software project
and commercial organizations, prior identification of projects that have organization
sponsorship or other form of alliance with commercial partners is not feasible. Also
limitations of the OSS hosting sites mailing lists used do not allow for identification of
projects that have commercial partners involved. The only way to obtain such information
is surveying the project manager.
Previous studies that employed transactions cost economics are heavily skewed towards
adopting a single case study methodology for investigating variations of discriminating
alignment hypotheses (Richman & Macher, 2006). Since most of the studies were
retroactive in nature (i.e., make-versus-buy or outsourcing decision already made) there
was no difficulty in identifying those organizations where such transactions took place
beforehand. The study selection criteria was developed in an attempt to include as many
projects involved in transactions with commercial partners as possible, but results in a
relatively moderate response rate.
In accord with table 2.1 classification, the data confirmed various categories of
stakeholders involved in open source. The majority of those who responded to the survey
(62%) indicated they work for open source voluntarily. Of the non-volunteers, 15% are paid
by a flagship open source company, 4% are paid by an open source foundation, 8% are
paid by a software packager or hardware/middleware vendor organization, and 11%
designated paid by other, which include any other party excluded from the previous
categories. To insure there are no significant differences in responses between volunteer
participants in a project and those paid by a commercial partner or other organization, a
chi-square test to assess between groups difference was conducted. Statistical test
indicates no significant difference; therefore the two groups were pooled for analysis.
To gain insights of collaboration length between open source project and commercial
organization(s), subjects were solicited to identify their project relation with a commercial
partner. The data show that engagement ranges from less than a year to more than ten
years for some projects. It can be inferred that some alliances between OSS community
and commercial organizations dated back for a decade or so. However, the majority
indicated a relationship length between three and less than four years. The continuation of
the relationship is also an indication of successful establishment of efficient hybrid form
between the parties that bear low cost for the transaction, since it is unlikely that
relationships would continue with experienced deficiencies from either side.
An additional assessed attribute of the hybrid model is demographics. This includes the
number of OSS foundations, OSS companies, IT vendors, or other organizations involved
in a project. The data reveal that OSS projects have between one and more than five
partner organization(s) working in collaboration. It also shows that the GNU General Public
License (GPL) and subsequent GNU General Public License version 3.0 (GPLv3)
dominated across the projects sampled, with over 65% of projects adopted the license. In
addition, some projects adopted a dual license that likely served multiple parties interests.
Dual licensing could be regarded as a strategic tactic that positions the open source project
as an appealing candidate for diverse stakeholders.
4.3.2 Non-response Bias
To account for non-response bias that might be introduced in the data a statistical
significance test was conducted to assess potential impact of the non-respondent group.
The data collection phase was designed to have single round of survey invitation.
Therefore, a traditional non-response assessment method was not feasible. As an option,
this study employs an alternative approach (Armstrong & Overton, 1977) fortesting for
non-response bias effect. The study treatment of late respondents to the survey consider
those who responded to the survey invitation after more than two weeks from the invitation
date as similar to the theoretical non-respondent group.
By conducting a two-tail test (a = 0.05) for the difference between early and late
respondents in the sample, results conclude that there is no significant difference between
the two groups and the non-respondent set is similar to those who participated in the
survey. Therefore, it is reasonable to assume that the non-response bias is minimal or has
negligible effect on the subsequent analysis of the data.
4.4.1 Dependent Variable
To identify the outcome of the hybrid OSS efficiency model, satisfaction with the
partnership was measured. The decision whether to engage a commercial partner in the
project development or maintain the purity of open source model is primarily a process
efficiency decision. Utilizing satisfaction with the partnership as a proxy for realizing
efficiencies that economize on transaction cost is a natural choice for examining the make-
versus-buy (fully open source developed vs. OSS- commercially developed project)
decision in this special setting. The study posits that satisfaction with the hybrid structure is
an indication of achieving a degree of institutionalization that effectively governs the
alliance between the project and commercial partners; thereby, lowering transaction cost
for both sides. If achieved, it represents an outcome of an all-winner scenario.
Satisfaction is measured using three items that capture open source project managers
perceived fulfillment of the project efficiency under joint relationship, as well as the
perceptions of fulfillment of project contributors who are paid by the commercial partner.
4.4.2 Independent Variables
The model includes four independent latent variables, namely, asset specificity,
uncertainty, opportunism, and product distinctiveness. Each latent variable (factor) is
measured through direct observed variables.
