Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Building social Legoland through collaborative crowdsourcing: marginality, functional diversity, and team success
(USC Thesis Other)
Building social Legoland through collaborative crowdsourcing: marginality, functional diversity, and team success
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
Running head: COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
BUILDING SOCIAL LEGOLAND THROUGH COLLABORATIVE
CROWDSOURCING: MARGINALITY, FUNCTIONAL DIVERSITY AND TEAM
SUCCESS
By
Rong Wang
________________________________________________________________________
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(COMMUNICATION)
August 2016
Running head: COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
Dedication
To my beloved family.
To all the great mentors over the years.
To all the dearest friends along this incredible journey.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
3
Acknowledgment
My interest in Openideo and open social innovation networks dates back to a cold
winter in Ottawa Canada in early 2011 when I was working at the International
Development Research Center (IDRC). It was there that I was first inspired by work in
social innovation, leading me along the long road to the collection of data and the
completion of an academic project for my Ph.D. Along this wonderful journey, I relied
on the advice and support of many brilliant individuals whom I wish to thank.
First and foremost, I am grateful and indebted to my dissertation chair, academic
advisor and mentor, Dr. François Bar. Without his guidance and support, this dissertation
would not be possible. I thank him immensely for inspiring me to find the perfect
dissertation topic that combined my research interests in ICTD, organizational
communication, and network analysis, and for pushing me to develop my own ideas and
thinking. I appreciate and admire both his commitment to his students and to the rigorous
pursuit of scholarship. Dr. Bar not only taught me how to be a competent scholar, but
also how to be a better person.
I also want to thank Dr. Janet Fulk, and Dr. Tom Valente, who served on my
dissertation committee. Both of them have inspired me in many ways beyond my
dissertation project. Dr. Fulk has always been a role model, ever since my first day at
USC Annenberg. She encouraged me to engage in critical thinking and to always keep in
mind the big picture of my research. Dr. Valente provided invaluable instruction on
network analysis and helped me to improve several research projects. I also want to
express my gratitude to my other mentors at USC: Dr. Patti Riley, Dr. Peter Monge, Dr.
Aimei Yang, Dr. Tom Hollihan, and Dr. Rebecca Weintraub. In addition, I want to thank
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
4
Anne Marie Campian and Christine Lloreda for all their administrative help in the past
five years.
My time at USC would never be the same without all my other wonderful
colleagues and friends: Wenlin Liu, Zheng An, Li Lu, Jingbo Meng, Jinghui Hou, Nancy
Chen, and Peter Knaack. Without their support, my journey here would not be so full of
color. Thanks for being my academic comrades, travel companion, and hiking buddies.
Outside of the USC community, I benefited from the generosity of many. I will
always have a soft spot in my heart for my networks at National University of Singapore
and IDRC: Dr. Weiyu Zhang, Dr. Leanne Chang, Dr. Millie Rivera, and Dr. Matthew
Smith. A special thank goes to Dr. Evan Due, whose support has helped build my
confidence along this journey. Thanks for always being the first to criticize my work and
provide constructive feedback. More importantly, thanks for putting up with my
frustration and exhaustion and for always being at my side.
The process of completing a dissertation and a doctoral degree is an experience
unlike any other. More than anything, this journey has reminded me how lucky I am to
have an amazing family to rely on for support and encouragement. I am indebted to my
dearest parents, Wang Jichun and Wu Yan, who never have any doubt in their daughter.
Thank you for your unconditional support and love. I am grateful to my grandparents, my
uncles and aunts, my cousins, and my nieces, who make me feel so loved even that I am
thousands miles away from home. Thank you all for believing in me and for standing by
me. I love you all so much!
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
5
Table of Content
List of Tables 7
List of Figures 8
Abstract 9
Chapter 1 Introduction 11
Organization of the dissertation 16
Chapter 2 Collaborative Crowdsourcing as A Model for Solving Social Issues 18
Different models of solving innovation problems 18
Current research on crowdsourcing 21
Crowdsourcing as an open model of social development 25
Chapter 3 Diversity and Collaboration in Crowdsourcing towards
Social Innovation 32
Study 1 Team As a form of collaboration in crowdsourcing 32
Team as a form of Collaboration in Crowdsourcing Social Contests 34
A network perspective of team collaboration in crowdsourcing 36
Study 2 Leveraging Diversity for Team Crowdsourcing Performance 52
Demographic diversity and functional diversity 53
Functional diversity, faultlines, and team success in collaborative
crowdsourcing 58
Chapter 4 Method and Analysis 69
Background of Openideo.com under study 70
Methodology 78
Data collection 78
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
6
Measurement 79
Analysis 83
Chapter 5 Results 86
Characteristics of Openideo collaborative crowdsourcing participants 86
Study 1: Network mechanisms underlying team collaboration in
crowdsourcing 94
Study 2: Functional Diversity, Team Faultline and Crowdsourcing Success 97
Chapter 6 Discussion and Conclusion 101
Building a sustainable open collaboration environment 103
Building successful team through geographic and functional diversity 109
Limitations and Future Research Agenda 114
Conclusion 118
References 121
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
7
List of Tables
Table 1 Summary of Hypotheses and Research Questions in Study 1 ............................. 51
Table 2 List of hypotheses from Study2 ........................................................................... 68
Table 3 Summary of descriptive statistics of .................................................................... 86
Table 4 A list of 34 Openideo Social Challenge............................................................... 93
Table 5 Summary of ERGM - Study 1 results .................................................................. 96
Table 6 Summary of logistics regression models – Study 2 results ............................... 100
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
8
List of Figures
Figure 1 One example challenge on Openideo ................................................................. 74
Figure 2 The system generated user design quotient on Openideo .................................. 77
Figure 3 Region distribution of all the Openideo members .............................................. 87
Figure 4 Country origins of Openideo members .............................................................. 88
Figure 5 Visualization of Openideo team collaboration networks ................................... 92
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
9
Abstract
This dissertation examines crowdsourcing in the context of solving social
development issues, through the lens of team collaboration. The goals are twofold. First,
guided by a network perspective it aims to uncover the team dynamics of collaborative
crowdsourcing. Specifically, it examines how marginality and preferential attachment
influence collaboration patterns among members of a crowd. Second, it investigates how
to leverage benefits of team collaboration through the lens of team diversity. It draws
from the literature on team diversity and Group Faultline theory to investigate what team
compositions can lead to better team performance. With data scraped from a global
crowdsourcing community, Openideo, this dissertation employed large-scale behavioral
data at the individual user level and team level to test the hypotheses. Exponential
random graph modeling (ERGM) was used to analyze user level data in Study 1 and
logistic regression was used to analyze team level data in Study 2.
Results in Study 1 show that certain dimensions of marginality significantly
influence how people choose team members to collaborate and solve social problems
collectively, such as individual’s project evaluation skills, collaboration skills, and their
network positions. Study 1 also shows that there were are influential members who are
more likely to be chosen in teams due to their geolocation in main innovation centers,
experience in creating teams, number of unique connections with others, and
crowdsourcing winning experience in the community.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
10
Study 2 found that demographic diversity measures such as evaluation skills
diversity and winning experience help improve team performance. It also found that team
diversity in terms of geographic diversity does not generate disruptive influences on a
team’s success. However, team diversity from members’ community tenure could
activate team faultlines and lead to lower team performance. A practical implication of
the two studies is that innovating organizations or a crowdsourcing community should
apply certain strategies to integrate newcomers and sustain user contributions to achieve
their social goals. A second implication is that network intervention strategies should be
applied to facilitate fluid collaboration across sub-communities in Openideo. The third
implication is that to build a winning team, we need to be aware that the effect of
diversity is context specific. Discussion on how to leverage the benefits of diversity while
avoiding the disruptive influences that may come along is provided.
Keywords: Collaborative crowdsourcing, marginality, preferential attachment, network
analysis, diversity, team collaboration, open collaboration, social challenges, Openideo,
Exponential random graph modeling (ERGM)
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
11
Carlson’s Law: “In a world where so many people now have access to education and
cheap tools of innovation, innovation that happens from the bottom up tends to be chaotic
but smart. Innovation that happens from the top down tends to be orderly but dumb”
Chapter 1 Introduction
Governments have traditionally been the primary actor in solving social
development problems. However since early 2000s, private sector and non-governmental
organizations have become increasingly more important players, due to their effort in the
implementation of information and communication technology (ICT) policies (Heeks,
2008). In the recent literature, development experts have been calling for a new “open”
and inclusive approach to solving long-standing social problems in society, which can be
inclusive of all stakeholders (Smith & Reilly, 2013). The whole notion of inclusiveness
and open development centers on the argument that everybody can contribute to the open
process of social development given the free distribution of content and the increasing
availability of ICTs. As Smith and Elder (2010) note, “open ICT ecosystems provide the
space for the amplification and transformation of social activities that can be powerful
drivers of development” (p.65).
One model of open development is crowdsourcing, which is defined as “a type of
participative online activity in which an individual, an institution, a non-profit
organization, or company proposes to a group of individuals of varying knowledge,
heterogeneity, and number, via a flexible open call [i.e. announcement], the voluntary
undertaking of a task” (Estellés-Arolas & Gonzalez-Ladron-de-Guerva, 2012, p. 9). This
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
12
study focuses on collaborative crowdsourcing through which self-selected members of
the crowd work together to solve social development problems (Chesbrough & Di Minin,
2014). It conceptualizes the use of crowdsourcing for solving social issues as a model of
open social innovation, which aims to achieve transformative social change. Open social
innovation examples can be found in applying crowdsourcing to address climate change
issues, social inequality issues, and public health issues. Examples of successful cases
include the use of Ushahidi in a political election for preventing voting fraud in Mexico
and crisis mapping in Haiti, the hackathon events hosted by the World Bank and partner
organizations to build open source applications for developing communities to report
water sanitization problems, and the Philippines with respect to its crowdsourcing act
which aims to include citizens in the lawmaking process (Root, 2012).
This dissertation is motivated by three important factors. First, crowdsourcing has
the potential to help address inequality, inclusion, and other development issues to
contribute to greater social goods. However few studies have examined the application of
the crowdsourcing model outside of the business innovation context. This research thus
conceptualizes crowdsourcing as a social innovation model that could be used to
aggregate diverse expertise and experience from members of a crowd to contribute to
solving social challenges.
Second, what drives crowdsourcing to be an effective model for solving social
issues is the socially constructed nature that allows for collaboration and generative co-
creation. This study focuses on uncovering network mechanisms underlying the
collaboration among participants and how collaboration affects crowdsourcing
performance. Furthermore, it addresses how to build a winning team for organizational
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
13
innovation through analyzing the relationship between team diversity and crowdsourcing
success.
Third, the literature on crowdsourcing has relied more on qualitative case studies
to discover what drives success for innovation; but these findings are hard to generalize.
With the availability of longitudinal data and novel computational models, scholars have
been calling for the use of multiple methods to examine the complex dynamic processes
of open collaboration (Brunswicker, Bertino, & Matei, 2015). This study applies network
analysis, and logistic regression to capture the multilevel dynamics of participants (e.g.:
node level, dyadic level, triadic level, team level, and community level), and unpack the
driving factors of performance and sustainability of developing innovative solutions.
This dissertation uncovers underlying collaboration mechanisms that drive team
formation in crowdsourcing, and empirically test what team compositions could lead to
better crowdsourcing performance. To achieve these two goals, Study 1 of this
dissertation draws from the literature on online collaboration and open collaboration to
test two major network mechanisms on collaborative innovation: marginality and
preferential attachment. Hypotheses were developed with respect to how endogenous and
exogenous variables influence the ways individual members of a crowd choose their team
collaborators. Exponential random graph modeling (ERGM) was used to test the
hypotheses on collaboration mechanisms.
The first study of the dissertation contributes to the literature on crowdsourcing
and open collaboration by taking a network approach to examine the potentials of
collaborative crowdsourcing. It focuses on understanding the dynamics between members
of a crowd rather than just attributes or behaviors of isolated individuals. This approach
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
14
demonstrates the value of a network theoretical framework for understanding the
processes of crowdsourcing. The results provide implications on how to form
collaboration with others who possess different levels of knowledge or expertise so we
can better leverage the benefits of openness to solve social development issues.
An additional interest in the dissertation is to uncover how to build a winning
team in collaborative crowdsourcing. This highlights the importance of examining the
relationship between team composition and task outcomes. Study 2 looks at the literature
from open innovation and team collaboration to investigate the effect of team diversity on
crowdsourcing success. The unit of analysis is set at the team level. It is worth
mentioning that Study 2 is built upon the assumption that diverse teams are more likely to
be innovative. Though much of the literature focuses on diversity of a crowd in general,
this study examines the diversity of teams who are engaged in collaborative
crowdsourcing. In particularly, literature on the benefits of diversity is reviewed to
hypothesize how different dimensions of diversity (demographics and cognitive
diversity) could help achieve a better solution at the team level. Furthermore, Group
Faultline theory is used to examine scenarios where diversity could lead to tensions and
divisions, thus generating disruptive influence on team performance in crowdsourcing.
Logistic regression is used to test all the hypotheses on the relationships between
diversity and team success. Study 2 contributes to the literature on crowdsourcing and
open collaboration by empirically testing what diversity factors could determine team
winning. Results show that diversity is valuable but is contextually specific. Implications
are drawn regarding how to build winning teams to collaborate in the context of
crowdsourcing for solving social challenges.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
15
Results from Study1 show that individual members of a crowd who possess
higher expertise are more likely to be selected into teams in collaborative crowdsourcing.
It is also the case that individuals located in marginal network positions (measured by
betweenness centrality) are more likely to join collaboration in crowdsourcing. The
network mechanism of preferential attachment is significant through the effects of
geolocation (measured as being located in major innovation centers or not), team leader
status, and degree centrality. These findings imply that though some influential members
are still the driving force in the network of collaborative crowdsourcing, members with
marginal attributes tend to be selected into teams as they might possess unique skill sets
that could contribute to the process of solving social challenges.
Results from Study 2 show that teams with higher diversity in terms of members’
evaluation scores and winning experience are more likely to win crowdsourcing
challenges. This highlights the value of having people of diverse expertise for solving
social challenges as their collaboration constitutes the generative co-creation through
building upon each other’s ideas. Furthermore, teams composed of individuals from
diverse geolocations are more likely to win. As most social challenges are general social
development problems, individuals located in one country or region could bring their
experience to another location. Therefore, geolocation diversity is important for applying
crowdsourcing to solve open social innovation issues. However, diversity attributed to
team members’ organizational tenure can generate some disruptive influences on team
performance. This suggests that crowdsourcing communities should develop certain
strategies to help newer members of the community to socialize with others.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
16
In summary, the dissertation examines how crowdsourcing can be implemented to
solve social innovation issues such as climate change, social inequality, community
development, and public health concerns. Through the lens of individual members of a
crowd collaborating in teams to engage in generative co-creation, this study uncovers the
collaboration mechanisms underlying the dynamics of crowdsourcing. Furthermore, it
empirically tests how demographics diversity and cognitive diversity at the team level
influence teams’ crowdsourcing task success. Implications are drawn to discuss under
what conditions we can aggregate the wisdom of a crowd and apply the collaborative
crowdsourcing model to contribute to greater social goods.
Organization of the dissertation
The dissertation is organized as follows. Chapter 2 provides a review of the
current literature on crowdsourcing and identifies important gaps, which motivates the
focus of this dissertation. The chapter conceptualizes that crowdsourcing can be applied
as a model for solving social innovation issues, through open models of social production
and collaboration. Chapter 3 examines the literature relevant to the current dissertation. It
begins with theoretical frameworks used in Study 1, i.e. network perspectives of
analyzing team formation and collaboration mechanisms. Drawing from research
conducted on online communities, team science, and open collaboration (i.e. open source
software, wiki, and other peer production endeavors), two network mechanisms –
marginality and preferential attachment, are reviewed to propose hypotheses about what
factors drive team collaboration in a crowdsourcing community. This chapter also
reviews studies related to team composition and team performance for Study 2.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
17
Hypotheses are derived from team diversity literature and also the Group Faultline
theory.
Chapter 4 provides detailed procedures of data collection and sampling. It also
includes measurement and data analysis techniques used in both studies. Chapter 5
presents the results from Study 1 and Study 2. Chapter 6 discusses the contribution of the
two studies to theories and existing literature on crowdsourcing and open collaboration,
and draws practical implications from the results. It also discusses the limitations of the
current dissertation and suggests some directions for future research.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
18
Chapter 2 Collaborative Crowdsourcing as a model for Solving Social Issues
Different models of solving innovation problems
Dunbar (1990) defines two key features of problem solving: 1) a problem exists
when a goal must be achieved while the solution is not immediately obvious; 2). the
processes of problem solving often involve attempting different ways of solving the
problem. He also argues that problem solving consists of four components: 1). an initial
state; 2). the goal state; 3). the actions or operations that the problem solver can use to get
to the goal state; and 4). the task environment that the solver is working on. The complete
set of states constitutes the problem space where problem solvers search for solutions
(Newell & Simon, 1972). It is often the case, however, that problem solvers do not have
the entire problem space represented in their mind when they are solving a problem. One
of the most important aspects of problem solving is to search for a path through the
problem space that will lead to the goal state. Problem solvers will use strategies or
heuristics that allow them to move through a problem space effectively (Dunbar, 1990).
One type of problem solving that has gained scholarly attention is process
innovation related, such as product research and development (R&D), cracking a tough
scientific problem, or evaluating a difficult public issue. Though the detailed mechanisms
of innovation-related problem solving are contextually specific, an innovation process
across industries and sectors share some characteristics. It often starts with the creation of
innovation opportunities, the evaluation of proposed solutions, and the selection of the
most promising opportunity among all the candidates (Terwiesch & Xu, 2008). Some
other innovation models have been studied as well, such as the garbage can processes that
take into account non-rational approaches (Cohen, March, & Olsen, 1972).
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
19
Conventionally, innovative problem solving has been conducted through two
main models. The first model is internal solutions through which tasks are performed by
the employees or experts within an organization. The second model is external
contracting, whereby a problem-solving task is given to a designated contractor. The
recent literature on innovation-related problem solving has looked at “user innovation”
and “open innovation” in exploring how diversity affects the effectiveness of
approaching and solving problems (Page, 2007). User innovation refers to the act of
individuals outside of an organization customizing the product to their own needs and the
organization incorporating the modifications into future mass-market iteration (Brabham,
2013; von Hippel, 2005). Open innovation is an extension of user innovation. It is
defined as embracing openness with external stakeholders to the problem-solving process
(Chesbrough, 2003; Schuurman, Baccarne & Mechant, 2013).
The rise of newer open models of innovation can be understood from the lens of
transaction cost theory, which examines the information and communication costs
involved in market and organizational transactions and ways of minimizing the costs
(Williamson, 1975, 1985). The transaction cost approach focuses on three levels of
analysis when it comes to organizational innovation: the overall structure of the
enterprise; what tasks should be performed within and outside an organization for
efficiency purposes; and how to manage human assets in the organization (Williamson,
1981). With technological advancement, network forms of organizing for innovation
have provided an alternative to market and vertically integrated organizations (Powell,
1990; Powell, Koput, and Smith-Doerr, 1996; Monge & Contractor, 2003). Newer
models of organizational innovations have thus emerged to leverage the benefits of
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
20
networks to reduce information and production costs by linking people and objects
locally and globally without the constraints of traditional national, institutional or
organizational boundaries (Monge & Contractor, 2003; Benkler, 2006; Tsai, 2001).
The openness embedded in networked innovation has been at the center for
crowdsourcing, which has been conceptualized as a problem-solving model.
Crowdsourcing enables an organization to take a function once performed internally by
its employees and/or contracted experts and open up a desired goal state to a crowd
(Brabham, 2008a, 2008b; Howe, 2006; Brito, 2008; Haklay & Weber, 2008; Fritz et al.,
2009). Crowdsourcing is a collaboration model enabled by networked technologies to
solve individual, organizational, and social problems through an open call to leverage the
collective intelligence of a crowd (Estellés-Arolas and Gonzalez-Ladron-de-Guerva,
2012; Brabham, 2012).
The conceptualization of crowdsourcing has the following key elements: an
organization that has a task to be performed; a crowd that is willing to participate in
completing the task for the benefit of organizations; an online platform that affords the
performance of the task, and also the interaction between the community and the
organization; and a governance structure that is a blend of the top-down hierarchical
management process and the bottom-up open innovation process (Brabham, 2013; Howe,
2006, 2008). Crowdsourcing can take different forms. It can take place in the form of
peer-production that the job is performed collaboratively or be undertaken by solo
individuals independently (Howe, 2006). It also varies in terms of tasks that need to be
solved, which could be about generating new ideas, creating a new design, or evaluating
a new product.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
21
Current research on crowdsourcing
Crowdsourcing is not new. However, with the advent of new media technologies,
crowdsourcing opportunities have greatly expanded (Afuah & Tucci, 2012). As an
innovation model, crowdsourcing is now implemented by different types of
organizations. The private sector applies crowdsourcing to solve a diverse set of tasks.
For example, the Canadian mining company Goldcorp released its proprietary geological
data and hosted an open contest to encourage participants from all around the world to
submit proposals regarding where the next 6 million ounces of gold would be found in a
low-performing Northern Ontario mine. It successfully attracted over 1400 participants
and Goldcorp was able to discover 44 new and productive goal targets (Tapscott &
Williams, 2006; Afuah & Tucci, 2012). Dell and Starbucks have also used crowdsourcing
to engage their user communities to suggest, discuss and vote on thousands of new
products and service ideas (Bayus, 2013; Sullivan, 2010). Other notable examples in the
private sector can be found in technology firms such as Microsoft, Samsung, Intel,
Facebook, Apple, Motorola, Sony, Nokia and Google.
Crowdsourcing practices for innovation have also been utilized in the public
sector as an online tool complementary to traditional public participation programs
(Lampe, LaRose, Steinfield, & DeMaagd, 2011). One well-known example is Iceland
crowdsourcing a constitution after its kitchenware revolution (Castells, 2012). Another
example is the Next Bus Stop Design project, which was an online competition to
encourage citizens to submit bus stop shelter designs and to vote on the design of peers to
determine a best proposal (Brabham, 2012). Very recently, New Zealand government
started a national referendum and applied crowdsourcing to attract over 10000 new flag
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
22
designs. Though the citizens voted to stick to their existing flag, the crowdsourcing effort
generated political benefits for its ruling party.
Crowdsourcing is also now being utilized in developing countries for various
purposes. In the Philippines, for example, crowdsourcing has even been legislated with
its own crowdsourcing act in 2013. The World Bank has employed crowdsourcing in its
Hackathon initiative, which was an open call to invite software developers to design
applications and help solve water challenges facing developing communities.
New organizational forms have been born out of crowdsourcing initiatives, such
as technology intermediary organizations which help firms to take advantages of
technology development for innovation (Spithoven, Knockaert, & Clarysse, 2009), or
online crowdsourcing communities (such as Innocentive, Threadless, and OpenIDEO)
that provide a platform for crowdsourcing contests (Brabham, 2012; Piller & West, 2014;
Sundic & Leitner, 2013).
The value of crowdsourcing for organizational innovation has been summarized
in the saying of “utilizing the wisdom of a crowd”, which is derived not from averaging
solutions, but from aggregating them (Surowiecki, 2004; Sundic & Leitner, 2013). For an
innovating organization, when its expertise and absorptive capacity is not sufficient
enough for solving a problem, distant research is required to locate knowledge outside of
its organizational boundary (Chesbrough & Boger, 2014). As a distributed innovation
model that builds upon openness and collective intelligence, crowdsourcing could
transform a distant search into a local search so that the innovating organization would
enjoy the benefits of a distant search without enduring its costs of distant transaction. It
improves the efficiency and effectiveness of problem solving (Afuah & Tucci, 2012).
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
23
However, there are several major gaps in the current literature on crowdsourcing.
First, though research on crowdsourcing has been growing in a variety of disciplines,
most of the literature comes from the disciplines of information science, business and
management and takes on a firm-centric view (Cummings et al., 2013; Keupp &
Gassmann, 2009; Yoo, Lee, & Choi, 2013). The emphasis is on the active involvement of
customers and other business partners in the problem-solving processes (Sundic &
Leitner, 2013), and how crowdsourcing can contribute to innovation in the private sector.
Crowdsourcing has also been applied by individuals to sponsor their entrepreneurship
endeavors, through platforms such as Kickstarter, GoFundMe, and Indiegogo
(Belleflamme, Lambert, & Schwienbacher, 2014; Mollick, 2014; Mollick & Nanda,
2015; Fleming & Sorenson, 2016).
More than a business model, crowdsourcing has the potential to spur public
participation and contribute to greater public goods (Brabham, 2012; Campbell, 2009;
Brito, 2008). Furthermore, effective design principles for the private sector may not be
directly applied in other organizational contexts (Koch, Fuller & Brunswicker, 2011).
Therefore, the scholarship on crowdsourcing should move beyond the heavy focus on
firm contexts and examine the implementation of crowdsourcing in different sectors. In
particular, there is a strong call for research in the context of using crowdsourcing to
solve social challenges (Chesbrough & Di Minin, 2014).
Second, crowdsourcing faces the quantity and quality issues while the literature
has not systematically addressed these challenges. Crowdsourcing contests often generate
a large quantity of contributions, which causes challenges for evaluation (Boudreau,
2010; Blohm et al., 2011). Furthermore, minimal collaboration has been found among
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
24
participants and thus the crowd often fails to offer well-considered feasible solutions that
incorporate multiple perspectives, risks and needs (Schenk & Guittard, 2011; Majchrzak
& Malhotra, 2013). Some scholars suggest that to address quality and quantity issues, a
staged approach should be introduced to crowdsourcing so that the benefits of the crowd
can be disaggregated into phrases (Leimeister et al., 2009; Hutter, et al., 2011). This
generates additional challenges in evaluating contributions from a crowd, particularly in
terms of coordinating tasks and relational management among members of a crowd,
which could take place at the individual user level, dyadic level, and team level (Tsoukas,
2009; Majchrzak & Malhotra, 2013). Given all these concerns, research on
crowdsourcing should place a stronger emphasis on examining the dynamics of
crowdsourcing, i.e. the collaborative processes of crowdsourcing.
Third, the existing literature on crowdsourcing relies heavily on qualitative case
studies to uncover under what conditions an organization would choose a crowdsourcing-
based approach for problem solving and what motivates an online community to perform
a crowdsourced task. These studies are primarily descriptive and suggest only a partial
perspective of crowdsourcing (Adamczyk, Bullinger & Moslein, 2012; Vuculescu &
Bergenholtz, 2014). Therefore, the existing literature fails to conclude what factors lead
to crowdsourcing success. To better examine what factors influence the processes and
outcomes of crowdsourcing, more diverse methodological approaches should be used
(Vuculescu & Bergenholtz, 2014).
Given these gaps, this study focuses on the application of crowdsourcing in
solving social innovation challenges, and examines forms and outcomes of collaboration
during a crowdsourcing process. The ultimate goal of social innovation crowdsourcing is
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
25
to achieve transformative social and environment change through aggregating collective
intelligence from a crowd with diverse expertise. Potential participants in a crowd could
be people who face severe social development issues themselves, or who feel sympathy
about certain societal issues. In the following section, the crowdsourcing approach is
conceptualized as an open development model for people to engage in collaborative
social innovation. Key elements of the model are discussed. Throughout the dissertation,
social development and social innovation are interchangeably used to refer to solving
social development problems through open models.
Crowdsourcing as an open model of social innovation
Applying crowdsourcing to solve social innovation issues is consistent with the
recent open development approach proposed by development experts, given the rise of
network society and advances in ICT access and uses (Castells, 1996, 2006; Benkler,
2006; Smith & Reilly, 2013). According to Castells (1996: 469): “As a historical trend,
dominant functions and processes in the information age are increasingly organized
around networks. Networks constitute the new social morphology of our societies, and
the diffusion of networking logic substantially modifies the operations and outcomes in
processes of production, experience of power and culture.” He further argues that in the
information age, the dilemma for development arises from the fact that social
development is put at the service of globalized informational capitalism (Castells, 1999).
Reilly and Smith (2013) note that Castells’ work on networked society has been
applied in ICT for development literature in two major ways. First, there is a need to
close the digital divide to ensure the mobilization of informational capitalism for social
development (Norris, 2000). Second, informational capitalism needs to be fundamentally
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
26
challenged in order to tackle the root causes of inequality, through the appropriation of
ICTs to create alternative forms of development. Both camps are based on Castells’
argument that access to ICTs is key to inclusion in the network society, which is a
fundamental prerequisite for solving development issues.
However, recent development in the ICTD literature argues the need to move
beyond access, which is no longer the defining feature of development in the information
age (Gurstein, 2005; Smith, Elder, & Emdon, 2011). The focus should be placed on the
way ICTs have worked to restructure social relations of production. Furthermore, given
that Castells’ work focuses on global informational capitalism, his work is less able to
account for the many social innovations that are emerging to take advantages of
networked computer technologies, such as the increasing significance of information
sharing and collaboration (Reilly & Smith, 2013; Smith & Elder, 2010).
Crowdsourcing has the potential to address such social innovation issues through
an open model of social production, as it is built upon networked information economy
which offers new opportunities of knowledge production and distribution and facilitates
bottom-up social changes. The goal of applying crowdsourcing in a broader development
context is to create a more equitable and inclusive capability to problem solving using
science and technology, which helps overcome the typical constraints embedded in
vertical-hierarchical organizations and institutions (Masum, Schroeder, Khan, & Daar,
2013; Singh & Gurumurthy, 2013). As Benkler (2004: 14) explains: “the availability of
free information resources makes participating in the economy less dependent on
surmounting access barriers to financing and social-transactional networks that made
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
27
working out of poverty difficult in industrial economies. These resources and tools thus
improve equality of opportunity”.
As an open model of social production, crowdsourcing for social innovation has a
number of key elements. Specifically, the conceptualization highlights the role
technology plays in open social innovation, the importance of generative co-creation and
also social benefits as the ultimate goal. One thing to note here is that crowdsourcing for
social innovation is a sub-set of crowdsourcing. And these following key elements
identified help to define the application of crowdsourcing in a different context, i.e. to
solve social development issues.
The first key element is that networked technology plays an important role in
achieving social benefits through open access, open content and open collaboration
(Smith & Reilly, 2013). This is supported by the argument from recent literature on ICTD
that networked technologies (e.g. mobile devices, and social network platforms) enable
not only free or low-cost access to information content, but also empower everybody to
actively participate and collaborate to provide insights on human development issues
(Reilly & Smith, 2013; Smith, Spence, & Rashid, 2011; Ling & Horst, 2011). As Bar and
Riis (2000) argue, “universal access to advanced networks would serve to cultivate a vast
pool of lay users who have the potential to make very meaningful contributions to the
innovation process” (p.103). Therefore, crowdsourcing is a socio-technical system, which
includes not only communication processes and individual capabilities, but also social
openness towards a common goal. Applying crowdsourcing to solve social issues entails
more inclusion and the achievement of networked innovation for social goods. One
interesting example is the United Nations has now started an online program called
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
28
“2015” to encourage people from across the world to use digital media and mobile phone
technology to participate stetting the next generation of anti-poverty goals (Kjorven,
2013).
A second element of crowdsourcing for social innovation is generative co-
creation within which individual participants collaborate and build upon each other’s
ideas to come up with better solutions (Majchrzak & Malhotra, 2013; Tsoukas, 2009). As
discussed earlier, this element is essential for solving innovation issues that requires input
from diverse sources. It involves a series of interactions and collaboration in which
crowdsourcing participants offer alternatives, and collectively modify and improve
examples, data, and proposals in order to co-create solutions which would not surface if
one single perspective was presented (Majchrzak & Malhotra, 2013). This highlights the
importance of team collaboration towards the collective goal imbedded in a
crowdsourced task for greater social goods. An excellent example is Openideo, a global
crowdsourcing platform for people to work together in teams to design solutions for the
world’s biggest social development problems, such as public health issues, social
inequality, climate change and other environmental concerns.
A third key element is that crowdsourcing for social innovation has social
development benefits as the ultimate goal, which highlights not only process but also the
end product of innovations (Chesbrough & Di Minin, 2014; Smith & Reilly, 2013). In the
context of open social innovation, crowdsourcing should be defined to provide an
opportunity to achieve social benefits in a manner that does not take place otherwise.
Combined with the other two elements, the implementation of crowdsourcing to solve
social issues challenges the traditional development model of donor-recipient frameworks
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
29
and does not necessarily rely on the one-way contribution of development experts to
solve social problems. This type of development model through utilizing the wisdom of a
crowd provides “an important counterweight to the neoliberal Washington consensus, as
well as later efforts to soften it” and a degree of freedom and flexibility “for designing
development-oriented interventions without strong dependencies on either markets or
states” (Benkler, 2013: viii-ix). Therefore in evaluating such crowdsourcing efforts the
feasibility and real social impact should be taken into account.
The key elements of crowdsourcing model for solving social challenges help to
clarify its difference from other forms of crowdsourcing applications, such as firm-
sponsored new product R&D (Cummings, Daellenbach, Davenport, & Campbell, 2013),
distributed human-intelligence tasking, and crowdsourced art projects (Brabham, 2012).
They also highlight the importance of analyzing under what conditions, collaboration
among crowdsourcing participants will take place and contribute to the ultimate goal of
social development.
Drawing from literature on open collaboration and online collective action, there
are several reasons why people want to participate in crowdsourcing social innovation.
First, given that the ultimate goal for applying crowdsourcing to solve social challenges is
to achieve social development outcomes, participants are driven by pro-social motives. In
open collaboration contexts such as Wikipedia and other peer production communities,
altruism has found to be a major pro-social motive (Osterloh & Rota, 2007; Nov, 2007;
Antikainen & Vaataja, 2010; von Krogh, Haefliger, Spaeth, & Wallin, 2012). Participants
of crowdsourced social contests are collaborating towards a common goal through
voluntary contribution (Budhathoki & Haythornthwaite, 2013; Nam, 2012).
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
30
Second, another reason is that complex social issues can be divided into small
tasks and thus the threshold for contributing is quite low. In contemporary media
environment rich of communication and information, contributing to public goods is no
longer a discrete decision of whether or not and it can take different forms (Bimber,
Flanagin, & Stohl, 2005; 2012). Making contribution to an open social innovation
challenge could take the following forms: simply reviewing a post to boost its
viewership, engaging in a one-time discussion with others of similar interest, ad-hoc
editing of an idea, and spontaneous comments to a proposal. Though in some cases, a
certain level of commitment is required, the cost of making a contribution to a social
cause is relatively low.
Third, there can be some tangible benefits to participants who volunteer to
contribute to open social innovation. These benefits constitute selective incentives in
collective action literature, which could be material self-interests, solidarity benefits
which arise from social interaction with peers, and purposive incentives like self-
satisfaction (Olson, 1965; Marwell & Oliver, 1993; von Hippel & von Krogh, 2003).
These benefits are private outcomes to reward contributors. Research on open
collaboration in the Linux community and Wiki communities has shown that getting
attention in a virtual community and seeking potential employment opportunity are major
selective incentives (Weber, 2004; Benkler, 2006). These benefits also apply to the
context of open social innovation. Some other studies on crowdsourcing identified other
motivational factors such as gaining peer recognition (Hargadon & Bechky, 2006; Lerner
& Tirole, 2002), and improving skills (Brabham, 2011; Benbya & Belbaly, 2010).
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
31
Innovating organizations that implement crowdsourcing to solve social challenges
can be corporations, government organizations, non-governmental organizations, or inter-
government agencies. The motivations for these organizational actors can be under the
notion of corporate social responsibility (Bulnes, 2011), branding (Schleich & Prell,
2015), developing alterative business models (Chesbrough, 2006), or contributing to open
development through networked innovation (Smith & Reilly, 2013). In some cases,
organizations build alliances with each other to create an open call together and provide
sponsorships to crowdsourcing challenges.
In the next chapter, theories related to team collaboration and open collaboration
are reviewed to examine how collaborative crowdsourcing takes place, what are the
underlying collaboration mechanism in crowdsourcing towards social goods, and what
factors influence crowdsourcing success at the team level.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
32
Chapter 3 Diversity and Collaboration in Crowdsourcing
Towards Social Innovation
Study 1 Team As a form of collaboration in crowdsourcing
Crowdsourcing technologies bring changes not only to how project workforce is
sourced, but more importantly to the ways that workforce is organized, coordinated, and
engaged in collaboration (Dissanayake, Zhang, & Gu, 2014). The theoretical basis for
crowdsourcing being generative of innovation is the value of expertise diversity from the
crowd (Majchrzak & Malhotra, 2013; Surowiecki, 2004). To incorporate the benefits of
functional diversity into innovation processes, collaboration among participants should be
encouraged. Collaboration-based crowdsourcing occurs when self-selected members of a
crowd work together to solve a problem and the result is one solution from the collective
(Cullina, Conboy, & Morgan, 2015). In crowdsourcing, collaboration and competition
often co-exist at various development stages of a project (Park, Son, Lee, & Bae, 2013)
One important form of collaboration in crowdsourcing is through team building.
A crowdsourcing community can be viewed as a social network of contestants who work
in teams competing against each other and/or work independently on their own. To solve
innovation tasks, any member of a crowd can take the initiative to set up a team and
recruit team members. To unlock the potential of collective intelligence, teams do not
form organically. Rather, they are organized. For example, teams for design projects are
assigned in a way to ensure diversity by mixing people of different skills (Sawney, 2003).
Evidence was also found in product development teams that there are underlying
mechanisms that affect whom people choose to work with (Bell, Villado, Lukasik, Belau,
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
33
& Briggs, 2011). Furthermore, literature on muilti-player games also found that members
of different skills tend to form teams together to increase their chances of winning
(Thomas & Brown, 2009).
However in an open contest environment, individual styles of generating content
are still prevalent (Zhao & Zhu, 2014). Examples have been found in Threadless,
iStockPhoto, and TopCoder. Vukovic (2009) points out that most of the existing
crowdsourcing systems fall short on facilitating collaboration mechanisms, due to a lack
of a flexible and proactive team discovery and building mechanisms. Nevertheless, the
importance of collaboration between solvers has been increasingly emphasized (Bullinger
et al., 2010; Hutter et al., 2011) and some even argue that crowdsourcing will become
more focused on collaboration (Rajala, Westerlund, Vuori, & Hares, 2013). So far, no
studies have examined the team dynamics in collaborative crowdsourcing from a
communication perspective. More specifically, little is known regarding the structure of
collaboration among crowdsourcing participants who work towards a common goal to
contribute to public goods. However, it is crucial to understand what drives team
dynamics and uncover the relationship between team collaboration and crowdsourcing
performance. Furthermore, it is important to uncover what team compositions could lead
to better crowdsourcing performance. Therefore, the goal of this first study is to examine
what factors drive members of a crowd to form team collaboration with whom so they
can better utilize the benefits of networked participation, and what factors influence the
task outcomes at the team level.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
34
Team as a form of Collaboration in Crowdsourcing Social Contests
The values of joining teams to participate in crowdsourcing contests have been
discussed in recent studies on crowdsourcing and open collaboration. Rokicki, Zerr and
Siersdorfer (2015) argue that team mechanism can be leveraged to improve cost
efficiency of crowdsourcing competition, as well-organized teams can shorten the time
required for completing a certain task by dividing the workload among team members.
They also found that team collaboration, which builds upon diverse expertise, allows for
the solution of more complex tasks. Kelley and Johnston (2012) suggest that social
motivation of team membership helps boost individual’s activities to produce quality
design outcomes, due to a sense of community building.
However, team-based collaborative crowdsourcing could also face some major
challenges, such as potential miscommunication, lack of trust, attrition inside a team, and
even lower quality of work (Dissanayake, Zhang, & Gu, 2014; Lykourentzou, Antoniou,
& Naudet, 2015; Skopik et al., 2010). Due to the lack of deep ties and lack of common
experience in learning from each other, working in teams could end up having higher
level of relational conflict (Mohammed & Angell, 2004). Furthermore, team members
who represent different knowledge domains need to find an effective way of moving
knowledge across boundaries (Carlile, 2002).
The value of joining teams to participate in crowdsourcing for social innovation is
grounded in the argument that social challenges are often interdisciplinary complex
problems. One successful example is the use of collaborative open source biotechnology
platforms in developing communities, which helps to improve crop yield or advance
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
35
medical discoveries for marginalized communities (Masum, Schroeder, Khan, & Daar,
2013). Another example is the Ebola challenge on Openideo, where individuals from
both developing and developing communities collaborated to participate in critical
knowledge gathering, and generate ideas on how to improve healthcare delivery and
restrict the spread of the disease global wide. Through team-based collaboration, both
projects were successful in generating innovative solutions.
As a form of virtual collaboration, teams in crowdsourcing can be defined as a
temporary work group composed of culturally diverse and geographically dispersed
individuals (Jarvenpaa & Leidner, 1998). These virtual teams are not unique, as they can
be found different contexts of open collaboration. However, they differs from traditional
forms of teams in the sense that at least one team member is located in a different
location and the team relies on networked technologies for task coordination and
communication. Organized around specific challenges, these teams have narrowly
defined goals and agenda. In a way, they function like task forces that keep regular
deadlines with a limited lifetime (McGarth, 1984).
Virtual teams in the context of collaborative crowdsourcing have their own
features. First of all, there are elements of both competition and collaboration, which
motivate people to join more than one team to increase winning chances (Park, et al.,
2013). Team membership thus is not exclusive. Second, members are often self-
organized without a hierarchical structure (Rokicki, Zerr, & Siersdorfer, 2015). In other
words, anybody can create a team and recruit others from the crowd to join and
collaborate. There is no necessary division of labor among team members. Third,
membership is voluntary and ad-hoc, without formal membership requirements (Scekic,
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
36
Truong, & Dustdar, 2015). With fluid team boundaries and temporary membership,
anybody can join a team at any time and contribute in any form, either through
submitting a post or providing feedback on other’s ideas (Majchrzak, More, & Faraj,
2012).
A network perspective of team collaboration in crowdsourcing
The structure of team collaboration in an open environment can be examined from
a network perspective (Benkler, 2006; Rainie, & Wellman, 2012). A network perspective
has demonstrated its value in understanding collaboration patterns in contexts such as
open source software communities (Shen & Monge, 2011), Wikipedia communities
(Keegan, Gergle, & Contractor, 2013), and other peer production communities (Cheliotis
& Yew, 2009; Wang & Cheliotis, 2016). Focusing on relations among actors, a network
can be defined in terms of two dimensions: nodes (i.e. actors), and links (i.e relations
between all the actors) (Hanneman & Riddle, 2005).
In a crowdsourcing community, the nodes are members from a crowd who joined
at least one team to participate in crowdsourcing contests, and the links are team
affiliations between any pairs of team participants. As Dahlin (2011) suggests, innovation
operates on networks of ideas and networks of people. The network perspective enables
us to analyze the dynamics of collaborative crowdsourcing between team members,
rather than just the attributes or behaviors of the nodes or projects.
Networked benefits of team collaboration can be explained by the concept of
generative co-creation, which is a foundational requirement for innovation from diverse
sources (Carlile, 2002; Tsoukas, 2009). Generative co-creation is defined as a series of
interactions in which different assumptions and perspectives are discussed to surface and
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
37
resolve critical tradeoffs that were unresolvable previously (Carlile, 2002; Leonard-
Barton, 1995; Majchrzak et al., 2012; Tsoukas, 2009). Generativity is the ability, power,
or function to originate, produce or propagate something (Avital & Te’eni, 2009; Weick,
2007). Novel approaches of solving social issues require the cross-fertilization of ideas
across intellectual and organizational boundaries (Dahlin, 2011). Through connecting
with each other and exchanging ideas in teams, team participants are not just processing
or transferring knowledge but transforming the way knowledge is created and shared
(Carlile, 2002). Through the generative co-creation, team collaboration could help
address the quality issues crowdsourcing faces.
However, studies have not fully examined what drives the process of generative
co-creation in collaborative crowdsourcing. More specifically, what are the network
mechanisms underlying crowdsourcing team collaboration? A crowdsourcing platform is
not a given socio-technical system that would automatically enable generative co-
creation, but rather should be viewed as a shaper, which under certain network
mechanisms optimizes open innovation solutions (Majchrzak & Malhotra, 2013). The
following sections will review two different mechanisms underlying team collaboration
in crowdsourcing: marginality and strategic selection.
Marginality as a mechanism of forming team collaboration
Literature on crowdsourcing has examined the relationship between marginality
and problem-solving effectiveness. For example, Jeppesen and Lakhani (2010) found a
positive and significant relationship between the chance of an external problem solver
winning a crowdsourcing contest and the solver’s self-assessed technical expertise
distance from the focal problem field. With a focus on intra-firm crowdsourcing,
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
38
Villarroel and Ries (2010) also found that contributors with lower positions in the
corporate hierarchy and those who were located farther away from the corporate
headquarters were more likely to perform better with innovation tasks.
The effect of marginality on effective innovation rests on the assumption that
problem solvers on the margin have the advantage of not being burdened with
organizational rules and can bring in new perspectives and heuristics to solve the
problem. The value of marginality in open social innovation is more compelling given
that such innovation often addresses the needs of under-served population and aims to
reduce information asymmetry and exclusion (Chesbrough & Di Minin, 2014). However,
marginality does not necessarily lead to better performance in solving social issues. The
reason is that social challenges are intertwined with existing power issues in the society
(Smith & Reilly, 2013).
In order for open social innovation to succeed through crowdsourcing, the
benefits of marginality need to be incorporated into the collective level to prompt
collaboration towards public goods (Surowiecki, 2004; Yuan, Fulk, Monge & Contractor,
2009). This requires other members in a community to be aware of who the marginal
members and the also need for diversity. As Tortoriello, McEvily, and Krackhardt (2014)
argued, it is important for a community or organization to have access to diverse sources
of knowledge, but also equally important that numerous members have the awareness of
who needs or possesses what types of knowledge. In the context of innovation, evidence
has been found that organizational members actively seek for knowledge diversity
(Singh, Kryscynski, Li, & Gopal, 2015; Sorenson, Rivkin, & Fleming, 2006).
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
39
Individuals who are located in the margin are viewed as catalysts of innovation
(Tortoriello et al., 2014), as they can provide alternative views for problem solving, and
enable innovation participants to build upon each other’s contributions for a better
solution. Therefore, this study conceptualizes marginality as a team feature that affects
team collaboration patterns in crowdsourcing. It is through marginality that dyadic level
collaboration takes place to drive the crowdsourcing to higher levels of innovation to
achieve organizational effectiveness (Hong & Page, 2001; Malhotra & Majchrzak 2014;
Boudreau, 2012).
The literature has identified four major processes by which individuals arrive at
marginal positions. The first relates to technical expertise; the second relates to social
attributes; the third relates to ranking in an organization or a community; and the fourth
relates to location. In this study, the four dimensions of marginality will be measured to
examine what drives collaboration patterns and task success in open social innovation.
First, technical marginality is defined as being distant in terms of technical skills
in solving innovation issues (Jeppesen & Lakhani, 2010). In collaborative crowdsourcing,
we would expect people of different levels of technical expertise to work together. The
heterogeneity of expertise indicates the diversity in knowledge domains and opinions,
which is the source of creativity (Erickson et al., 2012; Schenk & Guittard, 2011).
Buecheler et al. (2010) argue that the theoretical basis for crowdsourcing being
generative of innovation is the value of expertise diversity, which could trump ability.
The value of having some individuals with lower technical expertise as team
members lies in that they may come from a field that is distant from the focal field of the
crowdsourced social challenge, and they are not closely affiliated with taken-for-granted
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
40
perspectives or existing problem solving models. Therefore they are well positioned to
offer novel solutions or heuristics that are not present in existing landscape (Fernandes &
Simon, 1999). In crowdsourcing for social innovation, there is no easy answer to complex
social challenges. As Jeppensen & Lakhani (2013) argue, in an open collaboration
environment, individuals at the technical margin are able to access previously
inaccessible innovation issues and display their willingness to employ their different
perspectives on a given challenge to offer successful solutions. Given that technical
marginality might be conductive to crowdsourcing success, they will be viewed as
potential collaborators in a crowdsourcing community. People at the technical margin
thus will be actively pursued as collaborators by other members in a crowd. In this study,
it is proposed:
H1: Members of a crowd on the technical margin will be more likely to be selected into
teams.
The second dimension is social marginality, which is defined as being distant
from the core establishment, i.e. being socially excluded from one’s own professional
community (Jeppesen & Lakhani, 2010). There are significant disadvantages to being
marginal in a social context, such as limited access to information and resources,
isolation, and lack of social support (Wellman, 2007). However, a potential advantage of
social marginality is the possibility of having alternative and novel views of problem-
solving solutions (Jeppesen & Lakhani, 2010; Page, 2007). Being socially distant in a
community allows individual members to possess alternative knowledge and approaches
of problem solving. In the literature, the concept of social marginality has been tested
with gender. For example, women who are considered as in “the out circle” of an
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
41
established scientific field are an untapped talent pool. They are more likely to provide
high-quality solutions for broadcast search, which aims for generating ideas (Etzkovitz et
al. 2000).
This study focuses on social marginality defined by network position. The reason
is that being included in a network but occupying periphery roles shows that individuals
on the margin are making an effort to follow socializing rules and also keeping up to date
with activities. This balance of inclusion and partial exclusion allows them to view and
approach a social challenge in a more unconventional way but at the same time also
possess the ability to connect with others.
One important network attribute to measure social marginality is reversed social
ranking or betweenness centrality, which measures the frequency a node lies on the
shortest path connecting all other nodes in the network (Freeman, 1979). Betweenness
centrality indicates the extent to which a node occupies a strategic position to bridge
different cliques, which carries social capital through structural holes (Burt, 1992, 2009).
In the context of team collaboration towards social innovation, participants who occupy a
strategic position have better access to strategic contacts and other relational resources,
and they can function as gatekeepers in a community even if the volume of contacts may
not be large (Grewal, et al., 2006; Valente, 2010). However, potential productive
marginal nodes might be capable of analyzing and approaching a problem in a novel way
and complement how community gatekeepers view it. They might also be on the only
path that connects to certain sub-communities in a network. Selecting individuals with
lower betweenness centrality scores could help reduce information redundancy and may
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
42
provide access to unique information that could not be reached through central nodes in a
network. Therefore,
H2: Members of a crowd on the social margin will be more likely to be selected into
teams.
The third dimension is rank marginality, defined as being lower positioned in the
community hierarchy (Villarreol & Reis, 2010). Literature on crowdsourcing has looked
at different types of labor divisions in intra-firm crowdsourcing, such as technicians and
support staff as low rank, professionals and specialists as medium rank, and managers,
executives and directors as high rank (Villarreol & Reis, 2010). However, the concept of
rank marginality has not been tested in other crowdsourcing settings. In a crowdsourcing
community, which is a form of virtual community, community reputation is often
measured by looking at membership history (Shen & Monge, 2011). Members of lower
rank can be defined by the duration they have remained in a community. Newer members
of a crowd are less experienced in socializing with their peers in the community. They
will be more motivated to engage with others and find more effective ways of getting
their ideas noticed. With fresh knowledge about the community, they may hold
alternative views of how to solve a challenge. Their input combined with more senior
members’ experience would facilitate the innovation process. Members of a crowd will
be more likely to select these individuals of less tenure as they can bring in unique views
to a team and create additional value in the generative co-creation process. Therefore,
H3: Newer members of a crowd will be more likely to be selected into teams.
The last dimension is geographic marginality, defined as being spatially distant
relative to the innovation center. This could be the crowdsourcing organization’s
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
43
headquarter office, or where the majority of the crowd is located (Villarreol & Reis,
2010). In an open social innovation community, anybody can have the opportunity to
contribute regardless of their location. Crowdsourced tasks for social innovation are to
solve social development issues, which require knowledge about local communities or
culture for effective solution. Individuals that are located closer to centers of innovation
are more likely to be exposed to traditional innovation processes. Located in global south
or spatially distant from innovation centers provides individuals unique experience about
social development, which could prove valuable in generating novel solutions.
Individuals from a distant location will be more likely to be selected into a team as they
can help collective decision-making by examining an innovation issue from a more
comprehensive perspective. Therefore,
H4: Members of a crowd on the geographic margin will be more likely to be selected into
teams.
Preferential attachment as a mechanism of forming team collaboration
Another important network mechanism that drives team collaboration in
crowdsourcing communities is preferential attachment, which describes the “rich-get-
richer” phenomenon (Barabási, 2002; Powell, et al., 2005). This mechanism indicates that
early achievers are disproportionally favored by reward system, as summarized by the
“Mathew Effect” (Merton, 1968). In a crowdsourcing community where participants are
free to choose from a large pool of potential collaborators, people make strategic
selections to expand their networks (Dissanayake, Zhang, & Gu, 2014). Literature from
other open collaboration context also suggests the same. For example, developers in open
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
44
source software communities tend to interact with experts who have better reputation
(Kuk, 2006), or team leaders (Shen & Monge, 2011).