Asset specificity construct is operationalized with four observed variables. It corresponds to
the amount of investment in specialized assets dedicated to the partnership between open
source project and commercial partners. It defines parties commitment to the alliance and
the objective to promote an efficient long-term relationship. It also leads to the
establishment of efficient processes for moderating uncertainty, and guarding against
potential opportunism occurrence. The uncertainty construct is estimated using four items
that capture behavioral and environmental uncertainties facing hybrid OSS projects.
The opportunism construct is a seven-item scale that captures the essences of parties
opportunistic behaviors exhibited during the alliance. It assesses the way OSS project
perceived commercial partner activities and actions in the project. Product distinctiveness
is the joint activities by the relationship parties to develop differentiable product and related
activities to achieve such goal. This construct was measured by a four-item scale.
Table 4.1 displays latent variables and their respective observed items. Reliability analysis,
assessed by Cronbachs alpha for each factor, is also included in the table.
Table 4.1 Efficiency Model Constructs Reliability and Measures
Construct Observed Variables Reliability Adopted From
Asset Specificity ASPC1 0.69 (Zaheer &
(ASPC) ASPC2 Venkatraman,
Uncertainty UNCRT1 0.74 (Walker &
(UNCRT) UNCRT2 Weber, 1987)
Opportunism OPRT1 0.92 (Chiou & Shen,
(OPRT) OPRT2 2006)
Table 4.1 (Cont.)
Construct Observed Variables Reliability Adopted From
Product DSTN1 0.80 (Stone-Romero &
Distinctiveness DSTN2 Stone, 1997)
4.5 Data Analysis and Results
4.5.1 Test of Research Model and Hypotheses
This study employs both exploratory and confirmatory statistical analysis. Structural
equation modeling (SEM) technique is used to specify and assess the theoretical model.
SEM analysis is a multi-step analysis that involves compound techniques including:
multiple regression, analysis of variance, and factor analysis. One of the principal benefits
of using structural equation modeling is set by SEMs capabilities of enabling simultaneous
analysis of the complete model variables and causal effect. Having the study objective of
realizing efficiency drivers in hybrid OSS model and being able to examine multiple factors
effects on project efficiency concurrently, rather than in several individual regression
equations, strengthens the overall predictive power of the model.
Moreover, the hypothesized model seeks to validate support for each hypothesis. SEM
lends itself as capable of hypothesis testing by allowing for nullification of hypotheses and
eventually rejecting unsupported ones.
The statistical analysis and assessment is carried out using maximum likelihood (ML)
estimation procedure of LISREL 8.8 (Joreskog & Sorbom, 1996) for simultaneous
multivariate regression computation. The rationale for adopting this technique stems from
efficient operation carry out of covariance matrix analysis (Byrne, 1998). Also ML is
considered both scale free and scale invariant, by being considerably robust at handling
violations of the multivariate normality assumption (Kline, 2005). Likewise, SEM review and
recommended approach indicates that the technique is robust with small sample size and
still holds even with sample size of 50 (J. C. Anderson & Gerbing, 1988).
4.5.2 Measurement Model
As part of model specification, a confirmatory factor analysis was conducted to assess
constructs operationalization. In addition, factor score loadings were assessed to insure
that each measured item is measuring a single construct and inter-factor loadings are
negligible. The measurement model factor score regressions are presented in table 4.2. In
addition, the computed mean, standard deviation, and correlations of measured variables
are depicted in table 4.3.
The factors correlations between the five constructs (ASPC (asset specificity), UNCRT
(uncertainty), OPRT (opportunism), DSTN (product distinctiveness), and STSF
(satisfaction)) in the model are calculated. The model maintains the assumption that zero
correlations between errors (e,) and residuals (reSj) for each item exist. The measurement
model is illustrated in figure 4.2.
Table 4.2 Efficiency Measurement Model Factor Score Regression
DSTN1 DSTN2 DSTN3 DSTN4 UNCRT1 UNCRT2 UNCRT3
DSTN 0.19 0.60 0.02 0.13 -0.01 0.00 0.00
UNCRT -0.01 -0.03 0.00 -0.01 0.42 0.20 0.15
ASPC 0.00 -0.01 0.00 0.00 0.01 0.01 0.00
OPRT 0.00 0.01 0.00 0.00 0.01 0.00 0.00
STSF 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Table 4.2 (Cont.)