The rationale behind the network attachment logic of preferential attachment is
that when people make strategic partner selection, they seek to connect with those who
are resourceful in the community. Open collaboration is often viewed as having no
formal governance structures, as collaboration towards a common goal tends to occur
with relatively equitable patterns of self-organized contribution and less hierarchical
organizational forms (Weber, 2004). However, inequalities of participation and deeply
entrenched leadership are found in open innovation communities. Shaw and Hill (2014)
analyzed the participation patterns in a Wiki community and found that as the community
grows to attract more contributors a small group of leaders exercise a monopoly over the
mechanisms of authority and could influence organizational goals in the community.
Though crowdsourcing as a problem-solving model could function with a broad
democratizing potential, there could also be a hierarchy of decision-making or
community influence. Matzler, Strobl, and Bailom (2016) found that leaders in a
crowdsourcing community could increase cognitive diversity in decision-making, access
and aggregate decentralized knowledge, and encourage individuals to contribute their
knowledge. Some even argue that leadership is required to mitigate the negative effects
of virtual team collaboration (Pillis & Furumo 2007). Having leadership is not
necessarily a bad thing for collaborative crowdsourcing. Having some degree of oversight
helps ensure productivity of collaborative work through governing how individuals and
teams interact and engage towards creativity (Boughzala et al., 2012).
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
45
In a crowdsourcing community where collaboration is desired, leadership can be
defined by whether a user has initiated at least one team to recruit others to work towards
a common goal. Though leaders in open collaboration communities do not possess the
same level of power or authority as in traditional organizational settings, leadership in
crowdsourcing indicates the willingness to take responsibility for talent recruitment and
task coordination (Hackman, 2011; Kim, Cheng, & Bernstein, 2014). The role of team
leaders helps to facilitate collective decision-making, which can be achieved through both
process facilitation and content facilitation (Boughzala, et al., 2012). In process
facilitation, team leaders manage how team members communicate and share information
to contribute towards a common goal and contribute to final product indirectly (Anson,
Bostrom, & Wynne, 1995; Griffith, Fuller, & Northcraft, 1998; Bessiere, Ellis, &
Kellogg, 2009); while in content facilitation they help to resolve potential conflicts and
provide structural guidance to the collaboration process (Eden, 1990).
This study focuses on testing whether leadership influences the structure of
collaboration. Literature on open collaboration has found a significant relationship
between team leadership and team dynamics, such as leader-member communication,
how team members allocate and access resources, and team cohesion (Goh & Wasko,
2010; Hoyt & Blascovich, 203). Focusing on one specific dimension of team dynamics,
which is the structure of collaboration, it allows us to uncover to what extent leadership
matters in influencing collaborator selection in crowdsourcing. Though there is often no
clear hierarchy in teams, the role of team leader may indicate a higher level of
commitment to the collective cause or higher popularity. Team leaders in a
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
46
crowdsourcing community will be more likely to be selected into other teams in the
future. Therefore, the following hypothesis is proposed:
H5: Team leaders in crowdsourcing contests are more likely to join or seek collaboration
ties with others.
Another factor of influencing preferential attachment is team members’ network
positions. Network scholars have extensively examined the measurement of centrality to
capture different dimensions of node influence in a network (Valente, 2010). One
dimension of centrality to capture the influence and activeness of nodes is degree
centrality, which refers to the number of links a node has to all the other nodes in a
network (Wasserman & Faust, 1994). Degree centrality has been viewed as an effective
measure of a node’s power potential, which affects their ability to call on resources of the
network as a whole (Hanneman & Riddle, 2006). In a directed network, in-degree
centrality shows the degree of prominence or importance; while out-degree centrality
indicates the level of being active. In an undirected network, there is no need to
differentiate in-degree and out-degree centrality, and degree centrality only refers to the
total number of connections a node has. It captures how influential a node is, which could
be generated by its popularity or activity.
In the context of team collaboration towards social innovation, participants are
positioned in an undirected network. To put in a more specific context, degree centrality
in a team collaboration network indicates how many unique members a node has
collaborated with through different projects. The higher degree centrality, the more
central they are in terms of connecting with other contacts, having access to resources,
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
47
and exchanging information with others. Higher connectivity shows a more advantageous
network position. Therefore, it is proposed:
H6: Members of a crowd with advantageous network positions are more likely to join or
seek collaboration ties with others.
Another factor underlying preferential attachment is the availability of
employment information in a crowdsourcing community. The application of
crowdsourcing in social challenges attracts contributors from different sectors, such as
people who work in government organizations, inter-governmental organizations,
academic institutions, policy tanks, non-governmental/nonprofit organizations, or a self-
employed environment. Though open collaboration often occurs in an anonymous
environment where people do not necessarily know each other, providing certain
information could help endorse your technical expertise or institutional background.
Providing employment information also helps to construct cognitive social
structures, i.e. knowing who knows what in a team. As team literature identifies, one key
challenge in knowledge sharing is how to locate experts and expertise (Brandon &
Hollingshead, 2004). There are different ways of describing individual’s expertise and
skill sets. One way of providing such information is through self-reported employment
data. Individuals who provide employment information are also providing information on
expertise and experience distribution, which is an important cue for division of labor
among team members. Even though providing employment information does not
necessarily equal to the endorsement of expertise, as some members could be lower level
or support staff in an organization, such information could provide a basis for developing
further relationships. Therefore,
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
48
H7: Members of a crowd who provided employment information are more likely to be
selected into teams.
The last factor underlying preferential attachment in a crowdsourcing community
is innovation performance. Working in virtual teams often comes with high uncertainty.
To ensure collaboration outcomes, people tend to work with others with good reputation.
In traditional team environment, individual’s reputation can be evaluated by education,
certificate, and other conventional signals. In open collaboration context, such
information is not readily available and thus people often look for other signals of such as
potential collaborators’ work quality or past performance (Shen & Monge, 2011; Weber,
2004; Benkler, 2006). Therefore,
H8: Members of a crowd with a record on winning crowdsourcing contests are more
likely to be selected into teams.
Core-periphery and transitivity in crowdsourcing collaboration network
H1-H8 test the relational attributes at the dyadic level, i.e. how members of a
crowd select their team members to engage in collaborative crowdsourcing. There will
also be some structures at higher levels in the network. Therefore, the following two
hypotheses focus on uncovering the overall structure of team collaboration in
crowdsourcing.
The first characteristic of the network under investigation is transitivity, which
captures how teams and groups function (Faust, 2008; Burt, 1992). One theory that has
been proven useful in examining network transitivity is balance theory, which argues that
people preferred a balanced environment with people around them (Heider, 1958).
Balance theory is built upon the assumption that people will try to reduce cognitive
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
49
dissonance, which refers to discomfort feelings when surroundings are out of balance
(Festinger, 1954). In a network, reducing intransitive triads can help balance relationships
among nodes. The benefit of a higher tendency towards transitivity can be explained by
the strength of weak ties (Granovetter, 1973; Valente, 2010). As Granovetter (1973)
argues, higher transitivity means forbidden triads (i.e. if A and B are connected, B and C
are connected, but A and C are disconnected) are uncommon, and thus there are few
weak ties in the network. The scarcity of weak ties enhances information capacity in the
network (Valente, 2010). It helps connect individuals of diverse expertise to exchange
information and leverage functional diversity at the team level. Furthermore, a network
with a tendency towards transitivity tends to be more cohesive and effective in
collectively solving problems (Valente, 2010). When individuals work in teams to
participate in crowdsourcing contests, choosing team member’s existing team members
helps to build trust and reduce potential risk. Therefore,
H9a: There will be a tendency towards transitivity in the crowdsourcing collaboration
network.
The second network characteristic is the tendency for divided sub-communities,
which can be examined with modularity analysis (Girvan & Newman, 2002; Williams et
al., 2015; Kolaczyk & Casárdi, 2014). A higher modularity score indicates that the nodes
in a network can be decomposed into modular sub-networks, defined by their network
positions (Blondel, Guillaume, Lambiotte, & Lefebvre, 2008).
Given the discussion above that both marginality and preferential attachment
could affect how crowdsourcing participants collaborate with each other, it suggests that
both people who are considered as more influential and people who are categorized as
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
50
occupying marginal positions will be more likely to attract more collaboration links. This
indicates that an overall network of team collaboration in crowdsourcing might show a
sophisticated community structure. Within each sub-community, nodes are more densely
connected to each other or more frequently interact with each other than to nodes in the
rest of the network. It is not necessarily the case that there will be strong connections
between the sub-communities. The modularity score for a network increases when more
edges fall within sub-communities, and decreases when more edges fall between
communities (Girvan & Newman, 2002; Newman & Girvan, 2004; Newman, 2006).
Therefore,
H9b: Crowdsourcing collaboration network will have high modularity scores.
In summary, the hypotheses related to marginality test how members of a crowd
are more likely to reach out to marginal individuals in an open innovation environment.
The hypotheses related to preferential attachment captured the other side of the open
innovation story, which means that despite the utopian picture of open innovation, there
still will be some structural features embedded in the community where influential
individuals will be more connected compared to others and become a driving force of the
collaboration network. The last two hypotheses therefore focus on testing how the co-
existence of marginality and preferential attachment can generate a certain structural
feature in the network of collaborative crowdsourcing. See Table 1 for the summary of
hypotheses from Study 1.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
51
Table 1 Summary of Hypotheses and Research Questions in Study 1
Study 1 Team As a form of collaboration in crowdsourcing
Marginality and
team collaboration
H1: Members of a crowd on the technical margin will be
more likely to be selected into teams.
H2: Members of a crowd on the social margin will be more
likely to be selected into teams.
H3: Newer members of a crowd will be more likely to be
selected into teams.
H4: Members of a crowd on the geographic margin will be
more likely to be selected into teams.
Preferential
attachment and team
collaboration
H5: Team leaders in crowdsourcing contests are more
likely to join or seek collaboration ties with others.
H6: Members of a crowd with advantageous network
positions are more likely to join or seek collaboration ties
with others.
H7: Members of a crowd who provided employment
information are more likely to be selected into teams.
H8: Members of a crowd with a record of winning
crowdsourcing contests are more likely to be selected into
teams.
Overall network
structures
H9a: There will be a tendency towards transitivity in the
crowdsourcing collaboration network.
H9b: Crowdsourcing collaboration network will have high
modularity scores.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
52
Study 2 Leveraging Diversity for Team Crowdsourcing Performance
Ever since Surowiecki’s (2004) seminal book on crowdsourcing and wisdom of
the crowd, scholars have been proposing that it is diversity in a crowd that makes this
problem-solving model innovative. This argument has been tested across a wide range of
fields including computing, science, sports forecasting, stock forecasting, and world
events (Buecheler et al., 2010). However, how diversity is defined generates some
debate. Surowiecki (2004) emphasizes the diversity of opinions; while others focus on
the diversity of expertise, which may be derived from differences in knowledge domains,
contexts, product usage, discipline or specialty work areas (Erickson et al., 2012; Schenk
& Guittard, 2011; Sawhney, 2003). Diversity has also been defined in terms of distance
(Fisher, 2004; Pallot, Martínez-Carreras, & Prinz, 2010), and difference in team
members’ nationality (Gibson & Gibbs, 2006).
Furthermore, the debate is mixed with respect to how diversity affects team
performance. Despite the value of functional diversity for innovation, the other side of
the story is that difference is often related to self-categorization and this can lead to social
disintegration and decreased effectiveness (Ancona & Caldwell, 1992; Chatman, Polzer,
Barsade, & Neale, 1998). It is thus overly simplistic to say that higher diversity will lead
to greater creativity.
To uncover the effect of diversity on team outcomes, diversity needs to be
examined in terms of organizational functions (Cohen & Bailey, 1997). Motivated by this
argument, study 2 draws from the literature on generative co-creation in crowdsourcing
and the Faultline theory to examine how team composition affects collaborative
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
53
crowdsourcing outcomes. It aims to test under what conditions demographic and
functional diversity factors can lead to better team performance.
Demographic diversity and functional diversity
To clarify what is meant by diversity, we need to review how it has been defined
in the broader body of literature such as group communication and organizational
communication. Diversity is an umbrella term that indicates the extent to which members
of a team are heterogeneous with respect to individual-level characteristics (Jackson,
1992; Mohammed & Angell, 2004). Diversity has been conceptualized as a unit-level,
compositional construct, which is attribute specific. When describing diversity of a given
attribute within a unit (i.e a team, a working group, or an organization), it refers to the
unit as a whole rather than how a particular member differs from all the other members
(Harrison & Klein, 2007). A unit could be diverse in one specific attribute while
homogenous with respect to another attribute.
We can postulate two general categories of diversity. The first is demographic
diversity, which refers to surface level differences among team members “in overt,
biological characteristics that are typically reflected in physical features” (Harrison,
Price, & Bell, 1998: 97). Examples of demographic diversity are age, gender, race,
ethnicity, nationality and organizational/team tenure (Milliken & Martins, 1996; Peters &
Karren, 2009). The second is functional diversity, which refers to “distribution of team
members across a range of relevant functional categories” (Bunderson & Sutcliffe, 2002,
p. 875). Functional diversity affects both a team’s processes and psychosocial traits
(Dougherty, 1987; Souder, 1987). It captures differences in terms of team members’
attributes, beliefs, values, knowledge or skills that were not readily detectable but can be
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
54
communicated through member interactions (Harrison, Price, & Bell, 1998). In some
studies, functional diversity is used interchangeably with deep level diversity or cognitive
diversity to emphasize peer members’ perception of each other (Van der Vegt & Janssen,
2003; Horwitz & Horwitz, 2007).
The effect of diversity on team performance has been examined extensively in
traditional forms of organizations. However, no consistent patterns have been found. The
general conclusion is that diversity can be a double-edged sword, which can lead to
positive task outcomes but also generate inconcilable divisions among heterogeneous
members (Horwitz & Horwitz, 2007). A certain degree of diversity brings in wider
expertise and knowledge about problems at hand. However, people define and
differentiate themselves in terms of group memberships (see Tajfel 1978 on social
identity theory and Turner 1982 for self-categorization theory). Categorization can be
triggered by higher level of diversity and lead to relational conflict (Mohammed &
Angell, 2007).
Take for example the effect of demographic diversity. Heterogeneity in
demographics can increase creativity and innovation at the team level through expanding
the breadth of knowledge team members possess. The advantage of demographic
diversity lies in the ability of the team to synthesizing members’ diverse opinions and
experience (Amabile, 1983; Jehn, Northcraft & Neale, 1999; Kickul & Gundry, 2001;
Northcraft, Polzer, Neale & Kramer, 1995; Schwenk & Cosier, 1980). In particular,
Kanter (1998) provided an extensive review on how demographic diversity can positively
influence innovation performances at the team level.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
55
However, demographic diversity has been closely tied to social categorization and
social differentiation (Bell, Villado, Lukasik, Belau, & Briggs, 2011; Turner, 1982).
Members of homogenous backgrounds tend to perceive each other as similar. Lacking
common language, surface level differences can lead to increased communication
problems, conflict team membership turnover, and decreased team cohesion (Jehm et al.,
1999; Pelled, 1998; Harrison, Price, & Bell, 1998; Terborg, Castore & DeNino, 1976;
Van Der Vegt, and Janssen, 2003). The detrimental effect of surface level diversity can
also reduce team effectiveness and efficiency (Harrison, Price, Gavin, & Florey, 2002;
Jackson & Ruderman, 1995; Milliken & Martins, 1996; Williams & O’Reilly, 1998).
Studies have shown that functional diversity can either help or hinder team
processes and performance. On the one hand, functional diversity can be positively
related to team performance and the effect is strongest for design and product
development teams where the goal is not necessarily about efficiency (Bell et al., 2011).
Functional diversity in traditional working teams may create higher quality solutions as
team members engage in “critical and investigative interaction processes in which team
members identify, extract, and synthesize their different perspectives” (Amason, 1996:
124). Hong and Page (2004) found that teams of higher cognitive diversity can even
outperform teams composed of all experts. Furthermore, Kilduff, Angelmar, & Mehra
(2000) found deep level diversity and team performance can mutually influence each
other in a positive way. However, on the other hand, conflict and lower efficiency are
common outcomes for teams with cognitive diversity, due to higher heterogeneity in
psychological characteristics or perceived expertise differences (Mohemmed & Angell,
2004).
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
56
Phillips and Loyd (2006) suggest that we need to examine the complexity of the
co-existence of demographic diversity and functional diversity in teams. They looked into
how the interaction between surface level and deep level diversity factors impacts on
emotional and behavior reactions of dissenting team members. They found that
demographic diversity can be perceived as more positive and accepting than surface
homogenous groups, and can foster more persistent and confident voicing of dissenting
perspectives, and display greater task engagement. They further suggest that demographic
diversity may be beneficial for team collaboration even when team members do not have
a different deep level task perspective to share. Demographic characteristics do not
always co-vary with underlying cognitive attributes (Bantel & Jackson, 1989; Hambrick
& Mason, 1984), and that the link between demographic and cognitive diversity may be
more complex than generally assumed (Kilduff et al., 2000; Lawrence, 1997).
A series of moderating variables have also been identified which could influence
the relationship between different dimensions of team diversity and outcomes.
Mohammed and Angell (2004) conclude that the relationship between diversity and
performance is moderated by team orientation and team process. Team orientation helps
to moderate the negative effect of gender diversity on relational conflict while team
process helps to weaken the deleterious effect of time urgency on relational conflict.
They also found that over time the negative effect of demographic diversity on
performance would decrease.
Harrison and Klein (2007) suggest that a closer examination of how diversity is
constructed could help explain why few consistent findings regarding the relationship
between diversity and team performance have emerged. They propose that diversity
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
57
should be analyzed in terms of its substance, pattern, and operationalization, and they
further conceptualize diversity as separation, variety, and disparity. First of all, diversity
can be used as an indicator for separation and thus can be measured as a binary variable.
Second, diversity can be indicative of variety and thus be measured as a categorical
variable to reflect differences in a particular category such as information, knowledge, or
experience among team members. Third, diversity can indicate disparity in terms of
members’ differences in their valued social assets or resources such as pay, social status,
and organizational tenure. This three-dimension typology of diversity contributes to the
literature theoretically and methodologically by differentiating team dynamics, and
capturing multilevel effects of diversity. It also helps to clarify what level of
measurement a study focuses on.
Another reason behind the lack of consensus on the relationship between diversity
and team performance is that most studies rely on either meta data or simulation data
(Bell et al., 2011; Horwitz & Horwitz, 2007; Kilduff, Angelmar, & Mehra, 2000). Very
few studies have examined under what conditions diversity can drive better team
performance with empirical data from survey, and field observation (Phillips & Loyd,
2006; Harrison et al., 2002; Østergaard, Timmermans, & Kristinsson, 2011).
The literature on how diversity and team performance are related highlights the
need to examine how team composition affects team level problem solving. The next
section reviews how to apply the literature on diversity to the context of collaborative
crowdsourcing, through the lens of generative co-creation (Majchrzak & Malhotra, 2013;
Tsoukas, 2009). To further clarify what factors lead to team success, the theory of Group
Faultline is introduced and discussed.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
58
Functional diversity, faultlines, and team success in collaborative crowdsourcing
So far very few studies have been conducted to examine how team diversity
affects crowdsourcing outcomes (see Bayus, 2013 for an exception). One explanation
could be that collaborative crowdsourcing itself is under-examined, which has been
discussed in an earlier section. Another reason is related to the concept of openness
embedded in collaborative crowdsourcing.
In open collaboration context, openness allows for the transparency of
information flow and task coordination. On the one hand, there clearly is a boundary
between different teams that are defined by their tasks, their expertise distribution, and
collaboration principles. On the other hand, there is a larger context of openness that can
make team boundaries porous. Crowdsourcing participants can move autonomously
among teams through voluntary membership. The boundary becomes even more porous
when crowdsourcing is applied to solve social issues, since its outputs and processes are
in the public domain where everybody can access it. This thus raises the question as to
how to measure team diversity in collaborative crowdsourcing and how to examine the
advantages team heterogeneity could bring in improving innovation performances.
Applying the concept of diversity to analyze the effect of team composition on
collaborative crowdsourcing faces another challenge, which is how to take into account
the newly proposed multi-staged approach of crowdsourcing. Scholars have suggested
that to fully leverage the benefits of collaborative crowdsourcing, a multi-staged
approach should be incorporated with the team mechanism. For example, at the first stage
members from a crowd can automatously offer ideas; at the second stage the crowd can
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
59
form teams to collectively refine firm-selected idea by offering comments and feedback
to each other; and at a third stage they can participate in voting best ideas (Leimeister et
al., 2009; Hutter et al., 2011). To uncover how team level diversity influences
performance during staged crowdsourcing contests, we need to evaluate the dynamic
processes of team formation over time and the duration of a team collaboration across
different stages to increase performance.
To analyze under what conditions team composition affects collaborative
crowdsourcing outcomes, this section applies the concept of generative co-creation to
capture the positive effects of functional diversity, and the Group Faultline theory to
capture disruptive influences of diversity.
Functional diversity and generative co-creation in team crowdsourcing
The rationale behind the use of crowdsourcing to solve certain social issues
instead of relying on experts of social development is the value of diverse expertise and
experience members of a crowd possess. Social problems are often complex and require
diverse perspective and heuristics to find solutions. In collaborative crowdsourcing
towards social innovation, how much a person contributes to the collective cause depends
on how her expertise and experience combine with and differ from those of other
problem solvers. As illustrated in Page (2008, pp.152-153), diversity can lead to effective
problem solving through the following procedure: one person searches for a solution until
she get stuck at a local optimum, then the next person searches beginning from that point;
throughout the process each person in a team builds off the best solution found by their
peers and eventually the best solution emerges at the team level when no problem solver
can find an improvement. Though this sequential search is a convenient example, it
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
60
shows the process of generative co-creation toward a collective goal (Majchrzak &
Malhotora, 2013; Tsoukas, 2009).
A number of studies have demonstrated that a crowd of independent strangers
could perform better than a small group of experts in completing crowdsourcing tasks
(Brabham, 2013; Ye at al., 2012; Madsen, Woolley, & Sarangee, 2012). The diversity-
trumps-ability argument highlights the contextual nature of an individual’s contribution.
You do not have to possess the best skillset to contribute; your contribution depends on
your diversity relative to others working on the same problem (Page, 2008). In other
words, diversity matters more than individual ability in collaborative crowdsourcing.
Team members with diverse perspectives and experiences can help each other improve
their local solutions through engaging in a variety of activities such as commenting on
each other’s ideas, offering alternative views, or concisely sharing ideas that are
sufficiently under-developed to inspire more complete ideas (Majchrzak & Malhotora,
2013).
Another related argument is that diversity can trump homogeneity (Page, 2008).
Imagine two teams working on the same crowdsourcing challenge on how to solve water
and sanitation issues in developing communities. The first team has members who all
have distinct perspective and set of heuristics, while all the members in the second team
are identical with the same set of skills and experience. For the homogenous team, they
would all have the same set of local optima. So when one of them finds an optimum,
nobody else from the team can improve on it. Therefore the identical local optimum will
become their collective optimum. However for the heterogeneous team, the possibility of
improvement is higher. When one team member applies her experience and expertise to
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
61
suggest, for example, using microfinance to develop hygienic sanitation facilities, another
member can build upon that idea and suggest how to build partnerships across borders.
Through the generative co-creation process of combining, exchanging, and synthesizing
ideas, the team with diverse members iteratively modifies their solution over time and
finally achieves the global optimum. Therefore, the following hypotheses are proposed to
examine whether diversity in members’ expertise and experience can lead to better task
performance.
H10: In collaborative crowdsourcing, teams with higher functional diversity in
members’ expertise are more likely to win.
H11: In collaborative crowdsourcing, teams with higher functional diversity in
members’ experience are more likely to win.
In collaborative crowdsourcing, another source of cognitive diversity is related to
team members’ network positions (Shalley & Perry-Smith, 2008; Kane, 2009). In
collaborative crowdsourcing, team dynamics require not only the novel combination of
existing knowledge possessed by diverse members, but also the generation of new
knowledge and solution. What also matters is who members of a team know. How a node
is positioned in a network defines the degree to which this node has access to diverse
resources and information (Valente, 2010). In open collaboration networks, individual
level centrality indicates an individual’s ability to exchange knowledge with others
(Wasko & Faraj, 2005). It affects how much information can flow through what paths.
In collaborative crowdsourcing, the discovery of new and more innovative ideas
requires members to take advantage of their networks to expand information sources. The
more diverse a team’s members network positions, the higher probability to be exposed
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
62
to diverse opinions and experience. Members of lower degree scores often occupy less
advantageous positions in a network. However it does not prevent them from contributing
to collective effort, as they may possess unique access to new information or resources.
Teams members of heterogeneous network centrality scores can aggregate information
from diverse groups in the community to engage in generative co-creation (Page, 2008).
Furthermore when there is no sufficient interest among potential contributors to a
collective goal, heterogeneity of team members’ network positions indicates a higher
probability that some members who are central in the network will be more motivated to
contribute and thus facilitate the achievement of a collective goal (Fulk et al., 1996;
Oliver, 1993).
As mentioned in Study 1, two dimensions of centrality are often used in
conjunction with each other to measure individual nodes’ network positions: degree
centrality and betweenness centrality. Heterogeneity in team members’ degree centrality
scores is defined by the variance in the number of connections they have in a network. It
captures the diversity in terms of team members’ activity or community influence.
Members of higher degree centrality scores can function as opinion leaders (Valente,
2010), and having members of lower degree centrality scores helps to bring in new and
unique perspectives to a team. Therefore, team members with diverse degree centrality
scores are more likely to leverage the benefit of information aggregation to combine and
improve existing ideas for problem solving (Page, 2008; Karen, 2009). When a team has
members of diverse betweenness centrality scores, it shows a higher probability that this
team can bridge different social groups and expand the information exchange network.
Both phenomena could lead to better team performance, indicated by a higher probability
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
63
of winning crowdsourcing contests through generative co-creation. Therefore, the
following hypothesis are proposed:
H12: In collaborative crowdsourcing, teams with higher functional diversity in
members’ centrality distribution are more likely to win.
Demographic diversity, team faultlines and crowdsourcing success
Lau and Murnighan (1998) introduced the concept of team faultlines based on the
assumption that greater diversity also makes groups more susceptible to disruptive
influences, potentially offsetting its benefits (Williams & O’Reilly, 1998). Faultlines
refer to the dividing lines based on one or more demographic attributes of team members
(Lau & Murnighan, 1998). The strength of team faultline depends on the extent to which
the hypothetical dividing lines can split a team into subgroups. Essentially, Faultline
Theory is about how to better understand negative effects of diversity through the lens of
multi-dimensional measures of diversity. Social categorization has been at the center of
the Faultline Theory (Knippenberg, Dawson, West, & Homan, 2011).
Team faultline research has advanced the literature on team composition and
helped to uncover how subgroup interactions within teams can influence team
performance. As summarized in Nishii and Goncalo (2008), examining diversity in terms
of faultlines rather than overall heterogeneity within a team helps to clarify the complex
relationship between diversity and creativity. The reasons are twofold. First, the strength
of team faultline influences social processes within a team and thus affects how a team is
able to capitalize on its diversity of perspectives and expertise. Groups with strong
faultlines tend to be more polarized and often suffer from relational conflict, lack of
communication, and behavioral disintegration (Lau & Murnighan, 2005; Li & Hambrick,
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
64
2005). Second, the approach of team faultline analyzes the effects of demographics in
combination of each other rather in isolation of each other. Contextual cues based on
multiple dimensions of member attributes can trigger the strongest and most disruptive
social dynamics not when every group member is unique but when team members’
differences segment teams into subgroups (Lau & Murnighan, 1998).
The faultline effect can be activated by a variety of factors. A salient factor could
be a specific social issue surrounding a team, shared affiliations among members,
geolocation, age, gender or other demographic information (Lau & Murnighan, 1998; Li
& Hambrick, 2005; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). Teams with
moderate diversity (i.e members who are similar within subgroups but different across
subgroups) face more challenges in fostering trust and managing relational conflicts than
teams with maximum diversity (Earley & Mosakowski, 2000). Take location as an
example. Polzer, Crisp, Jarvenpaa, & Kim (2006) found that geographic faultlines
heightened conflict and reduced trust, which led to decreased team performance. These
faultlines were stronger when a team was divided into two equally sized subgroups of co-
located members and when these subgroups were homogeneous in nationality.
Applying the faultline argument to the context of collaborative crowdsourcing,
geographic diversity within teams may seem advantageous on the surface; however,
spatial proximity can provide cues for ingroup categorization and become a source of
team faultline. This could generate negative influences on team performance. When
teams divide up their work, it is natural for members to use colocation as a decision rule
for who should work with whom and how effectively. However, it is not clear under what
conditions a group faultline would be activated and under what conditions geographic
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
65
diversity could bring certain benefits to team performance. In the original argument about
team faultline, Lau and Murnighan (1998, 2001) conclude that higher demographic
diversity such as geographic difference is detrimental to team success. However, some
recent literature has suggested that such faultlines can benefit teams for innovation
(Nishii and Goncalo, 2008). In the context of crowdsourcing for solving social issues,
having team members from different locations could prove helpful as everybody can
bring in their own experience and expertise from their local communities and build upon
each other’s ideas. The following hypothesis is thus proposed to examine the relationship
between geographic diversity and team success.
H13: In collaborative crowdsourcing, geographic diversity will positively
influence a team’s chance of winning.
Another source of team faultlines in collaborative crowdsourcing is diversity in a
member’s experience of being a team leader. Open collaboration often does not have a
clear division of labor and members engage in a collective goal voluntarily through ad-
hoc contribution (Shaw & Hill, 2014). However, there is a clear distinction between team
leaders and regular team members. Team leaders tend to take more responsibilities in
initiating a project and coordinating collective effort. Team leaders are more likely to be
influential users in a community who could help bring in more contributors. In a team
where half of the members have been team leaders themselves and where the other half
have always been regular team members, there might be a strong faultline dividing two
subgroups that favor ideas that were suggested by members of their own group. This
would prevent effective group decision-making. The strength of leadership faultline
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
66
depends on a disparity in subgroup size, which affects team dynamics and thus task
performance (Nishii & Goncalo, 2008).
However in a crowdsourcing environment where team memberships can be
competitive and collaborative at the same time, team leadership can have another
implication. Team leaders could join each other’s teams to learn how to improve their
own projects thus complicating the processes of open collaboration. They might be more
motivated than regular members but at the same time could be more reserved in terms of
how to contribute their ideas. In that case, it is not clear to what extent team leaders are
willing to form a coalition. To clarify the relationship between team leadership diversity
and team performance, the following hypothesis will be examined:
H14: In collaborative crowdsourcing, team leadership faultlines will negatively
influence a team’s chance of winning.
This study also examines the effect of tenure diversity on team performance.
Organizational tenure reflects the breadth of knowledge an individual possess about a
community. There is little consensus on how tenure affects team performance. Some
studies suggest that members’ mean organizational tenure affects team level efficiency
and thus team success, while other studies focus on testing the effect of team tenure. Bell
et al. (2011) concluded that the lack of effect from organizational tenure on team
outcomes could be attributed to the operationalization of the variable. They pointed out
that organizational tenure was never studied as variety conceptualization. This study
proposes the following hypotheses to test the effect of both mean organizational tenure
and also members’ tenure diversity.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
67
H15a: In collaborative crowdsourcing, team members’ mean organizational
tenure will positively influence its chance of winning.
H15b: In collaborative crowdsourcing, team diversity with respect to team
members’ organizational tenure will positively influence its chance of winning.
This study also examines how team size might influence the relationship between
team diversity and outcomes. Literature on traditional forms of collective action has
concluded that teams of larger size face more difficulty in coordinating individuals’
contributions and overcoming free riding challenges (Marwell & Oliver, 1993; Oliver &
Marwell, 1998; Olson, 1965; Peteete & Ostrom, 2004). However recent literature on
newer open approaches of online collaboration suggests that larger size indicates more
inclusion and thus team size does not prevent team success (Smith & Reilly, 2013).
Collaborative crowdsourcing for social good is organized around user autonomy while
members of a crowd work towards a common goal. When users self-organize themselves
into working team, the capability of openness depends on the size of a team. Including
more members in a team helps improve the skill sets the team possesses and can
positively affects how team members effectively coordinate peer contribution and
synthesize into a final product. Therefore,
H16: In collaborative crowdsourcing, team size will positively influence a team’s
chance of winning.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
68
Table 2 List of hypotheses from Study2
Study 2 Leveraging Diversity for Team Crowdsourcing Performance
Functional diversity H10: In collaborative crowdsourcing, teams with higher functional
diversity in members’ expertise are more likely to win.
H11: In collaborative crowdsourcing, teams with higher functional
diversity in members’ experience are more likely to win.
H12: In collaborative crowdsourcing, teams with higher functional
diversity in members’ centrality distribution are more likely to
win.
Demographic diversity
and control variables
H13: In collaborative crowdsourcing, geographic diversity will
positively influence a team’s chance of winning.
H14: In collaborative crowdsourcing, team leadership faultlines
will negatively influence a team’s chance of winning.
H15a: In collaborative crowdsourcing, team members’ mean
organizational tenure will positively influence its chance of
winning.
H15b: In collaborative crowdsourcing, team diversity with respect
to team members’ organizational tenure will positively influence
its chance of winning.
H16: In collaborative crowdsourcing, team size will positively
influence a team’s chance of winning.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
69
Chapter 4 Method and Analysis
The dissertation is based on behavioral data collected from the global
crowdsourcing community Openideo (openideo.com) in November and December 2015.
With Python, the dissertation applies data mining techniques to scrape user level and
team level information from August 2010 to December 2015, capturing all the historical
data in the online community under study. All the data scraping was conducted in
November 2015. To date, studies on crowdsourcing and open social innovation have
relied heavily on qualitative research methods by conducting interviews or qualitative
case studies. More recent analysis applies survey, experiment, simulation, and other
quantitative methods to study the dynamics of such innovation. Scholars have been
calling for analysis of large-scale behavioral data to uncover dynamics of open
collaboration. The current dissertation applies quantitative research methods to analyze
crowdsourcing dynamics at the user, team, and network levels.
The value of using extracted web data can be summarized as twofold. First, it
allows for a more accurate measurement of collaboration structure and user behavior by
tracking digital traces that were left since Openideo was created. Second, it allows us to
explain and predict performance and sustainability of collaborative collaboration with
longitudinal data. By tracking down all the winning teams and winning ideas throughout
the life span of Openideo, the dissertation is able to examine in more detail what factors
can explain the dynamics of collaborative crowdsourcing and what team compositions
lead to better crowdsourcing success.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
70
Background of Openideo.com under study
Openideo.com was chosen as the case study for the current dissertation for the
following reasons. First, it is a global crowdsourcing community that focuses on solving
big social challenges. It provides a platform for people to work together to design
solutions for global challenges, such as public health issues, social inequality, climate
change and other environmental concerns. It fits the focus of this study, which is the
application of crowdsourcing for contributing to social goods. Second, it encourages
collaboration among crowdsourcing contests participants. This allows the analysis of how
people collaborate with each other in the context of open social innovation. Third, it
generates system ranking for its all members based on how active a user is across various
phases of social challenges and it also captures winning history at both individual and
team level. This provides a rich data set for measuring user expertise and performance.
Openideo was officially launched in August 2010. Affiliated with the global
award-wining design firm, IDEO, it originated from a MBA course project at Harvard
Business School proposed by Tom Hulme. The key concept was to “upend IDEO’s high-
touch, high-involvement design consulting business model” (Lakhani, Fayard, Levina, &
Pkorywa, 2013: 6) and leverage its resilient learning environment to utilize the wisdom of
thousands of external individuals who participate collaboratively or competitively to
solve important innovation issues. To some extent, the innovation model in Openideo
was inspired by the success of free/open source software communities that have impacted
business models, and the rise of crowdsourcing platforms such as InnoCentive,
TopCoder, Threadless, and Quirky. As Hulme points out, Openideo is built upon an
innovation network approach that takes on IDEO’s own internal expertise around staged
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
71
processes of innovation and also the benefit of having external creative communities to
contribute (cited in Lakhani et al., 2013). The idea of Openideo was designed when the
founders were based in London, with some other key members located in the Bay Area.
When creating an account on Openideo, individual users can use their Facebook
account or an email address. They can provide background information about their
interests, profession, location, and a short bio. The amount of information people choose
to disclose on their profiles varies greatly, with some providing merely their names while
others providing detailed information about their vision and experience. All the user
information is visible to anybody, including all Openideo members and website visitors.
The major communication channel available on Openideo is through comments
for each post. Users can follow a thread of comments and engage in live conversations.
They can subscribe to email notifications when new comments are added to a thread.
Another communication channel is through offline meetups, which are often organized
by people located in the same city.
In Openideo, members engage in crowdsourcing challenges. A challenge is
usually a three to five month collaborative process that focuses on a specific social
development issue and it creates a space for Openideo community members to contribute,
refine and prototype solutions. The goal of these challenges is to raise awareness about
global social issues, inspire local communities to engage in global conversation about
how to tackle these issues, and to select winning ideas that will be implemented in
different communities to generate social impact, with funding from Openideo partner
organizations.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
72
Each social challenge on openideo typically goes through the following sequence
of development. The first stage is the research phase during which Openideo community
members post anecdotes, photos, videos, or potential solutions that may already exist to
develop empathy and understand people’s needs before diving into solutions. The second
stage is the ideation phase during which members post specific solution concepts based
on what they have proposed at the research phrase. The third stage is the refinement
phase during which Openideo community members test their individual or team
prototypes with real people, incorporate their feedback, and then test again. The last stage
is the evaluation phase during which evaluation of all the refined projects and ideas will
be conducted, and the top ideas will be selected by community members in Openideo.
The diverse participants in Openideo engage in four stages of a challenge. At
each stage, members are encouraged to create teams, invite team members, comment and
build on other people’s research and concepts. Teams can be created at different phases
of a social challenge and may not appear at all phases, depending on how much progress
team members make as a collective. The duration of each stage of development varies.
For example, the research stage often varies from one to two months, the idea stage
varies from 2 weeks to a month, and the refinement and evaluation stages both take two
weeks.
To further elaborate the collaboration process of Openideo, the following is an
example of a completed social challenge (Figure 1). The goal of this challenge was to
design ways of establishing better recycling habits at home. The Recycling challenge
started with 324 contributions, which were short posts on insights, examples and stories
about recycling from different countries. The contributions at the first stage could be
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
73
simply sharing personal experiences, stories on existing solutions, or reflections from
interviewing some experts. Then it moved onto the second stage of development, i.e.
ideation where a total of 203 ideas were conceptualized into creative solutions to address
the challenge question. These ideas were focused on demystifying recycling, how to
make recycling fun, and how to engage children and young people, and so on. All these
ideas went through the applause stage within which Openideo members commented and
provided feedback on their favorite ideas. Some other effort at this stage involved user
testing, and iterations of idea development. Following the applause phrase, the challenge
reached the third stage, i.e. the refinement stage where 25 shortlisted ideas were
identified which went through further development to achieve greater impact. The last
stage was the evaluation to select winning ideas, based on the viability and potential for
social impact. A total of 163 evaluations were submitted by the Openideo community,
which helped to finalize 8 winning ideas. People who participated in the evaluation
included Openideo community members, Coca-Cola Enterprises who sponsored the
challenge, and recycling experts invited by Openideo.
The 8 winning ideas from the Coca-cola recycling challenge are quite diverse.
One proposal is about building a customizable and user-friendly recycling bin which
could help to change people’s perception of trash. Two winning proposals were built
upon the idea of social networking and community building, aiming to design social
networking apps using gamification, incentives, pledges and social status to encourage
people to recycle. Some other proposals suggested building partnerships with Coca Cola
Foundation to match the energy offset from recycling with donations of subsidized or
free electricity units to people in low-income household, starting a social movement to
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
74
persuade people not to consumer products that will end up in landfill on Wednesdays,
and designing an icecream truck also functioning as a mobile recycling center.
Figure 1 One example challenge on Openideo
Throughout the evolution of Openideo, some new features were added and some
stages of development were named differently. However a typical challenge still goes
through four distinctive stages of development, which all emphasize collaboration at the
community level that could take different forms.
Given its social mission and the imperative to draw on an open innovation model,
Openideo provides no financial compensation for community participation. Each social
challenge has at least one sponsor organization, who can be for-profit or non-profit.
There are also inter-governmental organizations (such as UN agencies and World Bank)
and governmental organizations (such as US state department and Center for Disease
Control) that sponsor the challenges. The sponsor organizations work with Openideo to
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
75
decide on what mission to crowdsource to the community and the extent to which the
sponsor can be involved in the development. Though sponsor organizations do not
provide any tangible rewards to crowdsourcing participants, they sometime offer
assistance for winning proposals such as helping with the implementation and evaluation.
People can contribute to Openideo challenges with different roles. There are
creators who contribute new content to the community, critics who are skilled at
constructively providing comments on others’ ideas but do not necessarily want to create
new content, and also lurkers who just want to observe what is going on in Openideo. To
encourage all tiers of users to contribute, some features were created to allow people to
participate without high commitment such as using the applause button to indicate the
support of an idea (Lakhani et al., 2013). In addition, Openideo also has a group of users
who are community managers, whose role is to make sure the community remains
“collaborative, inclusive, community-centered, and optimistic while the posted
challenges steadily moved through the various phases and the appropriate connections
were maintained with the sponsors” (Lakhani et al., 2013: 8). Openideo does not have a
clear disclosure of who the community managers are. However, community managers
can be identified when they posted administrative updates under each challenge, such as
“thanks for all the contributors from this team” or “we have announced our top ideas for
this challenge, please go to the challenge main page to check out our winners”.
There are two different types of participants on Openideo.com. The first type are
solo participants who contribute to social challenges across different phases of a project
individually. Though they may engage in conversation with others by commenting on
each other’s proposals, they do not join any teams on Openideo. They prefer to work
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
76
alone. The second type is team participants who can be team creators, team members or
both. One motivation for people to join teams is that Openideo allows multiple team
affiliations, which help to increase an individual’s chances of winning crowdsourcing
contests. Furthermore, the organizational culture in Openideo encourages collaboration
on solutions among participants. At the end of each challenge top ideas are selected and
will be supported by challenge sponsors for implementation. This generates some degree
of competition. However, in general Openideo emphasizes collaborative crowdsourcing
towards transforming collective ideas into social impact. It states that diversity is a
cornerstone for effective collaboration (Openideo, 2016).
In Openideo, anybody can take the initiative to create a team under a challenge
and can feel free to continue working in the same team across different challenges. There
is no limit as to how many collaborators one team can have, so team sizes range from 2
members to 33 (M = 4.97, S.D. = 3.65). Team leaders identify potential team members by
reviewing what comments peers in Openideo have provided or what ideas they have
submitted. There are also cases in which two individuals create two different teams to
compete for one contest but join each other’s team to collaborate. Other team members
can also suggest whom to recruit in a team. There are no official rules regarding team
roles, division of labor or how a team should function. In addition, the level of
commitment required for team members is also undefined and flexible.
To capture how often a member contributes to the challenges and how they
collaborate with each other, Openideo generates the Design Quotient (DQ) for all the
members. DQ corresponds to how active a user is across various phases of Openideo
challenges. Therefore, there are a total of four categories of DQ: research score, idea
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
77
score, evaluation score, and collaboration score. With a strong emphasis on members’
collaboration, DQ scores increase every time someone comments or builds on other
people’s research and concepts. DQ can be used by members in Openideo to identify who
they want to work with, or publicly identify their own design expertise and strengths. See
Figure 2 for the screenshot of an example DQ distribution.
Figure 2 One Openideo user’s design quotient scores
To summarize, Openideo is built upon a staged approach of crowdsourcing and
encourages collaborative innovation among community members. The goal of its
crowdsourcing contests is to design social challenge solutions, with sponsorship from
different organizations. Each crowdsourcing challenge is organized in the format of a
contest while at the same time Openideo encourages participants to collaborate with each
other. It fits the testbed for the current dissertation, which focuses on uncovering the
processes and outcomes of collaborative crowdsourcing towards social innovation.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
78
Methodology
Data collection
The data collection was conducted in four steps. First, all the 34 public challenges
on Openideo were selected to locate all the active users, who contributed to at least one
challenge. This generated a list of 7,233 Openideo members. Second, public profiles of
all the sampled community members were collected, including the date a user joined
Openideo, team affiliations, number of contributions a user won, number of challenges a
user won, geolocation, and design quotient. Third, team collaboration network data were
constructed by assigning a tie to any two members who joined a team together. Tie
weight was recorded to reflect the collaboration frequency among members. A list of 946
teams and 2100 team members were identified. For the network analysis, the sample of
all the identified team members was used. Fourth, another network was constructed at the
team level, which recorded the sharing of team members.
Data scraping raises concerns about missing data. To handle this issue, the scraper
developed for Openideo data was run a number of times to ensure as few missing data as
possible. However, due to the Python character encoding, users with non-ascii characters
(such as foreign language characters) were excluded during data mining. To fill out
missing data of these users, human coding was conducted by the author to collect user-
level attributes data.
In addition, data cleaning was conducted to keep all user attribute data stored in
consistent format. For example, all geographic information available was recoded to three
variables: country, city and region. For individuals who provided more than one
geolocation, the first geolocation was used to indicate the main location. Employment
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
79
information was also recoded into one binary variable to indicate whether or not a user’s
employment information was available.
Measurement
Technical expertise was measured with four dimensions, based on the design
quotient for all the Openideo members: research score, idea score, collaboration score,
and evaluation score. Each dimension indicates the expertise level a user possesses in the
Openideo community. These variables were used to measure technical competence in
Study 1 and also used to construct diversity measurement in Study 2.
Tenure was measured as the number of days since a user joined Openideo until
the day the data were analyzed (which was January 16
th
2016). This variable was used in
Study 1 to measure tenure. In study 2, tenure was used to measure team level diversity
and also mean tenure among a team’s members. One thing to note is that the
measurement of tenure in the dissertation could be an overstatement of the data, as it was
calculated as how many days since a person registered an account on Openideo. This
measurement, however, cannot capture for sure the activeness of individual users or their
experience. It is also possible that some members may no longer be active in the
community or already dropped out.
Geographic marginality was measured as a categorical variable. 1442 team
members at Openideo provided detailed information on where they were located, which
was about 20% of the whole Openideo population. User geolocation was identified from
their Openideo profile page. The majority of the team players were based in US (N = 699)
and UK (N = 196). Given that key administrative members of Openideo were also
located in US and UK, user geolocation was coded as a binary variable: located in main
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
80
Openideo innovation sites, or not. All the users located either in US or UK were coded as
1 and people located elsewhere were coded as 0, which indicates site marginality (N =
547). Missing data were coded as missing values.
Social marginality was measured by betweenness centrality in the team
collaboration network, which was to indicate the popularity or activeness of a user, which
ranged from 1 to 311 (M = 12.63, S.D. = 12.16). Following Freeman (1979), it was
measured by recording how frequently a node lies on the shortest path connecting
everyone else in the network. This variable was used in both Study 1 and Study 2.
Team leader was coded as a binary variable (1 = team leader, 0 = not team
leader). If one user created at least one team in Openideo, this user was coded a team
leader (N = 629).
Community influence was measured by the degree centrality in the team
collaboration network. This was to indicate the power a user has in the community. Since
the team collaboration network in Openideo is defined as symmetric, degree centrality
was measured by the number of unique connections a node has with others. The tie
frequency was not weighted in the measurement, as the unique connection suggests the
influence a node has over others in terms of information spread and resource coordination.
This variable was used in both Study 1 and Study 2.
Employment information availability was coded as a binary variable, with 1 =
available, and 0 = not available. A total of 811 sampled team users provided their detailed
employment information to endorse their expertise and experience. This variable was
used in Study 1 only.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
81
Number of winning proposals was measured at individual user level in terms of
how many proposals an Openideo user won as of when the data were collected. When a
team won, all its members get the credit. It captures the number of winning proposals
across all the different phrases of all the crowdsourcing challenges in Openideo. In some
cases, the winning proposals were submitted by individuals; while in other cases the
winning proposals were submitted by a team thus making all the team members winners.
This variable ranged from 0 to 26 (S. D. = 1.03). It was used in Study 2 to measure
winning experience diversity at the team level by calculating the standard deviation.
Winner was measured as a binary variable: 1 = winner (who won at least once)
and 0 = other participants who never won in Openideo crowdsourcing.
Number of winning challenges was measured at the individual user level as well
to capture how many unique Openideo challenges a user won. All team members get the
credit when the team’s proposal won a challenge. This variable ranged from 0 to 9 (S. D.
= .61). It was also used in Study 2 to capture the winning experience diversity at the team
level by calculating the standard deviation of all team members’ scores.
Expertise diversity was measured by the standard deviation of team members’
skill scores in the following technical expertise areas: research, idea, collaboration, and
evaluation. This variable was measured at the team level and was used in Study 2 only.
Crowdsourcing success was measured at the team level. The variable was coded
as binary: 1 = a team won at least one Openideo challenge; 0 = a team did not win one
challenge.
Team size was measured as the number of members in each team.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
82
Team betweenness centrality was measured by recording how frequently a team
node lies on the shortest path connecting all the other teams in the team-team affiliation
network. A higher score indicates that the team occupies a more strategic position.
Team degree centrality was measured by recording the number of ties a team had
in the team-team affiliation network (Freeman, 1979). The higher the score, the more
influence the team may possess.
Team degree centrality diversity was measured by calculating the standard
deviation of all team members’ degree centrality scores. The higher the standard
deviation, the higher diversity a team possesses.
Team betweenness centrality diversity was measured by calculating the standard
deviation of all team members’ betweenness centrality scores. The higher the standard
deviation, the higher diversity within a team.
Team geographic diversity was measured to capture the degree of heterogeneity
in members’ geolocation distribution. Geographic information of all the users was coded
as a categorical variable to indicate what regions they were located. Therefore, the
geographic diversity was calculated following Teachman (1980) and Pelled, Eisenhardt,
and Xin (1999). The index was as follows, which captures how team members are
distributed among the possible categories of a variable:
In the index, the total number of available categories of a variable equals l, and Pi is the
fraction of team members falling into category i.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
83
Leadership diversity was measured by calculating the degree of heterogeneity in
terms of team members’ leadership experience in Openideo. Given that leadership was
coded as binary variable, the same index was applied to calculate leadership diversity
score at the team level.
Analysis
Data analysis was conducted in three steps. First, descriptive statistics of the team
collaboration network data and user attributes data were computed to provide an overall
picture of collaborative crowdsourcing in Openideo. Second, exponential random graph
models (ERGM) were used (Koehly & Pattison, 2005) to test H1-H9a in Study1. H9b
was tested with community detection analysis, which followed the algorithm Louvain
method (Blondel et al., 2008). Third, logistic regression was conducted at the team level
to test all the hypotheses in Study 2. ERGM and logistic regression were conducted in R.
Community detection and data visualization were done in Gephi.
ERGM methods estimate the probability of tie formation among dyadic nodes
compared to what would occur randomly by chance alone. Modeling estimation is based
on Markov Chain Monte Carlo (MCMC) procedures for optimal estimates of parameters
(Robins et al., 2007). An ERGM fits the data when the t values of all parameters are
lower than 0.01, indicating that the standard error of each estimated parameter is within a
tolerable range of its actual value, given the sample data from the populations under
study and randomly generated networks of the same size (Snijder et al., 2006; Weber &
Monge, 2011). A specific parameter is significant when the t value is within 1.96
standard errors of the estimated parameter in the model (p < .05, Robin et al., 2007, p.32).
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
84
The procedure of ERGM tests was conducted as follows. First of all, for all the
estimated models two basic network attributes were added in the modeling: edges which
tests the tendency towards connection at the overall network level; and transitivity which
was tested with the gwesp parameter, i.e. geometrically weighted edgewise shared partner
distribution (Handcock, Hunter, Butts, Goodreau, Krivitsky, & Morris, 2012). Second, to
test all the hypotheses, corresponding parameters were added. To test H1-H4, the
following parameters were added: four dimensions of expertise scores, betweenness
centrality, community tenure, and site marginality. To test H5-H8, team leader, degree
centrality, employment, and winning information were added as parameters. H9a was
tested with the effect of transitivity from ERGM.
Logistic regression analyses were conducted to examine relationships between
diversity and team success (α = 0.05). The analysis was conducted in the following order.