UNCRT4 ASPC1 ASPC2 ASPC3 ASPC4 OPRT1 OPRT2
DSTN 0.00 0.00 0.00 0.00 0.00 0.00 0.00
UNCRT 0.13 0.01 0.01 0.00 0.00 0.01 0.00
ASPC 0.00 0.12 0.30 0.09 0.10 0.00 0.00
OPRT 0.00 0.00 0.00 0.00 0.00 0.11 0.09
STSF 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Table 4.2 (Cont.)
OPRT3 OPRT4 OPRT5 OPRT6 OPRT7 STSF1 STSF2 STSF3
DSTN 0.01 0.00 0.00 0.00 0.00 0.01 0.02 0.00
UNCRT 0.02 0.01 0.01 0.01 0.00 0.01 0.01 0.00
ASPC 0.01 0.00 0.00 0.00 0.00 0.00 0.01 0.00
OPRT 0.32 0.13 0.11 0.13 0.05 -0.01 -0.02 0.00
STSF 0.00 0.00 0.00 0.00 0.00 0.33 0.64 0.02
Table 4.3 Efficiency Measured Variables Mean, Standard Deviation, and Correlation
F o 1 2 3 4 5 6
1. DSTN1 5.83 1.393 1.000
2. DSTN2 5.87 1.180 .771" 1.000
3. DSTN3 4.16 1.906 .232 .294" 1.000
4. DSTN4 5.41 1.328 .578" .697" .429" 1.000
5. UNCRT1 3.38 1.548 -.156 -.094 -.200" -.163 1.000
6. UNCRT2 3.11 1.599 -.113 -.083 -.153 -.049 .549" 1.000
7. UNCRT3 3.51 1.507 -.073 -.047 -.178 -.050 .465"" .458"
8. UNCRT4 4.07 1.576 -.200" -.140 -.197' -.195 .481 .338"
9. ASPC1 3.42 1.602 -.080 -.194 .116 -.118 .013 -.014
10. ASPC2 2.88 1.620 .000 -.092 -.010 -.084 .158 .109
11. ASPC3 3.28 1.632 .117 .019 .002 -.025 .204" .207"
12. ASPC4 3.61 1.827 .108 .065 .176 .078 .048 .045
13. OPRT1 2.74 1.604 .088 .014 -.013 .012 .221" .108
14. OPRT2 3.15 1.602 .114 .084 .045 .028 .134 .115
15. OPRT3 2.30 1.446 .020 -.036 .092 .015 .106 .142
16. OPRT4 2.43 1.374 .053 -.021 .054 .025 .032 .056
17. OPRT5 2.73 1.599 -.007 .019 .070 .080 .118 .148
18. OPRT6 2.48 1.566 .142 .131 .186 .142 .115 .155
19. OPRT7 3.23 1.574 .182 .231" -.015 .113 .083 .058
20. STSF1 5.73 1.392 .281 .253" .035 .211" -.092 -.131
21. STSF2 5.74 1.405 .254" .245" .004 .207" -.065 -.108
22. STSF3 5.48 1.467 .206" .192 .044 .141 -.093 -.184
Table 4.3 (Cont.)
7 8 9 10 11 12 13 14
7. UNCRT3 1.000
8. UNCRT4 .229' 1.000
9. ASPC1 -.148 .226' 1.000
10. ASPC2 .038 .074 .443" 1.000
11. ASPC3 -.030 .109 .231' .349" 1.000
12. ASPC4 -.051 -.053 .325" .427" .315" 1.000
13. OPRT1 .217' .233' .085 .207' .005 -.137 1.000
14. OPRT2 .134 .178 .038 .161 .168 -.059 .704 1.000
15. OPRT3 .136 .087 .028 .118 .109 -.085 .702" .710"
16. OPRT4 .130 .023 .173 .203' .072 -.037 .649" .584"
17. OPRT5 .087 .166 .095 .150 .040 -.149 .620" .593"
18. OPRT6 .128 .116 .124 .109 .155 -.030 .686" .534"
19. OPRT7 .081 .062 .097 .019 -.025 -.035 .447 " .561"
20. STSF1 .052 .022 -.107 -.059 -.024 .085 -.264" -.305"
21. STSF2 .054 .022 -.125 -.040 .014 .093 -.274" -.298"
22. STSF3 .056 -.127 -.055 .037 .049 .196' -.347" -.354"
Table 4.3 (Cont.)
15 16 17 18 19 20 21 22
15. OPRT3 1.000
16. OPRT4 .746 1.000
17. OPRT5 .757 .603 1.000
18. OPRT6 .767 .719 .654" 1.000
19. OPRT7 .484 .542 .445 .552" 1.000
20. STSF1 -.477 -.233' -.455" -.239" -.291 1.000
21. STSF2 -.484" -.254' -.472 -.244" -.294 .982 1.000
22. STSF3 -.468 -.250' -.453" -.304 -.346 .802 .807 1.000
**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).
Figure 4.2 Efficiency Measurement Model
The measurement model confirmatory factor analysis denotes chi-square (x2i9g) value of
259.5. The model is over-identified with 199 degrees of freedom. The model fit is assessed
using multiple fit indices. The Comparative Fit Index (CFI) and Incremental Fit Index (IFI)
are both 0.94. The Normed Fit Index (NFI) is 0.85; Goodness of Fit Index (GFI) is 0.81, the
Root Mean Square Residual (RMR) is 0.19, and the Standardized RMR is 0.07. A
comprehensive set of the measurement model fit indices is presented in table 4.4.
The squared multiple correlations of the observed variables and standardized regression
weights for the measurement model are displayed in tables 4.5 and 4.6 respectively. As
specified in table 4.5 observed variables' R2 values are relatively high, with the exception
of DSTN3 and ASPC3. Each of these measured items is able to explain at least 30% of the
variance related to their respective construct. Similar observation could also be made
regarding standardized regression weights presented in table 4.6. All items measuring the
five constructs are significant at a = 0.001. The estimates, standard errors, critical ratio,
and P-values for the constructs items are displayed in table 4.7.
Table 4.4 Efficiency Measurement Model Fit Indices
Degrees of Freedom 199
Minimum Fit Function Chi-Square 297.7 (P = 0.00)
Normal Theory Weighted Least Squares Chi-Square 259.5 (P = 0.0025)
Estimated Non-centrality Parameter (NCP) 60.5
90 Percent Confidence Interval for NCP (22.9; 106.3)
Minimum Fit Function Value 2.9
Population Discrepancy Function Value (F0) 0.6
90 Percent Confidence Interval for F0 (0.03; 1.06)
Root Mean Square Error of Approximation (RMSEA) 0.05
90 Percent Confidence Interval for RMSEA (0.03 ; 0.07)
P-Value for Test of Close Fit (RMSEA < 0.05) 0.3
Expected Cross-Validation Index (ECVI) 3.7
90 Percent Confidence Interval for ECVI (3.3; 4.1)
ECVI for Saturated Model 5.1
ECVI for Independence Model 19.7
x'mi for Independence Model with df=231 1921.9
Independence AIC 1965.9
Model AIC 367.5
Saturated AIC 506.0
Independence CAIC 2045.4
Model CAIC 562.8
Saturated CAIC 1420.6
Normed Fit Index (NFI) 0.85
Non-Normed Fit Index (NNFI) 0.93
Parsimony Normed Fit Index (PNFI) 0.73
Table 4.4 (Cont.)
Comparative Fit Index (CFI) 0.94
Incremental Fit Index (IFI) 0.94
Relative Fit Index (RFI) 0.82
Critical N (CN) 84.42
Root Mean Square Residual (RMR) 0.19
Standardized RMR 0.07
Goodness of Fit Index (GFI) 0.81
Adjusted Goodness of Fit Index (AGFI) 0.76
Parsimony Goodness of Fit Index (PGFI) 0.64
Table 4.5 Efficiency Measurement Model Squared Multiple Correlations
Table 4.6 Efficiency Measurement Model Standardized Regression Weights
STSF1 <-- STSF .99
STSF2 < STSF .99
STSF3 < STSF .81
ASPC4 < ASPC .55
ASPC3 < ASPC .47
ASPC2 < ASPC .79
ASPC1 < ASPC .56
UNCRT4 < UNCRT .55
UNCRT3 < UNCRT .57
UNCRT2 < UNCRT .68
UNCRT1 < UNCRT .82
OPRT7 < OPRT .60
OPRT6 < OPRT .83
OPRT5 < OPRT .80
OPRT4 < OPRT .81
OPRT3 < OPRT .92
OPRT2 < OPRT .76
OPRT1 < OPRT .80