First, Model 1 included the following cognitive diversity measures: research skills
diversity, idea skills diversity, collaboration skills diversity, evaluation skills diversity,
winning ideas diversity, winning challenges diversity, team members’ betweenness
centrality diversity, and team members’ degree centrality diversity. Model 1 was tested to
examine how these variables influenced a team’s chances to win crowdsourcing contests.
Second, Model 2 was built to include all the demographic diversity measures (i.e.
geographic diversity, leadership diversity, and team members’ community tenure
diversity). The following 4 control variables were included in Model 2 as well: team size,
team mean tenure, team’s degree centrality, and team’s betweenness centrality. Third, all
the significant variables from Model 1 and Model2 were included in the final model
(Model 3).
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
85
To summarize the hypothesized relationship between dependent and independent
variables, the equation is stated as:
Log(team winningit) = β0 + β1X1(functional diversity it) + β2X2(demographic
diversityit + β3X3(control variablesit)
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
86
Chapter 5 Results
Characteristics of Openideo collaborative crowdsourcing participants
Population features
The data scraping of the Openideo population generated a total of 7922 users.
Their tenure on Openideo ranged from 36 days to 2039 days, with the average tenure of
1005.14 days (S.D. = 607.71). Their research scores ranged from 0 to 6114 (M = 40.06,
S.D. = 151.92). Their idea scores ranged from 0 to 9057 (M = 46.08, S.D. = 173.61).
Their evaluation scores ranged from 0 to 2806 (M = 3.99, S.D. = 42.71), and
collaboration scores ranged from 0 to 33357 (M = 56.00, S.D. = 538.05). See more
detailed statistics in Table 3.
Table 3 Summary of descriptive statistics of Openideo population and team participants
Population level All team participants Team leaders
Number of members 7922 2100 625
Average Openideo
Tenure
1005.14 (S.D.
=607.71)
739.67 (S.D. = 459.19) 736.18 (S.D. =
440.36)
Average Research
Score
40.06 (S.D. =
151.92)
51.78 (S.D. = 225.13) 103.11 (S.D. =
378.75)
Average Idea Score 46.08 (S.D. =
173.61)
110.25 (S.D. = 313.52) 231.77 (S.D. =
525.17)
Average Evaluation
Score
3.99 (S.D. = 42.71) 7.97 (S.D. = 51.38) 17.58 (S.D. =
86.57)
Average Collaboration
Score
56.00 (S.D. =
538.05)
124.44 (S.D. = 712.99) 301.76 (S.D. =
1206.79)
Percentage of winners 7% (N = 580) 21% (N = 440) 26% (N = 161)
In terms of region distribution, 3297 users provided identifiable information, with
2429 from North America, 1002 from Europe, 506 from Asia, 207 from Africa, 185 from
South America, 172 from Australia and Oceania, 72 from Middle East, 47 from Central
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
87
American, and 5 from the Caribbean. All of these 3297 users were located in 127
countries. See more detailed statistics of region distribution in Figure 3. Majority of them
were located in United States (N = 2278), United Kingdom (N = 471), India (N=213),
Australia (N=154), Canada (N=148), Germany (N=93), Brazil (N=87), Netherlands (N=
62), Singapore (N=58), Colombia (N=55), Kenya (N=55), China (N=54), and the
Philippines (N=47). All the Openideo members were located in a total of 121 countries
and 296 cities. See the world map of their country origins (Figure 4), which shows the
global participation.
Figure 3 Region distribution of all the Openideo members
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
88
Figure 4 Country origins of Openideo members (121 countries)
Team participants characteristics
As of when the data were collected, a total of 2100 team participants were
located. This accounted for about 26% of the total population in Openideo. Among all the
team players, research scores ranged from 0 to 4541 (M = 51.78, S.D. = 225.13), idea
scores ranged from 0 to 9057 (M = 110.25, S.D. = 313.52), collaboration scores ranged
from o to 15114 (M = 124.44, S.D. = 712.99), and evaluation scores ranged from o to
1573 (M = 7.97. S.D. = 51.38). All these four dimensions on technical skills were
significantly correlated with each other with the sample of all the team players. The
strongest correlation was between research skills and collaboration skills (r = .81, p <
.001), followed by the correlation between evaluation skills and collaboration skills (r =
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
89
.79, p < .001), the correlation between research skills and evaluations (r = .75, p < .001),
the correlation between idea skills and collaboration (r = .67, p < .001). This highlights
the importance of collaboration in the Openideo community. Among all the other
correlation scores, the stronger one was between research skills and idea skills (r = .63, p
< .001), followed by the correlation between idea skills and evaluation skills (r = .58, p <
.001).
The average Openideo tenure of all team users was 739.67 (S.D. = 459.19),
indicating that on average an Openideo member has been in the community over two
years. The average tenure of team users was shorter than the population mean tenure.
Tenure was significantly related to all the four dimensions of technical skills, with the
strongest correlation with research skills (r = .28, p < .001), followed by evaluation skills
(r = .23, p < .001), collaboration skills (r = .18, p < .001), and idea skills (r = .13, p <
.001).
The number of proposal a team participant won ranged from 0 to 26 (M = .32,
S.D. = 1.03). The number of Openideo challenges a team participant won ranged from 0
to 9 (M = .25, S.D. = .61). On average, a user joined 24.45 teams, indicating active
engagement among sampled participants. Among all the team players, 440 won at least
one project. On average, a team participant won .32 idea and .25 project.
In terms of geolocation distribution, 706 people indicated they were located in
North America, 339 people in Europe, 139 in Asia, 98 in Africa, 47 in Australia and
Oceania, 34 in South America, 26 in Middle East, and 8 in Central America. As
mentioned earlier, majority of the team participants were located in the global north.
Main countries included US (N = 699), UK (N = 196), Australia, and German. India and
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
90
Uganda were the two developing communities with a significant number of participants.
Most of the participants reported that they work in research institutes (such as Middelsex
University London, NYU, UCSF), inter-governmental orgs (UNICEF, UNHCR), new
agencies (BBC), NGOs, and design firms. Very few people indicated that they were
affiliated with Openideo (N = 7).
Among all the team participants, 625 created at least one team in Openideo
(coded as team leaders). On average, team leaders had been in Openideo for 736.18 (S.D.
= 440.36), which did not differ much from the average tenure of all team users. Team
leaders’ average research score was 103.11 (S.D. = 378.75), average idea score was
231.77 (S.D. = 525.16), average collaboration score was 301.76 (S.D. = 1206.78), and
average evaluation score was 17.58 (S.D. = 86.57), which were all higher than the
corresponding scores for all the team users. 200 of these team leaders indicated that they
were located in the US. As shown in Table 3, on average skill scores for team leaders
were the highest, followed by the average of team participants, and by the average scores
of the Openideo population.
The team collaboration network in Openideo has a network density of .012, with
an average path length of 3.61 between any two users. The degree centralization was
14%, indicating a group of key members in the network were occupying central
positions. The average degree centrality in the network was 12.24, with the average
weighted degree centrality being 24.48. The clustering coefficient (i.e. the average of
each node’s network density) was .80, indicating that nodes in the crowdsourcing
collaboration network were connected in dense pockets of interconnectivity.
A modularity analysis was conduced with the Louvain method (Blondel et al.,
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
91
2008), which searches for a partition of all nodes into distinct sets (communities) to
maximize network modularity. The Louvain method defines community as a group of
nodes that interact more frequently with each other. Its partitioning algorithm that
maximizes modularity provides a robust estimation of community structure (see Girvan
& Newman, 202; Newman, 2006). The community detection results show that all the
2100 nodes can be decomposed into 63 communities, with a resolution of .73.
The visualization of the main network component (i.e. only nodes connected to
the largest sub-network) showed a clear structure with divided sub-communities (Figure
5). The color of each node represents a sub-community it belongs to, and node size is
proportional to its degree centrality. Nodes labeled with user names indicated members
who indicated their affiliation with Openideo, either as employees, local meetup
organizers or former interns. They were located in US (Chicago, San Francisco, New
York), UK (London), Spain (Barcelona), Austria (Vienna), and Kenya. The visualization
shows that some influential users (indicated by the size of the nodes) were connecting
different sub-groups within the collaboration network.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
92
Figure 5 Visualization of Openideo team collaboration networks
Team characteristics
Among all the 943 teams identified, 74 won at least one challenge in Openideo.
Team size ranged from 2 to 33 (M = 4.97, S.D. = 3.65). 623 teams appeared at the first
stage of challenge development (research stage), 214 appeared at the second stage
(ideation stage), 70 appeared at the third stage (refinement stage), and 33 appeared at the
winning stage. Among all the teams, 79 participated in more than one challenge together.
All of these teams worked on 34 Openideo challenges (Table 4), in the fields of
higher education, renewal energy, recycling, public health, reducing generational gap,
human rights, water sanitation, and food security.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
93
Table 4 A list of 34 Openideo Social Challenge
Project
ID
Project Name
1 The Higher Ed Challenge
2 How might urban slum communities become more resilient to the effects of climate change?
3 How might we improve education and expand learning opportunities for refugees around the
world?
4 How might we use technology to inspire all socioeconomic and multicultural groups to lead
healthier lives?
5 How might we use the power of communities to financially empower those who need it most?
6 How might communities lead the rapid transition to renewable energy?
7 How might we rapidly equip and empower the care community to fight Ebola?
8 How might parents in low-income communities ensure children thrive in their first five years?
9 How might we build better employment opportunities and pathways for young people around the
world?
10 How might we inspire and engage young people to support older adults through mentorship?
11 How might we establish better recycling habits at home?
12 How might we make low-income urban areas safer and more empowering for women and girls?
13 How might we inspire young people to cultivate their creative confidence?
14 How might we all maintain wellbeing and thrive as we age?
15 How might we gather information from hard-to-access areas to prevent mass violence against
civilians?
16 How might we create healthy communities within and beyond the workplace?
17 How might we inspire and enable communities to take more initiative in making their local
environments better?
18 How might we identify and celebrate businesses that innovate for world benefit - and inspire
other companies to do the same?
19 How can we manage e-waste & discarded electronics to safeguard human health & protect our
environment?
20 How can we equip young people with the skills, information and opportunities to succeed in the
world of work?
21 How might we support web entrepreneurs in launching and growing sustainable global
businesses?
22 How might we design an accessible election experience for everyone?
23 How might we restore vibrancy in cities and regions facing economic decline?
24 How can technology help people working to uphold human rights in the face of unlawful
detention?
25 How might we increase social impact with OpenIDEO over the next year?
26 How might we use social business to improve health in low-income communities?
27 How might we better connect food production and consumption?
28 How might we increase the number of registered bone marrow donors to help save more lives?
29 How might we improve maternal health with mobile technologies for low-income countries?
30 How can we improve sanitation and better manage human waste in low-income urban
communities?
31 What global challenge do you think innovation leaders should work to solve right now?
32 How might we increase the availability of affordable learning tools & services for students in the
developing world?
33 How can we raise kids' awareness of the benefits of fresh food so they can make better choices?
34 Create an inspirational logo for OpenIDEO
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
94
Study 1: Network mechanisms underlying team collaboration in crowdsourcing
The model for testing all the parameters was a good fit. H1 stated that members of
lower technical skills would be more likely to be chosen as team members. This
hypothesis was supported for the following measures: evaluation expertise (coefficient =
-3.45-04, SE = 7.79e-05, p = .0001), and collaboration expertise (coefficient = -8.51e-05,
SD = 1.91e-05, p = .0001). This indicates that participants who have lower scores in
either evaluation or collaboration skills are more likely to be chosen as team members.
However, the effect of the research score was opposite to the hypothesis, indicating
people who are more skilled in research are more likely to be selected into teams
(coefficient = 2.01e-04, SD = 5.21e-05, p = .0001). The same effect was found on the
idea score but it was not significant (coefficient = 1.70e-05, SD = 2.63e-05, p = .51797).
Therefore, H1 received some support.
H2 stated that diversity in social ranking in the community would lead to higher
likelihood of collaboration among any pair of members. It was supported with
measurement of betweenness centrality, indicating that people who possess lower
betweenness centrality scores are more likely to be chosen as team members (coefficient
= -9.03e-06, S.D. = 6.67e-07, p = .0001). H2 was supported.
H3 stated that newer members in the Openideo community would be more likely
to be chosen as team members. This hypothesis was not supported (coefficient = 2.35e-
06, S.D. = 1.23e-05, p = .85). The correlation was in the opposite direction of the
marginality hypothesis, indicating that the more senior a person is in the community, the
more likely this person would be selected as a team member. However the effect of
community tenure was not significant in predicting tie formation patterns.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
95
H4 stated that individuals located at a marginal site are more likely to be chosen
as team members. This hypothesis was not supported. The result shows that individuals
located in US or UK are more likely to be selected into teams (coefficient = .05, S.D. =
.01, p = .0001), which is in the opposite direction of the hypothesized geographic
marginality effect.
H5, H6, H7 and H8 tested the effect of strategic selection on team formation. H5
stated that team leaders are more likely to join other teams, either to create new teams or
be recruited by other teams. The results showed that team leaders are more likely to join
team collaboration than non-team leaders (coefficient = .04, S.D. = .01, p = .0004). H5
was supported. H6 stated that people that are more active or influential in Openideo,
measured by degree centrality in the network, are more likely to be chosen as team
members in the future. It was supported (coefficient = .02, S.D. = 5.11e-04, p = .0001).
H6 was supported. H7 stated that users who provide employment information are more
likely to be chosen as team members. The results showed the opposite tendency that
people who provided their employment information are less likely to be chosen as team
members compared to people who did not (coefficient = -.03, S.D. = .02, p = .0592). The
effect however was not significant. Nevertheless, H7 was not supported. H8 stated that
individuals with wining experience would be more likely to be selected into teams. It was
supported (coefficient = .03, S.D. = .01, p = .0309).
H9a and H9b are hypotheses related to the overall network characteristics of
collaborative crowdsourcing. Based on the descriptive statistics, overall the team
collaboration network in Openideo has less tendency towards density (coefficient = -9.47,
SE = .27, p = .0001), which is common in online collaboration network. The reason is
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
96
coordination takes effort and collaboration does not occur indiscriminately. The network
also showed a stronger tendency towards transitivity than a random network (coefficient
= 4.58, SE = .21, p = .0001). This indicates that team collaboration is built upon trust.
H9a was supported. In addition, the modularity co-efficient of the network (.73) was
quite high. It indicates a strong tendency of dividing the network into smaller clusters or
sub-communities. It shows that in the crowdsourcing collaboration networks, there were
some dense connections between nodes within sub-communities but sparse connections
between nodes in different sub-communities. H9b were both supported. See Table 5 for
the summary of ERGM estimation.
Table 5 Summary of ERGM - Study 1 results
Parameters Estimate S.E. p value
Edges -9.47 .27 < .0001
Generalized reciprocity 4.57*** .21 .000001
H1a Research skill – Not
supported
2.01e-04*** 5.21e-05 .52
H1b Idea skill – Not supported 1.70e-05 2.63e-05 < .0001
H1c Evaluation skill -
Supported
-3.45e-04*** 7.79e-05 < .0001
H1d Collaboration skill -
Supported
-8.51e-05*** 1.91e-05 < .0001
H2 Betweenness centrality –
Supported
-9.03e-06*** 6.67e-07 < .0001
H3 Tenure – Not supported 2.35e-06 1.23e-05 .85
H4 Geolocation – Supported .05*** .01 .0003
H5 Team leader – Supported .04*** .01 .0004
H6 Degree centrality -
Supported
.02*** 5.11e-04 < .0001
H7 Employment - Not
supported
-.03 .02 .06
H8 Winner - Supported .03* .01 .03
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
97
Study 2: Functional Diversity, Team Faultline and Crowdsourcing Success
This section summarizes results with respect to how team diversity affects task
success in collaborative crowdfunding. Logistic regression model 1 tested the effects of
the functional diversity, including the following variables: team diversity in members’
research skills, team diversity in members’ conceptualization skills, team diversity in
members’ collaboration skills, team diversity in members’ evaluation skills, team
diversity in terms of how many crowdsourcing projects members won, team diversity in
terms of how many wining ideas team members won, and team diversity in members’
degree centrality and betweenness centrality. The model was significant (Χ
2
= 27.6, df =
3, p < .0001). Among all the four measurements of expertise diversity, only evaluation
skill diversity was a significant and positive predictor for team success (B = .02, S.E
= .005, p = .0002). Team diversity in members’ idea conceptualization skills was also
positively related to team success, however the relationship did not achieve statistical
significance (B = 1.92e-04, S.E = 4.34e-04, p = .66). Team diversity in members’
research skills (B = -1.12e-03, S.E = 6.06e-04, p = .07) and collaboration skills (B = -
6.78e-03, S.E = 3.77e-04, p = .07) were both negatively related to team success.
Therefore H10 received some support.
The positive effect of winning experience diversity on team success was
significant when it was measured as the heterogeneity in team members’ winning ideas
(B = .38, S.E. = .10, p = .0001). However, when experience diversity was measured as
the difference in how many unique Openideo challenges each team member won, it
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
98
showed a significant and negative effect on team performance (B = -.87, S.E. = .35, p
= .003). Therefore, H11 received some support.
Diversity in members’ network positions was measured with two variables:
degree centrality diversity and betweenness centrality diversity. Degree centrality
diversity was not significant (B = .02, S.E = .02, p = .37). The effect of betweenness
centrality diversity was not significant either (B = 1.36e-05, S.E = 1.78e-05, p = .45).
Therefore, H12 was not supported.
Model 2 tested the effects of demographic diversity and several control variables.
In particular, the following variables were included: team geographical diversity, team
diversity in members’ organizational tenure, mean organizational tenure, and team size.
Team level network measures were added as control variable, including team degree
centrality and team betweenness centrality. Model 2 was significant (Χ
2
= 18.00, df=3, p
= .00045). Team geographic diversity was the strongest predictor in the model (B = 1.02,
S.E = .39, p = .009). H13 was thus supported. Leadership role diversity was not
significant (B = -.51, S.E = .85, p = .55). H14 thus was not supported.
Members’ tenure diversity was negatively related to team success and the effect
was significant (B = -2.28e-03, S.E = 8.34e-04, p = .006). Team mean tenure was
positively related to team success the effect was significant (B = .1.26e-03, S.E = 3.78e-
04, p = .0009). Therefore both H15a and H15b were supported. Team size was found to
be a positive and significant predictor for team success (B = .09, S.E = .04, p = .01),
supporting H16. Team betweenness centrality was not significantly related to team
success (B = -9.86e-05, S.E = 9.12e-05, p = .28). Team degree centrality was positively
related to team success and the effect was significant (B = .02, S.E = 5.67e-03, p = .002).
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
99
The final model predicting team success included all the significant variables
from model 1 and model 2. The model 3 was significant (Χ
2
= 32.80, df = 3, p < .0001).
Evaluation skills diversity remained to be a significant and positive predictor of team
success (B = .006, S.E. = .002, p = .002), so did winning ideas diversity (B = .43, S.E. =
.11, p = < .0001). Team geographic diversity also remained a significant and positive
predictor (B= .96, S.E. = .39, p = .02). Two measurements of diversity remained to be
negative predictors of team success: winning challenge diversity (i.e. heterogeneity in
teams of how many challenges a member won, B = -1.24, S.E. = .37, p = .0008), and
members’ tenure diversity (B = -.002, S.E. = .0008, p = .04). Both team size (B = .12,
S.E. = .04, p = .001) and members’ average tenure (B = .001, S.E. = .0005, p = .03)
remained positive predictors of team success. In the final model, the effect of team degree
centrality was not significant (B = .008, S.E. = .006, p = .18). All the results from ERGM
analysis were summarized in Table 6.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
100
Table 6 Summary of logistics regression models – Study 2 results
Variables Model 1 Model 2 Model 3
Estimate S.E. Estimate S.E. Estimate S.E.
H10a Research skill diversity –
not supported
-.001 .0006
H10b Idea skill diversity – not
supported
.0002 .0004
H10c Collaboration skill
diversity – not supported
-.0007
.0004
H10d Evaluation score
diversity - supported
.02*** .01 .01** .002
H11a Winning ideas diversity -
supported
.38*** .10 .43*** .11
H11b Winning challenges
diversity – not supported
-.87* .35 -1.24*** .37
H12a Betweenness centrality
diversity – not supported
.00001 .00002
H12b Degree centrality
diversity – not supported
-.02 .02
H13 Geographic diversity -
supported
1.02** .39 .96* .39
H14 Leadership diversity – not
supported
-.51 .85
H15 a Mean tenure – supported .001*** .0004 .001* .0005
H15 b Tenure diversity – not
supported
-.002**
.001
-.002*
.001
H16 Size - supported .09* .04 .12** .04
Team degree centrality .02** .006 .008 .01
Team betweenness centrality -.0001
.0001
Intercept -2.54*** 2.30 -4.53 .82 -4.66 .50
Chi square 27.60, p < .0001 18.00, p = .0005 32.80, p < .0001
AIC 488.31 461.67 445.36
Note: *** p < .001, ** p < .01, * p < .05
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
101
Chapter 6 Discussion and Conclusion
Given the rise of open collaboration and open social innovation, this dissertation
set out to examine the implementation of crowdsourcing in the context of collectively
solving social problems. The goal of this dissertation was twofold. First, it aimed to
uncover what factors drive team formation patterns in collaborative crowdsourcing.
Second, it focused on examining the relationship between team diversity and
crowdsourcing success. The aspect of collaborative crowdsourcing is under-examined in
the literature; however scholars have called for the investigation of collaboration
dynamics to uncover the potential of aggregating wisdoms of a crowd for better solutions.
Through empirically testing data collected from a global crowdsourcing community,
Openideo, this dissertation provides insights regarding the mechanisms underlying an
individual’s decision to form team collaboration ties with a shared collective goal and
how to build a winning team to solve social challenges.
The dissertation offers several important contributions to the literature on
crowdsourcing, open collaboration and open development. First, it extends existing
research on crowdsourcing by focusing on how to apply the crowdsourcing model to
open social innovation issues. It moves beyond the current focus on crowdsourcing for
business outcomes and conceptualizes crowdsourcing as a model for solving social
problems and achieving positive social outcomes. This type of crowdsourcing differs
from firm-centric innovation in the sense that the ultimate goal is to gain social benefits
through the aggregation of individuals’ contributions. The focus on open social
innovation resonates well with recent studies on open development models which are
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
102
built upon an increasingly open ecosystem of technology and tools (Smith & Reilly,
2013). It highlights the power of openness in solving social development problems,
without having to resort to contracted experts or sophisticated development agencies.
Second, this dissertation is the first attempt to systemically examine collaborative
crowdsourcing from a network perspective. The value of crowdsourcing is in aggregating
individuals’ wisdoms and allowing members of a crowd to build upon each other’s ideas
to keep improving solutions. With a focus on collaborative crowdsourcing, this study
applied a network perspective to uncover different mechanisms underlying team
formation among members of a crowd. It demonstrated the value of using network
theories and methodologies to examine the processes of crowdsourcing. Taking into
account the relational aspects of team formation, this dissertation moved beyond the mere
examination of the attributes or behaviors of individual members in a crowdsourcing
community and analyzed the dynamics among members of a crowd in terms of selecting
team members.
Third, this dissertation contributes to the literature by empirically testing what
team compositions lead to better team performance, through the lens of team diversity. A
number of team studies rely on metadata or simulation data to examine the relationship
between diversity and team task outcome. Furthermore literature on team diversity has
not been extended to the context of collaborative crowdsourcing or open collaboration.
With large-scale behavior data at the team and user levels, this dissertation focused on
testing two categories of diversity: demographics and cognitive. With logistic regression,
results provided important implications on what diversity factors could lead to disruptive
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
103
influences when building teams for crowdsourcing and what could contribute to the
generative co-creation processes and thus lead to better team performances.
In the following sections, detailed discussion on results from the two studies
conducted in the current dissertation is presented. It further demonstrates the
contributions this current research brings to the field of crowdsourcing and open
collaboration. Implications on crowdsourcing community sustainability and how to build
winning teams are provided.
Building a sustainable open collaboration environment
Study 1 was set to examine how marginality and preferential attachment drive the
structures of collaborative crowdsourcing. With a focus on the application of
crowdsourcing in open social innovation, where members of a crowd collectively design
solutions for social challenges, this study was based on team collaboration data from
Openideo. Network analysis was conducted to model what factors explain collaboration
patterns among team participants. ERGM results show that certain dimensions of
marginality drove crowdsourcing participants to work together towards a common goal,
and so did the mechanism of preferential attachment.
Division of labor based on technical skills
The effect of technical marginality in driving team collaboration was mainly
attributed to the following variables: evaluation skills and collaboration skills. It indicates
that less skilled members are more likely to be selected as team members. The possible
reason is that members lacking skills in a certain area may possess expertise in other
fields that can provide alternative innovation ideas for the team. For example, some
individuals may have lower evaluation skills or collaboration skills but they possess
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
104
unique experience about how a social challenge has been handled in their local
community and how the crowdsourcing team they are part of can benefit from that. Given
that the social challenges posed on Openideo are often community development issues
that are common to any society, these members are able to share their stories and
experiences with others who might be experiencing similar issues. The open sharing
environment allows Openideo members to build upon each other’s expertise and
experience to design a better solution.
These results also showed that different dimensions of technical expertise have
heterogeneous influence on team formation. For example, individuals of higher research
skills are more likely to be selected into teams. This indicates that individuals with good
skills in researching relevant social problems are considered as good assets to teams.
Though individual’s idea skills score had no significant effect on collaboration tie
formation.
The heterogeneity effect of technical expertise could be a sign of division of labor
in Openideo and shows the value of aggregating people of diverse expertise. For
example, in a team some members help analyze the social issues under study with their
unique knowledge, some invest more on aggregating these stories and developing solid
design ideas, and others focus on utilizing their evaluation expertise to contribute to the
challenge.
The effect of social marginality was significant, indicating that individuals less
central in the team collaboration were more likely to be selected into teams. Lower
betweenness centrality indicates a node occupies a network position of less advantage in
terms of bridging different groups in a community. However, the results showed that this
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
105
does not mean individuals of lower betweenness centrality cannot contribute to
collaborative innovation. They may possess unique knowledge and fresh views about
how to solve innovation issues. Given the open environment that has been afforded by
Openideo, members do not have any barriers regarding how to collaborate across team
boundaries. This could explain why betweenness centrality was a negatively predictor for
team formation.
The effect of community tenure on team collaboration was not significant.
However, it was positively related to tie formation probability. This is consistent with the
findings from other open collaboration communities such as free open source software
communities, Wiki communities, and other peer production communities. The longer a
user stayed in a community, the better knowledge he might possess, which could prove to
be an important asset for a team in terms of how to efficiently coordinate tasks on the
website and offline, and how to better navigate the different phases of a crowdsourcing
challenge. The insignificant effect of community tenure on team formation indicates that
in choosing team collaborators, there is no preference over newbies or senior members.
As long as the members could contribute to the challenge and the collaboration process
with their unique perspectives and expertise, how long they have been in Openideo does
not matter. This finding resonates well with the principle Openideo was built upon, which
is to bring design professionals and anybody else to a global community so they can work
together to design solutions for the world’s biggest challenges.
The effect of geographic marginality was not significant either. Openideo was set
to be an inclusive community, which encourages people from all over the world to bring
in their own experience and expertise to contribute to open social innovation. However
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
106
the results showed that people located in major innovation centers were more likely to be
selected as collaborators. This finding showed that the global collaboration principle in
Openideo has not been implemented well by encouraging people located in different
geographical locations to be part of the community. This raises some issues concerning
how to promote Openideo to be a more inclusive global community that is dedicated to
solving social challenges. In particular, strategies need to be developed to encourage the
inclusion of people located in the global South to be part of collaborative crowdsourcing
to contribute to solving the world’s biggest challenges.
Driving the collaboration through experience and community influence
The effect of preferential attachment in predicting tie formation probability in
collaborative crowdsourcing has been mainly attributed to team leadership, community
influence, and winning status. The results showed that compared to ordinary team
members, team leaders who took the initiative to coordinate tasks and recruit talent for
the collective were more likely to be selected into teams in the future. The threshold to
become a team leader is very low, i.e. everybody can become a team leader as long as
you identify an Openideo challenge and create a team to submit a final proposal. Though
there is no clear hierarchy in terms of what roles team leaders and non-team leaders play
in Openideo, their responsibilities differ in the sense that team leaders are often more
involved in coordinating other team members’ contributions and writing up the final
proposal to be published on Openideo. Therefore, the role of team leader may indicate a
higher level of commitment to the community and a higher level of expertise in managing
collective intelligence and handling potential risks rising from virtual team collaboration.
These characteristics could make them more likely to be selected as collaborators.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
107
Furthermore, individuals who are already popular and influential (indicated by
higher degree centrality) in the Openideo community were more likely to be selected into
teams in the future. The higher the degree centrality, the more central a node is in the
network. Central positions mean better access to information and different social groups,
which could help expand the collaboration network. This finding showed that individuals
who have higher degree centrality scores will continue to drive the collaboration network
in Openideo. Together with the marginality effect of betweenness centrality on tie
formation, it shows that network positions could play different roles in explaining how
collaboration takes place in Openideo.
The last variable explaining the preferential attachment mechanism is the
innovation performance in Openideo. The results showed that individuals who won at
least one Openideo challenge were more likely to be selected into teams. This is
consistent with the literature on online collaboration. In online context, collaboration is
often filled with higher uncertainty due to lack of cues. Therefore, people often look for
signals that could reveal a person’s expertise or past performance. In Openideo,
individual user’s profile page offers information on whether or not this person won any
challenge in the community.
The ERGM results also showed that employment information did not help
endorse a person’s expertise or experience. This could be explained by the fact that the
format of employment information is quite flexible. Some users put down student,
freelance, entrepreneur and other general categories as employment, which did not
provide detailed clues about their experience and expertise. Also, the proportion of team
players who provided employment information was relatively small, which was 39%.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
108
Facilitating fluid collaboration across different sub-communities
The overall network structure of team collaboration on Openideo also showed the
co-existence of marginality effect and the effect of preferential attachment, evidenced by
the co-periphery structure. This is consistent with network structures found in online
communities (Shaw & Hill, 2014). This indicates that members in Openideo strategically
selected whom they collaborated with to participate in crowdsourcing challenges. The
marginality effect showed that in accordance with the openness principle, members with
high marginality in certain dimensions were more likely to be selected into teams as they
may bring their unique skills set, knowledge or views into the collective pool. However,
members of higher influence in the community or located in innovation centers were also
more likely to be selected into teams, as they may possess key skill sets in facilitating and
coordinating individual contributions and thus can help improve team performance.
Furthermore, the team collaboration network in Openideo also showed a strong
tendency towards transitivity. This suggests that people tend to work with their
collaborator’s existing collaborators. Open collaboration environment enables free
information flow among members of a crowd and people have the capability to work with
whomever they want. However, openness does not come without risks or uncertainties.
To ensure task success or higher relational satisfaction, crowdsourcing participants seem
to prefer working with others with whom they share common collaborators.
The strong tendency towards the division of the nodes into sub-communities
raises the issue of how to facilitate and sustain distributed collaboration in a
crowdsourcing community. User participation in an open collaboration environment is
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
109
often ad-hoc and spontaneous. A higher tendency towards division might result in
clustered collaboration within sub-groups and a lack of collaboration across different sub-
communities, which could create barriers for information sharing and expertise exchange
in the community. As Smith and Reilly (2013) suggested, open collaboration models are
composed of complex processes, which evolve around content, people, and outcomes. In
other words, “openness is perhaps best understood as a collective process that is
continuously under development and review, rather than as a fixed endpoint that can be
constructed” (Harvey, 2011: 30). To facilitate and sustain the process of open
development models, certain network intervention strategies should be applied in
Openideo to encourage the fluid collaboration across different subgroups (Valente, 2010).
For example, opinion leaders can be identified to advocate for a new crowdsourcing
challenge. They are people who can influence others’ opinions, attitudes, beliefs,
motivations, and behaviors (Rogers & Cartano, 1962). They can also function as liaisons
between sub-communities. – add more elaboration.
Building successful team through geographic and functional diversity
Study 2 was set to examine what team composition could lead to better team
performance. Guided by the literature on team diversity and the Group Faultline theory,
this study tested which demographic diversity and cognitive diversity factors improve
team performance, which was measured by the possibility of winning Openideo
challenges (i.e. collectively locating the global optimum, as defined in diversity
literature). Logistic regression was conducted to test the hypotheses. Results showed that
the effect of diversity is context specific. It found that certain diversity measures can help
improve team performance through generative co-creation, as members’ diverse expertise
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
110
and experience help them build upon each other’s existing ideas and come up with a
better design collectively. However, higher diversity derived from team members’
community tenure differences could divide a team into subgroups, activate faultlines and
generate disruptive influences that would lead to worse team performance.
Leverage the benefits of diverse experience and expertise within teams
The benefits of higher team diversity were mainly attributed to the following
cognitive diversity measures: evaluation skills diversity and winning experience
diversity. The regression results showed that if members in a team possess higher
variation in terms of their evaluation skills, this team would be more likely to win
Openideo challenges. This could be explained by the fact that evaluation processes on
Openideo help members to revise their designs. However, crowdsourcing evaluation does
not rely on experts. Openideo encourages everybody to comment on existing ideas and
participate in the evaluation stage. Members with higher evaluation skills may indicate
that they are more active in giving feedback to others; while members with lower
evaluation skills are people who might excel in other tasks. Having members of
heterogeneous evaluation skills allows the team to have access to diverse skill sets and
thus increases its winning chances.
The effect of winning experience diversity was only positive when tested on how
many winning ideas each team member had on Openideo, but negative when tested on
how many unique challenges a team member won on Openideo. The results suggest that
individuals who failed or had few winning ideas can bring in lessons to team
collaboration and reflect on how to improve future designs. The positive effect on the
variation of winning ideas rather than winning challenges could be explained by the fact
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
111
that there were limited numbers of public challenges on Openideo (N = 36). Therefore,
when assessing how to build a team that is more likely to win, Openideo participants tend
to pay more attention with regard to the total number of ideas an individual submitted and
how many this person won so they can get a better sense of their peers’ skill sets or
history of participating in crowdsourcing.
Another diversity measure that positively influences team performance was
geographic diversity. The finding highlights the importance of having a global team so
people located in different regions could bring in their own stories and experiences. As
discussed earlier, people located in main innovation centers of Openideo (US and UK)
were more likely to be selected into teams. Study 2 found that a winning team should
include members from diverse geolocations. This further highlights the importance of
building a global community when people located in different countries and regions can
bring in their experience through open sharing and open collaboration. As de Beer and
Oguamanam (2013: 264) suggested, “openness applied to international development
bears significant promise for shifting the conceptual paradigms that dominated the latter
half of the twentieth century”, which is to bring in experiences from different societies
and cultures and utilize the collective intelligence in the global community.
Study 2 also found that team members’ status difference did not significantly
impact on team performance. This indicates that there might be a division between users
who had leadership experience in Openideo and who did not, but the difference generated
from leadership experience had no significant effect on team performance. The reason
might be that it is a team leader’s responsibility to encourage communication and
collaboration among team members so individual contributions can be coordinated
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
112
towards a common goal, which is to win a crowdsourcing challenge. The overall
collaborative and open environment enables everybody to get equalized autonomy to
engage in the innovation process. Therefore, team leadership in the context of
collaborative crowdsourcing does not entail hierarchical roles within a team.
Integrating newer members into a community
The disruptive influences on team performance came from differences in
members’ community tenure that activates the group faultlines. Members of higher tenure
tend to differ from members of lower tenure in terms of their experience in
crowdsourcing or in Openideo community in particular. This division has a significant
and negative impact on how the team performs. The difference in community tenure
could reflect different levels of knowledge about the community, different levels of
interpersonal skills in managing peer relationships, or different levels of knowledge about
crowdsourcing processes. When members of longer tenure prefer cooperating with each
other but not with newer members, this could lead to inter-subgroup conflict and thus
create barriers for group efficiency and effectiveness. Though Study 1 found that tenure
did not have a significant effect on choosing team members, tenure plays a significant
role in building a winning team.
Furthermore, it was found that teams with higher mean community tenure are
more likely to win in crowdsourcing. This suggests that to build a winning team,
members of longer experience in a crowdsourcing community are preferred. Together
with the tenure faultline finding, the results raise the question of how to integrate
newcomers in a crowdsourcing community. Open collaboration does not require formal
membership and user engagement can be spontaneous. Without an effective mechanism
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
113
for socializing newer members, a crowdsourcing community might face high turnover or
less regular contribution. The existence of tenure faultlines could make it worse.
In Openideo, offline meetups have been organized in some cities (such as
Chicago, Palto Alto, New York, Barcelona, and Vienna) which can function as a platform
for members of different tenures to engage in conversation and exchange ideas. In 2013,
Openideo formally defined these local offline meetups as “Openideo Chapters”. So far
these chapters have expanded to 40 cities. Openideo reviews applications on a regular
basis to continue establishing new chapters to build their global community.
Some other strategies can be implemented such as adding social network features
to Openideo so members can leverage the benefits of inter-dependence in the community.
For example, a forum can be created on Openideo where people can advertise their teams
and post information for talent recruitment. So far no formal recruitment mechanism is
available on Openideo. Team leaders review potential candidates’ existing team
affiliations and submitted ideas to select team members, which could place new comers
in a disadvantaged position. By establishing an effective recruitment system that provides
a more equalized chance for members, Openideo not only functions as a platform for
project coordination and development but also as a place built upon social networks to
nurture generative co-creation. Newer members in the community can easily locate talent
requirement for different teams and projects and establish connections.
Study 2 also found that team size is not a barrier for crowdsourcing success. This
is consistent with recent literature in collective action. In online communities, larger
teams may face more challenges in motivating and coordinating individual contributions.
However, the environment in collaborative crowdsourcing enables any member in the
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
114
community to take on different roles and customize the way of making contribution. This
suggests that forming a winning team is not constrained by the concern of size. However,
given the finding on the strong tendency towards divided sub-communities, larger teams
need to apply effective coordination and integration strategies to tackle the potential
division issues.
Existing literature on team diversity has not reached a consensus in term of how
demographics and cognitive diversity affect team performance. Study 2 applied the
theoretical framework that takes into account both the dispersion and the pattern of team
members attributes (Thatcher & Patel, 2012; Shaw, 2004). By extending the multi-
dimensional conceptualization of team diversity into the context of crowdsourcing, it
provides a glimpse into team collaboration dynamics and outcomes.
Limitations and Future Research Agenda
This dissertation has several limitations. First of all, the data were not
representative of all collaborative crowdsourcing communities, which constitute a diverse
population. The current dissertation is a quantitative case study with a particular
crowdsourcing community, thus we should exercise caution in generalizing these
findings. However, by studying a global crowdsourcing community that focuses on
utilizing collective intelligence to solve social issues, this dissertation contributes to the
limited knowledge of open development models and open social innovation. It
empirically tests under what conditions we can better aggregate the wisdom of a crowd
for social good. More research should be conducted to further examine the process of
collaborative crowdsourcing in the context of social innovation to fill the gap in the
literature.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
115
Second, this dissertation only employs public behavior data and network data on
Openideo and it lacks analysis on how crowdsourcing participants perceive their
collaboration process and outcomes. In the advent of modern technological environment
where communication data can be exploited in real time, large-scale behavioral data and
network data allow us to analyze the process of organizing collective endeavors from a
relational perspective. Moving beyond characteristics of collective action contributors,
network analysis takes into account a number of important relational properties of inter-
connected entities, such as tie strength (how often people collaborate), reciprocity (how
often people support each other’s crowdsourcing initiatives), and centralization (the
degree to which a key set of people control the collaboration flow in the network). It
allows us to uncover structural characteristics of user behavior.
However, measuring user perception and attitude could prove useful when
examining the dynamics of crowdsourcing, particularly how individuals perceive the
group faultlines and relational outcomes within teams. Future research will design a
survey to measure how members in Openideo perceive the value of team collaboration,
under what condition they prefer working with others, what motivates them to participate
in teams, what factors drive them to join multiple teams, what team composition they
view as more effective for collective success, and how they define relational and team
success. Some other questions will also be included, such as how they perceive the co-
existence of competition and collaboration in Openideo, and what factors influence their
decision of whom to work with. As Openideo sets out to bring social impact through
crowdsourcing efforts, questions on how the members perceive the real life impact will
also be examined.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
116
Another related research area is to conduct text analysis on users self-reported
bios, which provide detailed information on their interest, expertise, and expectation from
Openideo, among other things. Topic modeling techniques will be used in future research
to identify how users brand themselves in terms of specifying their expertise and
knowledge sets to others, what motivate them to join Openideo, and what they plan to
contribute. This will also further examine how employment information is related to
team member selection and innovation performance.
The third limitation lies in the assumption that the relationship between team
composition and task outcomes was linear. Literature on team diversity has found that the
effect of diversity on team performance could change over time. For example, gender
diversity has a negative effect on a team’s innovation performance however the negative
effect would decrease as members engage in more interaction to handle potential
communication and relational issues (Mohammed & Angell, 2004). Furthermore, team
interaction processes can impact on the relationship between team composition and team
performance. To better examine the effect of group faultlines and how to build winning
teams, future research will take into account other factors such as team orientation,
emerging leaderships, communication patterns, and relational outcomes to examine the
curvilinear relationships between team diversity and team performance. It will also take
into account time elements to uncover how the effects evolve across different stages of a
crowdsourcing challenge.
Fourth, the dissertation only focused on how to increase the chance of winning at
the team level. In Openideo, some members of the crowd won a challenge individually,
some won in teams, while some others won in both forms. In addition, some users joined
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
117
multiple teams. Therefore, further analysis should be conducted for all the Openideo
members to uncover the winning patterns. The following research questions can be
examined: Are people on the margin more likely to win as individuals or in teams? Does
team collaboration make individual members more likely to win crowdsourcing
challenges?
Last but not least, this dissertation only looks at the collaboration patterns at the
user level. There are two other types of collaboration missing in the current work. First,
what remains unexamined is the structure of the team-team affiliation network. The user
level collaboration captures the micro level generative co-creation activities, while the
team level collaboration could provide insights on how openness unfolds to a broader
environment. Future research will look into what members are more likely to join what
teams, what teams are more likely to function across different categories of challenges,
and whether there is pattern in connecting teams and challenges.
The second type of collaboration that will be examined in future work is strategic
alliances for open innovation at the organizational level. Openideo emphasizes the cross-
sector collaboration among different types of organizations. The sampled 34 Openideo
challenges were sponsored by 39 organizations, including Unilever, Oxfam, Barclays,
Jamie Oliver, USAID, National Broadcasting Company, National Environment Agency
Singapore, Queensland Government, IDEO, The Amplify Program, The Clinton Global
Initiative, MasterCard, European Commission, Coca-Cola Enterprises, Grameen Creative
Lab, Amnesty International, Steelcase inc., The Information Technology and Innovation
Foundation, The White house office of Science and Technology Policy (OSTP), and
others.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
118
These organizations belong to different sectors. Future research will examine the
collaboration patterns among these organizations in terms of who are more likely to
sponsor crowdsourcing social challenges together, what organizations are more likely to
set the agenda in crowdsourcing alliances, and what structures of the inter-organizational
collaboration lead to higher social impact of the challenges. Given the small sample size
of the crowdsourcing challenges, more examples of applying crowdsourcing to solve
social development problems will be identified from intermediary organizations, such as
United Nations Global Pulse and InnoCentive.
Conclusion
Guided by a network perspective and existing literature on team diversity, the
dissertation sets out to investigate an under-examined phenomenon: applying
crowdsourcing as a model for solving social development problems (such as climate
change, social inequality, community development, and public health concerns).
Crowdsourcing, an innovative way of problem solving by utilizing the wisdom of
crowds, affords great potential in aggregating individuals’ expertise and experience to
contribute to social goods. Suited in the global crowdsourcing community Openideo, the
dissertation employs user behavioral and network data throughout its life span of over 5
years.
Drawing upon studies on open collaboration and online communities, the first
study in the dissertation tested what factors influence team formation in collaborative
crowdsourcing. Specifically, it examines the co-existence of two network mechanisms in
influencing how members choose their team members: marginality and preferential
attachment. The results from ERGM show that individuals with higher research skills are
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
119
more likely to be chosen into teams and individuals located on the margin of evaluation
and collaboration sills are also more likely to be chosen into teams. This suggests that in
Openideo, there is a division of labor to better leverage the open collaboration
environment so team members can build upon each other’s ideas and contributions to
collectively win crowdsourcing challenges. Other results from Study 1 show that
individuals’ network positions can have different influences on how they select future
team members, thus highlighting the relational aspects of crowdsourcing. The preferential
attachment effects were attributed to members’ experience in creating and leading teams,
experiencing in winning challenges, and their proximity to Openideo innovation centers.
Study 1 provides empirical support on what factors drive collaboration patterns among
crowdsourcing patterns.
To further examine the dynamics of collaborative crowdsourcing, Study 2 was
conducted to uncover how to build winning teams. It investigates the relationship
between team composition and team performance. Literature on team diversity was
reviewed to propose what dimensions of diversity may generate disruptive influences on
task outcomes and what diversity can lead to better performances. Without making a
simple claim that higher diversity at the team level indicates more openness, Study 2
measures the multi-dimensional features of diversity and examines how to build a
winning team to contribute to social goods. Logistic regression results highlight the
importance of geographic diversity for open collaboration.
Study 2 also found that in a team where some members have more winning ideas
while others have no winning experience, this type of experience diversity could help the
team improve its chance of winning. Some warning sign will show up if a team has a
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
120
higher diversity from team members’ community tenure. The results show that this type
of heterogeneity at the team level will activate group faultlines to divide the team into
subgroups, and lead to worse team performance.
The dissertation fills the knowledge gap in the literature on crowdsourcing by
focusing on how to apply this problem solving model to social innovation, and by
examining crowdsourcing dynamics through team formation mechanisms. Implications
were drawn on how to sustain crowdsourcing participation, how to integrate newcomers
to the community, and how to tackle the challenge from sub-community division. The
discussion highlights the value of marginality and diversity in influencing crowdsourcing
success. It provides insights on under what conditions we can aggregate the wisdom of a
crowd and apply the collaborative crowdsourcing model to contribute to greater social
goods.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
121
References
Adamczyk, S., Bullinger, A. C., & Möslein, K. M. (2012). Innovation contests: A review,
classification and outlook. Creativity and Innovation Management, 21(4), 335-
360.
Afuah, A., & Tucci, C. L. (2012). Crowdsourcing as a solution to distant search. Academy
of Management Review, 37(3), 355-375.
Alchian, A.A., & Demsetz, H. 192. Production, information cost, and economic
organization. American Economic Review, 62: 777-795.
Aldrich, H., & Reuf, M. (2006). Organizations evolving (2nd ed.). Sage Publications.
Almirall, E., R., & Casadesus-Masanell, R. (2010). Open vs. integrated innovation: A
model of discovery and divergence. Academic Management Review. 35(1). 27–
47.
Amabile, T. M. (1983). The social psychology of creativity: A componential
conceptualization. Journal of personality and social psychology, 45(2), 357.
Amason, A. (1996). Distinguishing the effects of functional and dysfunctional conflict on
strategic decision making: Resolving a paradox for top management teams. The
Academy of Management Journal, 39(1), 123–148.
Ancona, D. G., & Caldwell, D. F. (1992). Bridging the boundary: External activity and
performance in organizational teams. Administrative science quarterly, 634-665.
Antikainen, M., Makipaa, M. & Ahonen, M. (2010). Motivating and supporting
collaboration in open innovation. European Journal of Innovation Management.
13(1). 100-119. DOI 10.1108/14601061011013258
Antikainen, M. J., & Vaataja, H. K. (2010). Rewarding in open innovation communities-
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
122
how to motivate members. International Journal of Entrepreneurship and
Innovation Management, 11(4), 440-456.
Anson, R., Bostrom, R., & Wynne, B. (1995). An experiment assessing group support
system and facilitator effects on meeting outcomes. Management Science, 41(2),
189-208.
Avital, M., & Te'eni, D. (2009). From generative fit to generative capacity: exploring an
emerging dimension of information systems design and task performance.
Information Systems Journal, 19(4), 345-367.
Baldwin, C. Y., & von Hippel E (2011) Modeling a paradigm shift: From producer
innovation to user and open collaborative innovation. Organization Science.
22(6): 1399–1417. http://dx.doi.org/10.1287/orsc.1100.0618
Bantel, K. A., & Jackson, S. E. (1989). Top management and innovations in banking:
Does the composition of the top team make a difference? Strategic Management
Journal, 10(S1), 107-124.
Bar, F., & Riis, A. M. (2000). Tapping user-driven innovation: A new rationale for
universal service. The Information Society, 16(2), 99-108.
Barabási, A. (2002). Linked: The new science of networks. Cambridge, Mass.: Perseus.
Baum, J. A. C., Calabrese, T. & Silverman, B. S. (2000). Don't go it alone: Alliance
network composition and startups' performance in Canadian biotechnology.
Strategic Management Journal. 21(3). 267-294.
Bayus, B. L. (2013). Crowdsourcing new product ideas over time: An analysis of the Dell
IdeaStorm community. Management Science, 59(1), 226-244.
Bell, S. T., Villado, A. J., Lukasik, M. A., Belau, L., & Briggs, A. L. (2010). Getting
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
123
specific about demographic diversity variable and team performance
relationships: A meta-analysis. Journal of management. 37(3). 709-743 DOI:
10.1177/0149206310365001
Belleflamme, P., Lambert, T., & Schwienbacher, A. (2014). Crowdfunding: Tapping the
right crowd. Journal of Business Venturing, 29(5), 585-609.
Benkler, Y. (2013) Foreword. In Smith, M. & Reilly, K. (ed). Open Development:
Networked Innovations in International Development. vii-ix
Benkler, Y. (2006). The wealth of networks: How social production transforms markets
and freedom. Yale University Press.
Bengtsson, L. & Ryzhkova, N. (2013). Managing Online Users in Open Innovation The
Case of a Nordic Telecom Company. Digiworld Economic Journal, 89(1): 37-53.
Benbya, H., & Belbaly, N. (2010). Understanding developers’ motives in open source
projects: a multi-theoretical framework. Communications of the AIS, 27, 589-610.
Bessiere, K., Ellis, J., & Kellogg, W. (2009). Acquiring a professional Second Life:
Problems and prospects for the use of virtual worlds in business. In Proceeding of
the Twenty Seventh Annual SIGCHI Conference on Human Factors In Computing
Systems.
Bimber, B., Flanagin, A., & Stohl, C. (2012). Collective action in organizations:
Interaction and engagement in an era of technological change. Cambridge
University Press.
Bimber, B. A., Flanagin, A. J., & Stohl, C. (2005). Reconceptualizing collective action in
the contemporary media environment. Communication Theory, 15, 365-388.
Blohm, I., Bretschneider, U., Leimeister, J. M., & Krcmar, H. (2011). Does collaboration
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
124
among participants lead to better ideas in IT-based idea competitions? An
empirical investigation. International Journal of Networking and Virtual
Organisations, 9(2), 106-122.
Blondel, V. D., Guillaume, J., Lambiotte, R., & Lefebvre, E. (2008). Fast unfolding of
communities in large networks. Journal of Statistical Mechanics: Theory and
Experiment, 10, P10008 (12pp). doi: 10.1088/1742-5468/2008/10/P10008. ArXiv:
http://arxiv.org/abs/0803.0476
Bons, P. M., & Fiedler, F. E. (1976). Changes in organizational leadership and the
behavior of relationship- and task-motivated leaders. Administrative Science
Quarterly, 21: 453–473.
Boudreau, K.J., (2012). Let a thousand flowers bloom? an early look at large numbers of
software app developers and patterns of innovation. Organization Science, 23(5),
1409–1427.
Boudreau, K. (2010). Open Platform Strategies and Innovation: Granting Access vs.
Devolving Control. Management Science. 56(10). 1849-1872. doi
10.1287/mnsc.1100.1215
Boudreau, K. J. & Lakhani, K. R. (2013). Using the Crowd as an Innovation Partner.
Harvard Business Review. April, 60-69.
Boudreau, K. J., Lacetera, N., & Lakhani, K. R. (2011). Incentives and problem
uncertainty in innovation contests: An empirical analysis. Management Science.
57(5): 843-863.
Boughzala, I., De Vreede, G. J., & Limayem, M. (2012). Team collaboration in virtual
worlds: Editorial to the special issue. Journal of the Association for Information
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
125
Systems, 13(10), 714.
Brabham, D. C. (2008a). Crowdsourcing as a model for problem solving: an introduction
and cases. Convergence: The International Journal of Research into New Media
Technologies, 14(1). 75–90.