DSTN4 < DSTN .75
DSTN3 < DSTN .35
DSTN2 < DSTN .93
DSTN1 < DSTN .82
Table 4.7 Efficiency Measurement Model Estimates
Estimate S.E. C.R. P
STSF1 < STSF 1.00
STSF2 < STSF 0.96 0.10 9.84 ***
STSF3 < STSF 0.57 0.17 3.55 ***
ASPC4 < ASPC 1.12 0.29 3.83 ***
ASPC3 < ASPC .85 0.25 3.43 ***
ASPC2 < ASPC 1.42 0.35 4.03 ***
ASPC1 < ASPC 1.00
UNCRT4 < UNCRT 0.68 0.14 4.75 ***
UNCRT3 < UNCRT 0.85 0.25 3.43 ***
UNCRT2 < UNCRT 1.42 0.35 4.03 ***
UNCRT1 < UNCRT 1.00
OPRT7 < OPRT 0.74 0.12 6.29 ***
OPRT6 < OPRT 1.02 0.11 9.46 ***
OPRT5 < OPRT 1.00 0.11 8.97 ***
OPRT4 < OPRT 0.87 0.10 9.09 ***
OPRT3 < OPRT 1.04 0.10 10.85 ***
OPRT2 < OPRT 0.96 0.11 8.44 ***
OPRT1 < OPRT 1.00
DSTN4 < DSTN 0.87 0.10 8.28 ***
DSTN3 < DSTN 0.57 0.17 3.38
DSTN2 < DSTN 0.96 0.10 9.84 ***
DSTN1 < DSTN 1.00
4.5.3 Structural Model
To test and cross validate the impact of asset specificity, uncertainty, opportunism, and
product distinctiveness on satisfaction (see figure 4.1), a full path analysis was conducted
to test efficiency model hypotheses. The confirmatory factor analysis is intended to assess
independent constructs effects that determine the predicted satisfaction with joint
relationship efficiencies between open source software project and commercial partners.
The specified structural model is illustrated in figure 4.3.
Figure 4.3 Efficiency Structural Model
The structural model fit is assessed by several fit indices. The model chi-square value is
259.5 with 199 degrees of freedom. The Comparative Fit Index (CFI) and Incremental Fit
Index (IFI) are both 0.94. The Normed Fit Index (NFI) is 0.85; Goodness of Fit Index (GFI)
is 0.81, the Root Mean Square Residual (RMR) is 0.19, and the Standardized RMR is 0.07
A compete list of the structural model fit indices is presented in table 4.11.
The analysis results show that all the observed variables measuring each of the five
constructs are significant at the 0.001 level. The estimates, standard errors, critical ratio,
and p-values are displayed in table 4.8. Results also show support for hypotheses three
(H3) and four (H4). However both H1 and H2 are not supported by the data.
The squared multiple correlations of the observed variables and standardized regression
weights for the measurement model are displayed in tables 4.9 and 4.10 respectively. As
shown in table 4.9 the dependent construct's (STSF) R2 value is 0.314. This indicates that
the model is capable of explaining more than 31% of the variance related to satisfaction
with efficiency of the hybrid OSS model.
Table 4.8 Efficiency Structural Model Estimates
Estimate S.E. C.R. P
STSF < ASPC .046 .131 .354 .723
STSF < UNCRT .101 .157 .641 .522
STSF < OPRT -.687 .160 -4.289 **
STSF < DSTN .434 .131 3.304 ***
STSF1 < STSF 1.000
STSF2 < STSF 1.015 .023 44.764 ***
STSF3 < STSF .867 .063 13.762 ***
ASPC4 < ASPC 1.000
ASPC3 < ASPC .720 .210 3.430 ***
ASPC2 < ASPC 1.150 .284 4.051 ***
ASPC1 < ASPC .854 .219 3.892 **
UNCRT4 < UNCRT 1.000
UNCRT3 < UNCRT 1.026 .250 4.101 ***
UNCRT2 < UNCRT 1.281 .285 4.490 ***
UNCRT1 < UNCRT 1.509 .327 4.612 ***
OPRT7 < OPRT 1.000
OPRT6 < OPRT 1.378 .214 6.437 ***
OPRT5 < OPRT 1.356 .216 6.283 ***
OPRT4 < OPRT 1.177 .186 6.327 ***
OPRT3 < OPRT 1.412 .206 6.848 ***
OPRT2 < OPRT 1.295 .213 6.084 ***
OPRT1 < OPRT 1.351 .216 6.256 ***
DSTN4 < DSTN 1.000
DSTN3 < DSTN .652 .200 3.261 .001
DSTN2 < DSTN 1.116 .129 8.672 ***
DSTN1 < DSTN 1.152 .140 8.230 ***