Brabham, D. C. (2008b). Moving the crowd at iStockphoto: the composition of the crowd
and motivations for participation in a crowdsourcing application, First Monday,
13(6). Available at:
http://www.uic.edu/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/2159/1969.
Brabham, D. C. (2012). The myth of amateur crowds: A critical discourse analysis of
crowdsourcing coverage. Information, Communication & Society, 15(3), 394-410.
Brabham, D. C. (2013). Crowdsourcing. Cambridge, MA: MIT Press.
Brandon, D. P., & Hollingshead, A. B. (2004). Transactive memory systems in
organizations: Matching tasks, expertise, and people. Organization Science, 15(6),
633-644.
Brass, D. J., & Burkhardt, M. E. (1993). Potential power and power use: An investigation
of structure and behavior. Academy of management journal. 36(3). 441-470.
Brito, J. (2008). Hack, mash, & peer: crowdsourcing government transparency. The
Columbia Science and Technology Law Review, 9: 119–157.
Brunswicker, S., Bertino, E., & Matei, S. (2015). Big Data for Open Digital Innovation–
A Research Roadmap. Big Data Research, 2(2), 53-58.
Budhathoki, N. R., & Haythornthwaite, C. (2013). Motivation for open collaboration
crowd and community models and the case of OpenStreetMap. American
Behavioral Scientist, 57(5), 548-575.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
126
Buecheler, T., Sieg, J.H., Füchslin, R.M., Pfeifer, R., (2010). Crowdsourcing, open
innovation and collective intelligence in the scientific method: a research agenda
and operational framework. In: Artificial life XII. Proceedings of the Twelfth
International Conference on the Synthesis and Simulation of Living Systems,
Odense, Denmark, pp. 19–23.
Bullinger, A. C., Neyer, A. K., Rass, M., & Moeslein, K. M. (2010). Community-based
innovation contests: where competition meets cooperation. Creativity and
innovation management, 19(3), 290-303.
Bulnes, D. M. (2011). Social innovation: Crowdsourcing and the new face of corporate
social responsibility. University of Southern California.
Bunderson, J. S., & Sutcliffe, K. M. (2002). Comparing alternative conceptualizations of
functional diversity in management teams: Process and performance effects.
Academy of management journal, 45(5), 875-893.
Burt, R. S. (1992). Structural holes: The social structure of competition. Cambridge, MA:
Harvard University Press.
Burt, R. S. (2009). Structural holes: The social structure of competition. Harvard
university press.
Butler, B. (2001). Membership size, communication activity and sustainability: A
resource-based model of online social structures. Inform. Systems Res. 12(4):
346–362.
Butler, B. S., Sproull, L., Kiesler, S., & Kraut, R. (2007). Community effort in online
groups: Who Does the Work and Why? In Leadership at a Distance, S. Weisband
& L. Atwater (eds.), Mahwah, NJ: Lawrence Erlbaum Associates.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
127
Byrne, B. M. (2013). Structural equation modeling with LISREL, PRELIS, and SIMPLIS:
Basic concepts, applications, and programming. Psychology Press.
Campbell, L. (2009). Dotmocracy: crowdsourcing, mashups, and social change,
Available at: http://www.mobilerevolutions.org/Dotmocracy.pdf
Carlile, P. R. (2002). A pragmatic view of knowledge and boundaries: Boundary objects
in new product development. Organization science, 13(4), 442-455.
Castells, M. (2012). Networks of outrage and hope: Social movements in the age of the
Internet. Polity Press: Cambridge, MA
Castells, M. (2006). The theory of the network society. Great Britain by MPG Books Ltd,
Bodmin, Cornwall.
Castells, M. (1996). The rise of the network society. Vol. 1 of The information age:
Economy, society and culture. Massachusetts and Oxford: Blackwell.
Chatman, J. A., Polzer, J. T., Barsade, S. G., & Neale, M. A. (1998). Being different yet
feeling similar: The influence of demographic composition and organizational
culture on work processes and outcomes. Administrative Science Quarterly, 749-
780.
Chesbrough, H.W. (2003). Open Innovation: The New Imperative for Creating and
Profiting from Technology. Harvard Business School Press: Boston, MA.
Chesbrough, H., & Bogers, M. (2014). Explicating open innovation: clarifying an
emerging paradigm for understanding innovation. In H.W. Chesbrough, W.
Vanhaverbeke, J. West, New Frontiers in Open Innovation. Chapter1, 3-28.
Oxford: Oxford University Press.
Chesbrough, H., & Di Minin, A. (2014). Open social innovation. In New Frontiers in
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
128
Open Innovation, H. Chesbrough, W. Vanhaverbeke, & J. West (eds). Chapter 9,
pp. 169-187.
Chesbrough, H.W., Vanhaverbeke, W., & West, J. (2014). New Frontiers in Open
Innovation. Oxford: Oxford University Press.
Chesbrough, H.W., Vanhaverbeke, W., & West, J. (2006). Open Innovation: Researching
a New Paradigm. Oxford University Press: Oxford.
Chesbrough, H.W. (2006), “Open innovation: a paradigm for understanding industrial
innovation”, in Chesbrough, H., Vanhaverbeke, W. and West, J. (Eds), Open
Innovation: Researching a New Paradigm, Oxford University Press, Oxford, pp.
1-34.
Cheliotis, G. (2009). From open source to open content: organization, licensing and
decision processes in open cultural production. Decision Support Systems. 47.
229-244.
Cheliotis, G., & Yew, J. (2009, June). An analysis of the social structure of remix culture.
In Proceedings of the fourth international conference on Communities and
technologies (pp. 165-174). ACM.
Chesbrough, H. W. & Appleyard, M. M. (2007). Open innovation and strategy. California
Management Review, 50(1), 57-76.
Cohen, M. D., March, J. G., & Olsen, J. P. (1972). A garbage can model of organizational
choice. Administrative science quarterly, 1-25.
Cohen, S. G., & Bailey, D. E. (1997). What makes teams work: Group effectiveness
research from the shop floor to the executive suite. Journal of management, 23(3),
239-290
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
129
Coleman, J. S. (1973). Mathematics of collective action. Transaction Publishers.
Coleman, J. S. (1990). The foundation of social theories. Cambridge, MA: Harvard
University Press.
Cullina, E., Conboy, K., & Morgan, L. (2015, August). Measuring the crowd: a
preliminary taxonomy of crowdsourcing metrics. In Proceedings of the 11th
International Symposium on Open Collaboration (p. 7). ACM.
Cummings, S., Daellenbach, U., Davenport, S., & Campbell, C. (2013). “Problem-
sourcing”: a re-framing of open innovation for R&D organizations. Management
Research Review. 36(10). 955-974. DOI 10.1108/MRR-07-2012-0177
Dahlin, E. (2011). There’s No “I” in Innovation. Contexts, 10(4), 22-27.
Dankbaar, B. (2007). Global sourcing and Innovation: The consequences of losing both
organizational and geographical proximity. European Planning Studies. 15(2).
271-288
David, P. A. & Rullani, F. (2008). Dynamics of innovation in an “open source”
collaboration environment: lurking, laboring, and launching FLOSS projects on
SourceForge. Industrial and Corporate Change. 17(4). 647-710.
doi:10.1093/icc/dtn026
De Beer, J., Sowa, I., and Holman, K. (2013). Frameworks for analysing African
innovation: Entrepreneurship, the informal economy and intellectual Property, in
De Beer, J., Armstrong, C., Oguamanam, C. and Schonwetter, T. (Eds.),
Innovation and intellectual property: Collaborative dynamics in Africa, UCT
Press, Cape Town, www.openair.org.za/content/open- air-publications.
Dissanayake, I., Zhang, J., & Gu, B. (2014). Virtual Team Performances in
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
130
Crowdsourcing Contests. Paper presented at the 12th Americas Conference on
Information Systems, Savannah
Dittrich, K., & Duysters., G. (2007). Networking as a means to strategy change: the case
of open innovation in mobile telephony. Journal of Product Innovation
Management 24(6). 510-521.
Donaldson, T. & Preston, L. E. (1995). The stakeholder theory of the corporation:
Concepts, evidence, and implications. Academy of management Review 20(1).
65-91.
Duarte, V., & Sarkarm S. (2011). Separating the wheat from the chaff – a taxonomy of
open innovation. European Journal of Innovation Management. 14(4). 435-459.
DOI 10.1108/14601061111174907
Dunbar, K. (1998). Problem solving. In W. Bechtel, & G. Graham (Eds.). A companion
to Cognitive Science. London, England: Blackwell, pp 289-298
Earley, C. P., & Mosakowski, E. (2000). Creating hybrid team cultures: An empirical test
of transnational team functioning. Academy of Management Journal, 43(1), 26-
49.
Eden, C. (1990). The unfolding nature of group decision support - two dimensions of
skill. In C. Edena & J. Radford (Eds.),
Erickson, L.B., Petrick, I., Trauth, E.M., (2012). Hanging with the right crowd: matching
crowdsourcing need to crowd characteristics. In the Proceedings of the 18th
Americas Conference on Information Systems, August 9–12, Seattle, Washington.
Estellés-Arolas, E., Gonzalez-Ladron-de-Guerva, F. (2012). Towards an integrated
crowdsourcing definition. Journal of Information Science 38 (2), 189–200.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
131
Etzkowitz, H., C. Kemelgor, B. Uzzi. 2000. Athena Unbound: The Advancement of
Women in Science and Technology. Cambridge University Press, Cambridge, UK
Hambrick, D. C., & Mason, P. A. (1984). Upper echelons: The organization as a
reflection of its top managers. Academy of management review, 9(2), 193-206.
Heeks, R. (2008). ICT4D 2.0: The next phase of applying ICT for international
development. IEEE, 41(6), 78–85.
Faems, D., De Visser, M., Andries, P. & Van Looy, B. (2010). Technology alliance
portfolios and financial performance: Value-enhancing and cost-increasing effects
of open innovation. Journal of Product Innovation Management. 27(6). 785-796
Faraj, S., Jarvenpaa, S. L., & Majchrzak, A. (2011). Knowledge collaboration in online
communities. Organization Science. 22(5). 1224-1239.
http://dx.doi.org/10.1287/orsc.1100.0614
Faust, K. (2010). A puzzle concerning triads in social networks: Graph constraints and the
triad census. Social Networks, 32(3), 221-233.
Fernandes, R., & Simon, H. A. (1999). A study of how individuals solve complex and ill-
structured problems. Policy Sciences, 32(3), 225-245.
Festinger, L. (1954). A theory of social comparison processes. Human relations, 7(2),
117-140.
Fleming, L., & Sorenson, O. (2016). Financing by and for the Masses: An introduction to
the special issue on Crowdfunding. California Management Review, 58(2). 5-19
Franke, N., & von Hippel, E. (2003). Satisfying heterogeneous user needs via innovation
toolkits: The case of Apache security software. Research Policy, 32(7), 1199–
1215.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
132
Freeman, R., & Rogers, J. (2002). Open source unionism: Beyond exclusive collective
bargaining. Working USA: Journal of Labor and Society, 5, 3-4.
Freeman, L. C. (1979). Centrality in social networks: I. Conceptual clarifications. Social
Networks, 1. 215-239.
Freeman, R. E. (1984). Strategic management: A stakeholder approach. Boston: Pitman.
Freeman, R. E., & Evan, W. M. (1990). Corporate governance: A stakeholder
interpretation. The Journal of Behavioral Economics, 19: 337-359.
Freeman, C. (1974). The Economics of Industrial Innovation. Penguin Books,
Harmondsworth.
Fritz, S., McCallum, I., Schill, C., Perger, C., Grillmayer, R., Achard, F., ... &
Obersteiner, M. (2009). Geo-Wiki. Org: The use of crowdsourcing to improve
global land cover. Remote Sensing, 1(3), 345-354.
Fulk, J. (2001). Global network organizations: Emergence and future prospects. Human
Relations-New York. 54(1). 91-100.
Fulk, J., Flanagin, A. J., Kalman, M. E., Monge, P. R., & Ryan, T. (1996). Connective and
communal public goods in interactive communication systems. Communication
Theory, 6(1), 60-87.
Fulk, J. & Yuan, Y.C. (2013). Location, motivation, and social capitalization via
enterprise social networking. Journal of Computer-Mediated Communication, 19:
20-37
Gassmann, O. (2006). Opening up the innovation process: towards an agenda. R&D
Management. 36(3). 223-228.
Gassmann, O. Enkel, E, & Chesbrough, H. D. (2010). The future of open innovation.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
133
R&D Management. 40(3). 213-221.
Gibson, C. B., & Gibbs, J. L. (2006). Unpacking the concept of virtuality: The effects of
geographic dispersion, electronic dependence, dynamic structure, and national
diversity on team innovation. Administrative Science Quarterly, 51(3), 451-495.
Girvan, M., & Newman, M. E. (2002). Community structure in social and biological
networks. Proceedings of the national academy of sciences, 99(12), 7821-7826.
Goh, S., & Wasko, M. (2010). Leader-member relationships in virtual world teams.
Proceedings of the 44th Annual Hawaii International Conference on System
Sciences, 1-11.
Grand, S., von Krogh, G., Leonard, D. & Swap, W. (2004). Resource allocation beyond
firm boundaries: a multi-level model for open source innovation. Long Range
Planning, 37. 591-610.
Granovetter, M. S. (1973). The strength of weak ties. American journal of sociology,
1360-1380.
Greenstein, S. (1996). Invisible hands versus invisible advisors: Coordination
mechanisms in economic networks. In E. Noam & A. Nishuilleabhain (Eds.),
Public networks, public objectives (135–160). Amsterdam: Elsevier Science.
Grewal, R., G. L. Lilien, G. Mallapragada. (2006). Location, location, location: How
network embeddedness affects project success in open source systems.
Management Sci. 52(7) 1043–1056.
Griffith, T. L., Fuller, M. A., & Northcraft, G. B. (1998). Facilitator influence in group
support systems: Intended and unintended effects. Information Systems Research,
9(1), 20-36.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
134
Gurstein, M. (2005). Editorial: Sustainability of community ICTs and its future. The
Journal of Community Informatics, 1(2). Available at http://ci-
journal.net/index.php/ciej/article/viewArticle/230/185
Haklay, M., & Weber, P. (2008). Openstreetmap: User-generated street maps. Pervasive
Computing, IEEE, 7(4), 12-18.
Hann, I. H., Roberts, J., & Slaughter, S. (2004). Why developers participate in open
source software projects: An empirical investigation. Paper presented at the
Twenty-Fifth International Conference on Information Systems, Washington, DC.
Hanneman, R., & Riddle, M. (2005). Introduction to social network methods. Available
from http://faculty.ucr.edu/~hanneman/nettext/
Harrison, D. A., Price, K. H., & Bell, M. P. (1998). Beyond relational demography: Time
and the effects of surface-and deep-level diversity on work group cohesion.
Academy of management journal, 41(1), 96-107.
Holmes, S., & Smart, P. (2009). Exploring open innovation practice in firm-nonprofit
engagements: a corporate social responsibility perspective. R&D Management.
39(4). 394-409.
Hong, L. & Page, S. E. (2001). Problem Solving by Heterogeneous Agents. Journal of
Economic Theory. 97: 123-163.
Horwitz, S. K., & Horwitz, I. B. (2007). The effects of team diversity on team outcomes:
A meta-analytic review of team demography. Journal of management, 33(6), 987-
1015.
Howe, J. (2006). The rise of crowdsourcing. Wired, 14(6). Available at:
http://www.wired.com/wired/archive/14.06/crowds.html.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
135
Howe, J. (2008). Crowdsourcing: Why the power of the crowd is driving the future of
business. Crown, New York.
Hoyt, C. & Blascovich, J. (2003). Transformational and transactional leadership in virtual
and physical environments. Small group research, 34(6). 678-715.
Huizingh, K. R. E. (2011). Open innovation: state of the art and future perspectives,
Technovation, 31(1). 2-9.
Ibarra, H. (1993). Network centrality, power, and innovation involvement: Determinants
of technical and administrative roles. Academy of Management Journal, 36: 471-
501.
Jackson, S. E. (1992). Diversity in the workplace: Human resources initiatives. Guilford
Press.
Jarvenpaa, S. L., & Leidner, D. E. (1998). Communication and trust in global virtual
teams. Journal of Computer-Mediated Communication, 3(4), Available at
http://onlinelibrary.wiley.com/doi/10.1111/j.1083-6101.1998.tb00080.x/full
DOI: 10.1111/j.1083-6101.1998.tb00080.x
Jehn, K. A., Northcraft, G. B., & Neale, M. A. (1999). Why differences make a
difference: A field study of diversity, conflict and performance in workgroups.
Administrative science quarterly, 44(4), 741-763.
Jeppesen, L. B. & Lakhani, K. R. (2010). Marginality and problem-solving effectiveness
in broadcast search. Organization Science. 21(5): 1016-1033.
Judge, T. A., Piccolo, R. F., & Ilies, R. (2004). The forgotten ones? The validity of
consideration and initiating structure in leadership research. Journal of Applied
Psychology, 89: 36–51.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
136
Kane, G. C., Majchrzak, A., Johnson, J., & Chenisern, L. 2009. A longitudinal model of
perspective making and perspective paking within fluid online collectives, in
Proceedings of the 30th International Conference on Information Systems,
Phoenix, AZ.
Kennedy, M. T., & Fiss, P. C. (2009). Institutionalization, framing, and diffusion: The
logic of TQM adoption and implementation decisions among U.S. hospitals.
Academy of Management Journal, 52: 897-918.
Ketchen Jr., D. J., Ireland, R. D., & Snow, C. C. (2007). Strategic Entrepreneurship,
Collabarative Innovation and Wealth Creation. Journal of Strategic
Entrepreneurship. 1: 371-385.
Keupp, M.M., & Gassmann, O., (2009). Determinants and archetype users of open
innovation. R&D Management. 39(4), 331–341.
Kickul, J., & Gundry, L. K. (2001). Breaking through boundaries for organizational
innovation: new managerial roles and practices in e-commerce firms. Journal of
Management, 27(3), 347-361.
Kilduff, M., Angelmar, R., & Mehra, A. (2000). Top management-team diversity and firm
performance: Examining the role of cognitions. Organization Science, 11(1), 21-
34.
Kim, J., Cheng, J., & Bernstein, M. S. (2014). Ensemble: exploring complementary
strengths of leaders and crowds in creative collaboration. In Proceedings of the
17th ACM conference on Computer supported cooperative work & social
computing (pp. 745-755). ACM.
Kirkman, B.L., Rosen, B., Teslul, P. E. & Gibson, C. B. (2006). Enhancing the transfer of
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
137
computer-assisted training proficiency in geographically distributed teams.
Journal of Applied Psychology 91(3).706-716.
Kjørven, O. (2013, May 7). Crowdsourcing the next global development agenda. The
Guardian. Retrieved from http://www.theguardian.com/media-network/media-
networkblog/
Koch, G., Füller, J., & Brunswicker, S. (2011). Online crowdsourcing in the public sector:
how to design open government platforms. In Online Communities and Social
Computing (pp. 203-212). Springer Berlin Heidelberg.
Koehly, L. M., & Pattison, P. (2005). Random graph models for social networks: Multiple
relations or multiple raters. Models and methods in social network analysis, 162-
191.
Kolaczyk, E. D. & Csárdi, G. (2014). Statistical analysis of network data with R (vol. 65).
Springer: Chicago
Hackman, J. R. (2011). Collaborative intelligence: Using teams to solve hard problems.
Berrett-Koehler Publishers.
Hanneman, R. A., & Riddle, M. (2005). Introduction to social network methods.
Riverside, CA: University of California, Riverside. Available at
http://faculty.ucr.edu/~hanneman/
Hargadon, A. B., & Bechky, B. A. (2006). When collections of creatives become creative
collectives: A field study of problem solving at work. Organization Science,
17(4), 484-500.
Harrison, D. A., Price, K. H., Gavin, J. H., & Florey, A. T. (2002). Time, teams, and task
performance: Changing effects of surface-and deep-level diversity on group
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
138
functioning. Academy of management journal, 45(5), 1029-1045.
Harrison, D. A., Price, K. H., & Bell, M. P. (1998). Beyond relational demography: Time
and the effects of surface-and deep-level diversity on work group cohesion.
Academy of management journal, 41(1), 96-107.
Harrison, D. A., & Klein, K. J. (2007). What's the difference? Diversity constructs as
separation, variety, or disparity in organizations. Academy of management review,
32(4), 1199-1228.
Hunter, D. R., Handcock, M. S., Butts, C. T., Goodreau, S. M., & Morris, M. (2008).
ergm: A package to fit, simulate and diagnose exponential-family models for
networks. Journal of statistical software, 24(3), nihpa54860.
Hutter, K., Hautz, J., Fuller, J., Mueller, J., Matzler, K., (2011). Communitition: the
tension between competition and collaboration in community-based design
contests. Creativity and Innovation Management. 20(1), 3–21.
Ibarra, H. (1993). Network centrality, power, and innovation involvement: Determinants
of technical and administrative roles. Academy of Management Journal, 36. 471-
501.
Inauen, M., & Schenker-Wicki, A. (2011). The impact of outside-in open innovation on
innovation performance. European Journal of Innovation Management. 14(4).
496-520.
Inkpen, A.C. & Dinar, A. (1998). Knowledge management processes and international
joint ventures. Organization Science, 9(4), 454–468.
Jackson, S. E., & Ruderman, M. N. (1995). Diversity in work teams: Research paradigms
for a changing workplace. American Psychological Association.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
139
Keegan, B., Gergle, D., & Contractor, N. (2013). Hot off the wiki: Structures and
dynamics of Wikipedia’s coverage of breaking news events. American Behavioral
Scientist, 0002764212469367.
Kelley, T. M., & Johnston, E. (2012). Discovering the appropriate role of serious games
in the design of open governance platforms. Public Administration Quarterly,
504-554.
Kim, J., Cheng, J., & Bernstein, M. S. (2014, February). Ensemble: exploring
complementary strengths of leaders and crowds in creative collaboration. In
Proceedings of the 17th ACM conference on Computer supported cooperative
work & social computing (pp. 745-755). ACM.
Koehly, L. M., & Pattison, P. (2005). Random graph models for social networks: Multiple
relations or multiple raters. Models and methods in social network analysis, 162-
191.
Kuk, G. (2006). Strategic interaction and knowledge sharing in the KDE developer
mailing list. Management Science, 52(7), 1031-1042.
Makhani, K., Fayard, A. Levina, N., Pkorywa, S. H. (2013) Openideo. Harvard Business
School.
Lakhani, K. R., Lifshitz-Assaf, H. & Michael Tushman,. M. (2013). Open Innovation and
Organizational Boundaries: Task Decomposition, Knowledge Distribution and the
Locus of Innovation. Chap. 19. In A. Grandori. (ed.), Handbook of Economic
Organization: Integrating Economic and Organization Theory (355–382).
Northampton, MA: Edward Elgar Publishing
Lampe, C., Larose, R., Steinfield, C., & DeMaagd, K. (2011). Inherent barriers to the use
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
140
of social media for public policy informatics. The Innovation Journal. 16(1): 1-17
Lawrence B. S. (1997). The black box of organizational demography. Organization
Science. 8:1-22
Lau, D. C., & Murnighan, J. K. (2005). Interactions within groups and subgroups: The
effects of demographic faultlines. Academy of Management Journal, 48(4), 645-
659.
Lau, D. C., & Murnighan, J. K. (1998). Demographic diversity and faultlines: The
compositional dynamics of organizational groups. Academy of Management
Review, 23(2), 325-340.
Lee, G. K., & Cole, R. E. (2003). From a firm-based to a community-based model of
knowledge creation: The case of Linux kernel development. Organization
Science, 14(6), 633–649.
Lee, S., & Monge, P. (2011). The coevolution of multiplex communication networks in
organizational communities. Journal of Communication, 61(4), 758-779.
Leimeister, J. M., Huber, M., Bretschneider, U., & Krcmar, H. (2009). Leveraging
crowdsourcing: activation-supporting components for IT-based ideas competition.
Journal of management information systems, 26(1), 197-224.
Leonard-Barton, D. (1995). Wellsprings of knowledge: Building and sustaining the
sources of innovation. Harvard Business Press
Lerner, J., & Tirole, J. (2002). Some simple economics of open source. The journal of
industrial economics, 50(2), 197-234.
Levine, S. S., & Prietula, M. J. (2014). Open collaboration for innovation: Principles and
performance. Organization Science. 25(5): 1414-1433.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
141
http://dx.doi.org/10.1287/orsc.2013.0872
Li, J., & Hambrick, D. C. (2005). Factional groups: A new vantage on demographic
faultlines, conflict, and disintegration in work teams. Academy of Management
Journal, 48(5), 794-813.
Lykourentzou, I., Antoniou, A., & Naudet, Y. (2015). Matching or Crashing? Personality-
based Team Formation in Crowdsourcing Environments. arXiv preprint
arXiv:1501.06313. Available at http://arxiv.org/abs/1501.06313
Lynn, J., Laurence, E., Heinrich, C. J., & Hill, C. J. (2001). Improving governance: A
new logic for empirical research. Washington, D.C.: Georgetown University
Press.
Madsen, T., Woolley, J., & Sarangee, K. (2012). Using Internet-based collaboration
technologies for Innovation: crowdsourcing vs. expertsourcing. In Academy of
Management Annual Meeting, August, Boston.
Majchrzak, A., More, P. H., & Faraj, S. (2012). Transcending knowledge differences in
cross-functional teams. Organization Science, 23(4), 951-970.
Majchrzak, A., & Malhotra, A. (2013). Towards an information systems perspective and
research agenda on crowdsourcing for innovation. The Journal of Strategic
Information Systems, 22(4), 257-268.
Malhotra, A., & Majchrzak, A. (2014). Managing crowds in innovation challenges.
California Management Review, 56(4). 103-124.
Markus, L. M. (2007). The governance of free/open source software projects: monolithic,
multidimensional, or configurational? Journal of Management and Governance.
11(2): 151-163.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
142
Martinez, M. G.; Lazzarotti, V.; Manzini, R.; & Mercedes Sánchez García, M. S. (2014).
Open innovation strategies in the food and drink industry: determinants and
impact on innovation performance. International Journal of Technology
Management. 66(2/3). 212-242
Marwell, G., & Oliver, P. (1993). The critical mass in collective action. Cambridge
University Press.
Masum, H., Schroeder, K., Khan, M., & Daar, A. S. (2013). Open source biotechnology
platforms for global health and development: two case studies. Open
development: Networked innovations in international development, 113-128.
Matzler, K., Strobl, A., & Bailom, F. (2016). Leadership and the wisdom of crowds: how
to tap into the collective intelligence of an organization. Strategy & Leadership,
44(1), 30-35.
McGrath, J. E. (1984). Groups: Interaction and performance (Vol. 14). Englewood Cliffs,
NJ: Prentice-Hall.
Michels, R. (1915). Political parties: A sociological study of the oligarchical tendencies
of modern democracy (E. Paul & C. Paul, trans.). New York, NY: Hearst.
Milliken, F. J., & Martins, L. L. (1996). Searching for common threads: Understanding
the multiple effects of diversity in organizational groups. Academy of
management review, 21(2), 402-433.
Mohammed, S., & Angell, L. C. (2004). Surface-and deep-level diversity in workgroups:
Examining the moderating effects of team orientation and team process on
relationship conflict. Journal of Organizational Behavior, 25(8), 1015-1039.
Mollick, E. (2014). The dynamics of crowdfunding: An exploratory study. Journal of
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
143
business venturing, 29(1), 1-16.
Mollick, E., & Nanda, R. (2015). Wisdom or madness? Comparing crowds with expert
evaluation in funding the arts. Management Science. Published online first.
http://dx.doi.org/10.1287/mnsc.2015.2207
Monge, P. R., & Contractor, N., S. (2003). Theories of communication networks. New
York, NY: Oxford University Press.
Nam, T. (2012). Suggesting frameworks of citizen-sourcing via Government 2.0.
Government Information Quarterly, 29(1), 12-20.
Newell, A., & Simon, H. A. (1972). Human problem solving (Vol. 104, No. 9).
Englewood Cliffs, NJ: Prentice-Hall.
Newman, M. E. (2006). Modularity and community structure in networks. Proceedings of
the national academy of sciences, 103(23), 8577-8582.
Newman, M. E., & Girvan, M. (2004). Finding and evaluating community structure in
networks. Physical review E, 69(2), 026113.
Nieto, M. J. & Santamaria, L. (2007). The importance of diverse collaborative networks
for the novelty of product innovation. Technovation. 27(6-7): 367–377.
Nishii, L. H., & Goncalo, J. A. (2008). Demographic faultlines and creativity in diverse
groups. Research on managing groups and teams, 11, 1-26.
Northcraft, G. B., Polzer, J. T., Neale, M. A., & Kramer, R. M. (1995). Diversity, social
identity, and performance: Emergent social dynamics in cross-functional teams.
Nov, O. (2007). What motivates wikipedians?. Communications of the ACM, 50(11), 60-
64.
Park, C. H., Son, K., Lee, J. H., & Bae, S. H. (2013, February). Crowd vs. crowd: large-
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
144
scale cooperative design through open team competition. In Proceedings of the
2013 conference on Computer supported cooperative work (pp. 1275-1284).
ACM.
Pillis, E. D., & Furumo, K. (2007). Counting the cost of virtual teams. Communications
of the ACM, 50(12), 93-95.
Norris, P. (2002). The Digital Divide, in Digital Divide: Civic Engagement, Information
Poverty and the Internet Worldwide, 3-25, New York: Cambridge University
Press.
Oliver, P. E., & Marwell, G. (1988). The Paradox of Group Size in Collective Action: A
Theory of the Critical Mass. II. American Sociological Review, 1-8.
Olson, M. (1965). The logic of collective action: public goods and the theory of groups.
Cambridge, MA: Harvard University Press.
O’Mahony, S. C., & Ferraro, F. (2007). The Emergence of Governance in an Open
Source Community, Academy of Management Journal. 50(5): 1079-1106
Openideo. (2016). How Openideo works. Available at http://openideo.com/how-open-
ideo-works
Ostrom, E. (1999). Coping with tragedies of the commons. Annual Review of Political
Science, 2, 493–535.
Ostrom, E. 1990. Governing the commons: the evolution of institutions for collective
action. Cambridge University Press, Cambridge, UK.
Osterloh, M., & Rota, S. (2007). Open source software development—Just another case
of collective invention?. Research Policy, 36(2), 157-171.
Østergaard, C. R., Timmermans, B., & Kristinsson, K. (2011). Does a different view
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
145
create something new? The effect of employee diversity on innovation. Research
Policy, 40(3), 500-509.
Padgett, J. F., & Powell, W. W. (2012). The emergence of organizations and markets.
Princeton University Press.
Page, S. (2007). The Difference: How the Power of Diversity Creates Better Groups,
Firms, Schools, and Societies. Princeton University Press.
Pallot, M., Martínez-Carreras, M. A., & Prinz, W. (2010). Collaborative distance: A
framework for distance factors affecting the performance of distributed
collaboration. International Journal of e-Collaboration, 6(2), 1-32.
Pelled, L. H., Eisenhardt, K. M., & Xin, K. R. (1999). Exploring the black box: An
analysis of work group diversity, conflict and performance. Administrative
science quarterly, 44(1), 1-28.
Peters, L., & Karren, R. J. (2009). An examination of the roles of trust and functional
diversity on virtual team performance ratings. Group & Organization
Management.
Pfeffer, Jeffrey. New directions for organization theory: Problems and prospects. New
York: Oxford University Press, 1997.
Pfeffer, J., & Salancik, G. R. (1978). The external control of organizations. New York:
Harper & Row.
Phillips, K. W., & Loyd, D. L. (2006). When surface and deep-level diversity collide: The
effects on dissenting group members. Organizational behavior and human
decision processes, 99(2), 143-160.
Piller, F., & West, J. (2014). Firms, users, and innovation: an interactive model of coupled
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
146
open innovation. New Frontiers in Open Innovation, 29-49.
Pisano, G. P. (1991). The Governance of Innovation: Vertical Integration and
Collaborative Arrangements in the Biotechnology Industry. Research Policy 20:
237–249.
Polzer, J. T., Crisp, C. B., Jarvenpaa, S. L., & Kim, J. W. (2006). Extending the faultline
model to geographically dispersed teams: How colocated subgroups can impair
group functioning. Academy of Management Journal, 49(4), 679-692.
Porter, M. E., & Kramer, M. R. (2011). Creating shared value. Harvard business review,
89(1/2), 62-77.
Poteete, A. R., & Ostrom, E. (2004). Heterogeneity, group size and collective action: The
role of institutions in forest management. Development and change, 35(3), 435-
461.
Powell, W. W. (1990). Neither market or hierarchy: Network forms of organization.
Research on Organizational Behavior. 12. 295-336.
Powell, W. W., White, D. R., Koput, K. W., & Owen-Smith, J. (2005). Network dynamics
and field evolution: The growth of interorganizational collaboration in the life
sciences1. American journal of sociology, 110(4), 1132-1205.
Powell, W. W., Koput, K. W., & Smith-Doerr, L. (1996). Interorganizational collaboration
and the locus of innovation: Networks of learning in biotechnology.
Administrative science quarterly, 116-145.
Provan, K. G., & Kenis, P. (2008). Modes of network governance: Structure,
management, and effectiveness. Journal of public administration research and
theory. 18(2). 229-252.
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
147
Rajala, R., Westerlund, M., Vuori, M., & Hares, J. P. (2013). From Idea Crowdsourcing to
Managing User Knowledge. Technology Innovation Management Review, 3(12),
23-31.
Rainie, L., & Wellman, B. (2012). Networked: The new social operating system. MIT
Press.
Reed, R., Storrud-Barnes, S., & Jessup, L. (2012). How open innovation affects the
drivers of competitive advantage: Trading the benefits of IP creation and
ownership for free invention. Management Decision, 50(1), 58-73.
Reilly, K. M. A & Smith M. L. (2013). The Emergence of Open Development in a
Network Society. In Smith, M. & Reilly, K. (ed). Open Development: Networked
Innovations in International Development. Chapter 2 pp. 15-50
Rindova, V., Barry, D., and Ketchen Jr., D. J. (2009). Entrepreneuring as emancipation,
Academy of Management Review. 34(3): 477-491.
Robins, G., Pattison, P., Kalish, Y., & Lusher, D. (2007). An introduction to exponential
random graph (p*) models for social networks. Social networks, 29(2), 173-191.
Rokicki, M., Zerr, S., & Siersdorfer, S. (2015). Groupsourcing: Team Competition
Designs for Crowdsourcing. In Proceedings of the 24th International Conference
on World Wide Web (pp. 906-915). International World Wide Web Conferences
Steering Committee.
Root, A. (2012). Crowdsourcing act introduced in the Philippines. Archived from
http://www.crowdsourcing.org/editorial/crowdsourcing-act-introduced-in-the-
philippines-/20450
Rowley, T. J. (1997). Moving beyond dyadic ties: A network theory of stakeholder
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
148
influences. Academy of Management Review. 22(4). 887-910.
Sawhney, N. (2003). Cooperative innovation in the commons: rethinking distributed
collaboration and intellectual property for sustainable design innovation
(Doctoral dissertation, Massachusetts Institute of Technology).
Scekic, O., Truong, H. L., & Dustdar, S. (2015). Supporting Multilevel Incentive
Mechanisms in Crowdsourcing Systems: an Artifact-centric View. In
Crowdsourcing (pp. 91-111). Springer Berlin Heidelberg.
Schenk, E., Guittard, C., (2011). Towards a characterization of crowdsourcing practices.
Journal of Innovation Economics 7 (1), 93–107.
Schleich, R. J., & Prell, J. (2015). Social Open Innovation in Online Brand Communities
with Particular Regard to the Social Exchange Theory and Brand Management.
Lund University, Available at https://lup.lub.lu.se/student-
papers/search/publication/7439529
Schuurman, D. & Baccarne, B. & Merchant, P. (2013). Open Innovation: A Typology of
User Involvement in the Context of the Web2.0-Paradigm. Digiworld Economic
Journal, 89(1): 17-35
Schwenk, C. R., & Cosier, R. A. (1980). Effects of the expert, devil's advocate, and
dialectical inquiry methods on prediction performance. Organizational Behavior
and Human Performance, 26(3), 409-424.
Shaw, A. & Hill, B. M. (2014). Laboratories of oligarchy? How the Iron Law extends to
peer production: Laboratories of oligarchy. Journal of Communication. 64(2):
215–238. doi:10.1111/jcom.12082.
Shah, S. (2006). Motivation, governacne, and the viability of hybrid forms in open source
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
149
software development. Management Science. 52(7). 1000-1014. doi:
10.1287/mnsc.1060.0553
Shalley, C. E., & Perry-Smith, J. E. (2008). The emergence of team creative cognition:
the role of diverse outside ties, sociocognitive network centrality, and team
evolution. Strategic Entrepreneurship Journal, 2(1), 23-41.
Shaw, A., & Hill, B. M. (2014). Laboratories of oligarchy? How the iron law extends to
peer production. Journal of Communication, 64(2), 215-238.
Shen, C., & Monge, P. (2011). Who connects with whom? A social network analysis of an
online open source software community. First Monday, 16(6). Available at Who
connects with whom? A social network analysis of an onl…
Singh, H., Kryscynski, D., Li, X., & Gopal, R. (2015). Pipes, pools, and filters: How
collaboration networks affect innovative performance. Strategic Management
Journal. Published online EarlyView in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/smj.2419
Singh, P. J & Gurumurthy, A. (2013). Establishing Public-ness in the Network: New
Moorings for Development—A Critique of the Concepts of Openness and Open
Development. In Smith, M. & Reilly, K. (ed). Open Development: Networked
Innovations in International Development. Chapter 7, pp. 173-196.
Skopik, F., Schall, D., & Dustdar, S. (2010). Modeling and mining of dynamic trust in
complex service-oriented systems. Information Systems, 35(7), 735-757.
Smith, M. L., Elder, L., & Emdon, H. (2011). Open development: a new theory for
ICT4D. Information Technologies & International Development, 7(1), iii-iix.
Smith, M. L., Spence, R., & Rashid, A. T. (2011). Mobile phones and expanding human
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
150
capabilities. Information Technologies & International Development, 7(3), pp-77.
Smith, M., & Elder, L. (2010). Open ICT ecosystems transforming the developing world.
Information Technologies and international development, 6(1), 65-71.
Smith, M. & Reilly, K. (2013). Open Development: Networked Innovations in
International Development. MIT Press.
Snijders, T., Van de Bunt, G., & Steglich, C. (2009). Introduction to stochastic actor
based models for network dynamics. Social Networks, 32(1),44-60.
Snijders, T. A., Pattison, P. E., Robins, G. L., & Handcock, M. S. (2006). New
specifications for exponential random graph models. Sociological methodology,
36(1), 99-153.
Sorenson, O., Rivkin, J. W., & Fleming, L. (2006). Complexity, networks and knowledge
flow. Research policy, 35(7), 994-1017.
Souder, W. E. (1987). Managing new product innovation. Lexington, MA: Lexington
Books
Spaeth, S., Stuermer, M. & von Krogh, G. (2010). Enabling knowledge creation through
outsiders: towards a push model of open innovation. International Journal of
Technology Management. 52(3). 411-431. DOI - 10.1504/IJTM.2010.035983
Spithoven, A., Knockaert, M., & Clarysse, B. (2009). Open innovation: The role of
collective research centres in stimulating innovation in low tech SMEs. Flanders
DC. Vlerick Leuven Gent Management School. Retrieved from
http://www2.flandersdc.be/download/nl/20082642/file/flandersdc_openinnovation
_crcs_smes.pdf
Sullivan, E. (2010), “A Group Effort: More Companies Are Turning to the Wisdom of the
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
151
Crowd to Find Ways to Innovate,” Marketing News, February 28, 22-28.
Sundic, M. & Leitner, K. (2013). Crowdsourcing as an innovation strategy: A study on
innovation platforms in Austria and Switzerland. Digiworld Economic Journal,
89(1): 55-72.
Surowiecki, J. (2004). The wisdom of crowds: Why the many are smarter than the few
and how collective wisdom shapes business, economies, societies and nations.
Doubleday: New York
Tajfel, H. E. (1978). Differentiation between social groups: Studies in the social
psychology of intergroup relations. Academic Press.
Tapscott, D., & Williams, A. D. (2008). Wikinomics: How mass collaboration changes
everything. Penguin.
Taylor, A. (2005). An operations perspective on strategic alliance success factors: An
exploratory study of alliance managers in the software industry. International
Journal of Operations & Production Management. 25. 469-490. doi:
10.1108/01443570510593157
Teachman, J. D. (1980). Analysis of population diversity measures of qualitative
variation. Sociological Methods & Research, 8(3), 341-362.
Terborg, J. R., Castore, C., & DeNinno, J. A. 1976. A longitudinal field investigation of
the impact of group composition on group performance and cohesion. Journal of
Personality and Social Psychology, 34: 782–790
Terwiesch, C. & Xu, Y. (2008). Innovation Contests, Open Innovation, and Multiagent
Problem Solving. Management Science. 54(9): 1529-1543
Thomas, D., & Brown, J. S. (2009). Why virtual worlds can matter. International Journal
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
152
of Learning and Media. 1(1): 37-49
Thomas, J. B. & Trevino, L. K. (1993) Information processing in strategic building: A
multiple-case approach. Journal of Management Studies. 30(5). 779-814.
Tortoriello, M., McEvily, B., & Krackhardt, D. (2014). Being a catalyst of innovation: the
role of knowledge diversity and network closure. Organization Science, 26(2),
423-438.
Treem, J. W., & Leonardi, P. M. (2012). Social media use in organizations: Exploring the
affordances of visibility, editability, persistence and association. Communication
Yearbook, 36.
Trott, P., & Hartmann, D. (2009). Why open innovation is old wine in new bottles.
International Journal of Innovation Management. 13(4). 715-736.
Tsai, W. (2001). Knowledge transfer in intraorganizational networks: Effects of network
position and absorptive capacity on business unit innovation and performance.
Academy of management journal, 44(5), 996-1004.
Tsoukas, H. (2009). A dialogical approach to the creation of new knowledge in
organizations. Organization Science, 20(6), 941-957.
Turner, J. C. (1982). Towards a cognitive redefinition of the social group. Social identity
and intergroup relations, 15-40.
Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S. (1987).
Rediscovering the social group: A self-categorization theory. Basil Blackwell.
Ulhøi, J. P. (2004). Open source development: a hybrid in innovation and management
theory. Management Decision, 42(9), 1095-1114.
Valente, T. V. (2010). Social network influences on health behavior: Models, methods,
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
153
and applications. Oxford: Oxford University Press.
Van der Vegt, G. S., & Janssen, O. (2003). Joint impact of interdependence and group
diversity on innovation. Journal of management, 29(5), 729-751.
Van Knippenberg, D., Dawson, J. F., West, M. A., & Homan, A. C. (2011). Diversity
faultlines, shared objectives, and top management team performance. Human
Relations, 64(3), 307-336.
Vanhaverbeke, W., Cloodt. (2006). Open innovation in value network. In H. D.
Chesrough, W. Vanhaverbeke, J. West. (eds.), Open innovation: Researching a
new paradigm. Oxford: Oxford University Press
Villarroel, J. A. & Reis, F. (2010). Intra-corporate crowdsourcing (ICC): Leveraging upon
rank and site marginality for innovation. Proceedings of CrowdConf 2010. San
Francisco, CA.
von Hippel, E. (2005). Democratizing Innovation. Cambridge, MA, USA: MIT Press.
von Hippel, E., & von Krogh, G. (2003). Open source software and the "private-
collective" innovation model: issues for organization science. Organization
Science, 14(2), 209-233.
Von Krogh, G., Haefliger, S., Spaeth, S., & Wallin, M. W. (2012). Carrots and rainbows:
Motivation and social practice in open source software development. MIS
quarterly, 36(2), 649-676.
von Krogh, G., Spaeth, S., & Lakhani, K. R. (2003). Community, joining, and
specialization in open source software innovation: A case study. Research Policy,
32(7), 1217–1241.
Vuculescu, O., & Bergenholtz, C. (2014). How to Solve Problems with Crowds: A
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
154
Computer-Based Simulation Model. Creativity and Innovation Management,
23(2), 121-136.
Vuković, M. (2009, July). Crowdsourcing for enterprises. In Services-I, 2009 World
Conference on (pp. 686-692). IEEE.
Wallin, M.W., & von Krogh, G. (2010). Organizing for open innovation: focus on the
integration of knowledge. Organizational Dynamics 39 (2), 145–154.
Wang, R. & Cheliotis, G. (2016). Institutional and Entrepreneurial Engagement in
Commons-Based Peer Production. International Journal of Communication.
Wasko, M. M., & Faraj, S. (2005). Why Should I Share? Examining Social Capital and
Knowledge Contribution in Electronic Networks of Practice. MIS Quarterly,
29(1), 35-57.
Wasserman, S., & Faust, K. (1994). Social network analysis: Methods and applications
(Vol. 8). Cambridge university press.
Wasserman, S., & Galaskiewicz, J. (eds.) (1994). Advances in social network analysis:
Research in the social and behavioral sciences. Vol. 171. Sage Publications,
Weber, M. S., & Monge, P. (2011). The flow of digital news in a network of sources,
authorities, and hubs. Journal of Communication, 61(6), 1062-1081.
Weber, S. (2004). The success of open source. Cambridge, MA: Harvard University
Press.
Weick, K. E. (2007). The generative properties of richness. Academy of management
journal, 50(1), 14-19
Wellman, B., & Berkowitz, S. D. (eds.) (1988). Social structures: A network approach.
Vol. 2. CUP Archive,
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
155
West, J. (2003). How open is open enough? Melding proprietary and open source
platform strategies. Research Policy. 32(7). 1259-1285.
West, J., & Gallagher, S. (2006). Patterns of open innovation in open source software. In
H. Chesbrough, W. Vanhaverbeke, & J. West (Eds.), Open Innovation:
Researching a New Paradigm (pp.82-106). Oxford: Oxford University Press
West, J., & Lakhani, K. (2008). Getting clear about communities in open innovation.
Industry & Innovation, 15(2). 223- 231. DOI: 10.1080/13662710802033734!
West, J., Salter, A., Vanhaverbeke, W., & Chesbrough, H. (2014). Open innovation: The
next decade. Research Policy, 43(5), 805-811.
Williams, H. T., McMurray, J. R., Kurz, T., & Lambert, F. H. (2015). Network analysis
reveals open forums and echo chambers in social media discU.Ssions of climate
change. Global Environmental Change, 32, 126-138.
Williams, K. Y., & O’Reilly III, C. A. (1998). A REVIEW OF 40 YEARS OF
RESEARCH. Research in organizational behavior, 20, 77-140.
Williamson, O. E. (1985). The economic intstitutions of capitalism. Simon and Schuster.
Williamson, O. E. (1981). The Economics of Organization: The Transaction Cost
Approach. American Journal of Sociology, 87: 548-577.
Williamson, O. E. (1975). Markets and hierarchies. New York, 26-30.
World Bank. (2014). Water Hackathon: Lessons Learned. Retrieved from
http://www.scribd.com/WorldBankPublications/d/97458967-Water-Hackathon-
Lessons-Learned
Ye, H., Kankanhalli, A., Huber, M., Bretschneider, U., Blohm, I., Gaswami, S.,
Leimeister, J., Krcmar, H., (2012). Collaboration and the quality of user generated
COLLABORATIVE CROWDSOURCING AND TEAM SUCCESS
156
ideas in online innovation communities. In Annual Academy of Management
Meetings, August, Boston
Yoo, J., Lee, K., & Choi, M. (2013). Crowdsourcing for Device Manufacturers in the
Convergent Media Industry. Communications & Strategies, (89), 73-93.
Yuan, Y. C., Fulk, J., Monge, P. R., & Contractor, N. (2009). Expertise directory
development, shared task interdependence, and strength of communication
network ties as multilevel predictors of expertise exchange in transactive memory
work groups. Communication Research. 37(1): 20-47
Zhao, Y., & Zhu, Q. (2014). Evaluation on crowdsourcing research: Current status and
future direction. Information Systems Frontiers, 16(3), 417-434.
Abstract (if available)
Abstract
This dissertation examines crowdsourcing in the context of solving social development issues, through the lens of team collaboration. The goals are twofold. First, guided by a network perspective it aims to uncover the team dynamics of collaborative crowdsourcing. Specifically, it examines how marginality and preferential attachment influence collaboration patterns among members of a crowd. Second, it investigates how to leverage benefits of team collaboration through the lens of team diversity. It draws from the literature on team diversity and Group Faultline theory to investigate what team compositions can lead to better team performance. With data scraped from a global crowdsourcing community, Openideo, this dissertation employed large-scale behavioral data at the individual user level and team level to test the hypotheses. Exponential random graph modeling (ERGM) was used to analyze user level data in Study 1 and logistic regression was used to analyze team level data in Study 2. ❧ Results in Study 1 show that certain dimensions of marginality significantly influence how people choose team members to collaborate and solve social problems collectively, such as individual’s project evaluation skills, collaboration skills, and their network positions. Study 1 also shows that there were are influential members who are more likely to be chosen in teams due to their geolocation in main innovation centers, experience in creating teams, number of unique connections with others, and crowdsourcing winning experience in the community. ❧ Study 2 found that demographic diversity measures such as evaluation skills diversity and winning experience help improve team performance. It also found that team diversity in terms of geographic diversity does not generate disruptive influences on a team’s success. However, team diversity from members’ community tenure could activate team faultlines and lead to lower team performance. A practical implication of the two studies is that innovating organizations or a crowdsourcing community should apply certain strategies to integrate newcomers and sustain user contributions to achieve their social goals. A second implication is that network intervention strategies should be applied to facilitate fluid collaboration across sub-communities in Openideo. The third implication is that to build a winning team, we need to be aware that the effect of diversity is context specific. Discussion on how to leverage the benefits of diversity while avoiding the disruptive influences that may come along is provided.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
The evolution of scientific collaboration networks
PDF
Ecology and network evolution in online innovation contest crowdsourcing
PDF
Crowdsourcing for integrative and innovative knowledge: knowledge diversity, network position, and semantic patterns of collective reflection
PDF
The evolution of multidimensional and multilevel networks in online crowdsourcing
PDF
Communicating organizational knowledge in a sociomaterial network: the influences of communication load, legitimacy, and credibility on health care best-practice communication
PDF
Social value orientation, social influence and creativity in crowdsourced idea generation
PDF
The formation and influence of online health social networks on social support, self-tracking behavior and weight loss outcomes
PDF
We're all in this (game) together: transactive memory systems, social presence, and social information processing in video game teams
PDF
Collaborative care capacity: developing culture, power relationships and leadership support for team care in a hospital
PDF
Effects of economic incentives on creative project-based networks: communication, collaboration and change in the American film industry, 1998-2010
PDF
Improving information diversity on social media
PDF
Crowdsourced art: activating creative participation in online spaces
PDF
Staying ahead of the digital tsunami: strategy, innovation and change in public media organizations
PDF
Network influences in health initiatives: multimedia games for youth in Peru
PDF
A multitheoretical multilevel explication of crowd-enabled organizations: exploration/exploitation, social capital, signaling, and homophily as determinants of associative mechanisms in donation-...
PDF
Collaborative social networks in student affairs: an exploration of the outcomes and strategies associated with cross‐institutional collaboration
PDF
Understanding alliance building and communication across multiple technology platforms among California's immigrant-serving NGOs
PDF
Doing the hard work: including diversity and inclusion in your public relations tool kit to build stakeholder relationships
PDF
Essays on bioinformatics and social network analysis: statistical and computational methods for complex systems
PDF
Modeling social and cognitive aspects of user behavior in social media
Asset Metadata
Creator
Wang, Rong
(author)
Core Title
Building social Legoland through collaborative crowdsourcing: marginality, functional diversity, and team success
School
Annenberg School for Communication
Degree
Doctor of Philosophy
Degree Program
Communication
Publication Date
07/26/2016
Defense Date
05/02/2016
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
collaborative crowdsourcing,diversity,ERGM,exponential random graph modeling,marginality,network analysis,OAI-PMH Harvest,open collaboration,Openideo,preferential attachment,social challenges,team collaboration
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Bar, François (
committee chair
), Fulk, Janet (
committee member
), Valente, Thomas (
committee member
)
Creator Email
rongw@usc.edu,wrlaura@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c40-280635
Unique identifier
UC11280304
Identifier
etd-WangRong-4630.pdf (filename),usctheses-c40-280635 (legacy record id)
Legacy Identifier
etd-WangRong-4630.pdf
Dmrecord
280635
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Wang, Rong
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
collaborative crowdsourcing
diversity
ERGM
exponential random graph modeling
marginality
network analysis
open collaboration
Openideo
preferential attachment
social challenges
team collaboration