University of Southern California Dissertations and Theses

Social value orientation, social influence and creativity in crowdsourced idea generation

(USC Thesis Other)

Social value orientation, social influence and creativity in crowdsourced idea generation

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

SOCIAL VALUE ORIENTATION, SOCIAL INFLUENCE AND CREATIVITY IN
CROWDSOURCED IDEA GENERATION

by

Bei Yan

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 2018

Copyright 2018 Bei Yan
ii
Dedication
To my family – Zhiqiang Yan, Keling Zhu, Shunkun Zhang, Wenzhou Wang, and my advisor,
Andrea Hollingshead.

iii
Acknowledgements
I can’t remember how many times I’ve imagined writing the acknowledgement of my
dissertation. But when it finally came true, I was hesitating about what I should say. Indeed, too
much happened in my life in the past six years, and too many people I owe a debt of gratitude.
First and foremost, I would like to thank my family, my parents, grandma, husband and
extended family members. It is such a blessing to belong to a family full of love. Thank you all
for your unconditional support, although I know you are still not sure what I am keenly working
on. ☺
The person I can’t thank enough in my academic career is, of course, my advisor, Andrea
Hollingshead. Like I’ve said so many times in different occasions, Andrea, I can’t imagine a
better advisor than you are. Thank you for keeping a perfect balance between granting me
freedom and being responsive and supportive and for making me feel deeply respected and
acknowledged. And perhaps more importantly, thank you for being such a fun, sincere and
understanding person. Our occasional conference hangouts could not be more enjoyable. I have
to say, you are my lifelong role model both in my professional pursuit and academic career. You
did not just enlighten me to be a good researcher, but also to be a good person (with attitude!).
I’ve been part of the Annenberg Networks Network (ANN) since the first day of my PhD
to the last. To me, ANN is my academic family. I appreciate the generous support from Drs.
Janet Fulk and Peter Monge throughout the years. Dr. Janet Fulk’s Organization and Technology
class (I hope I still remember the name correctly, Janet) was a key inspiration for my PhD. It was
her class that taught me what is research and intellectual thinking. Additional thanks again goes
iv
to Dr Peter Monge, for keeping his office door open for me when I was lost in my academic
pursuit. I hope he did not mind that I joked about not being able to get his jokes at my hooding
ceremony.
There are many great minds that have illuminated me along my way. First are my
dissertation and qualifying committee members, Drs. Peter Carnevale, Lian Jian and Peter Kim. I
sincerely appreciate their valuable intellectual input to my research. It has been such an honor
working with them. Second are my course instructors who might have forgotten I was one of the
students seating in their lectures years ago – Drs. Sarah Banet-Weiser, Sarah Bonner, Manuel
Castells, Peer Fiss, Thomas Goodnight, Larry Gross and Sonia Livingstone. I would like to thank
them for instructing me how to think, critique, break and create.
I would also like to extend my deepest gratitude to my master’s and first year mentor, Dr.
Patricia Riley for her consistent encouragement and support during my PhD application and
beyond. My special thanks also goes to Drs. Jonathan Aronson and Ken Sereno for their support
in my PhD and job application process.
Finally, my friends and academic buddies, Nahoi Koo, Diana Lee, Ruqin Ren, Prawit
Thainiyom, Lena Uszkoreit, Wei Wang, Chi Zhang and Lin Zhang – my PhD life is incomplete
without our collective times of consolation, complaints and procrastination. Weekend hangouts
with Jingwei Li, Jiepeng Rong and their baby are the secret escapes in my spare time. Well, this
acknowledgement is going a little long-winded, but I am sure this is not an exhaustive list of
gratitude in such a significant period of my life. My sincere thanks are with all that have helped
me during my endeavor. Now it’s time to put an end to my graduate time. I am looking forward
to the future.

v
Table of Contents
Dedication ....................................................................................................................................... ii
Acknowledgements ........................................................................................................................ iii
List of Tables ................................................................................................................................. vi
List of Figures ............................................................................................................................... vii
Abstract ........................................................................................................................................ viii
CHAPTER 1: INTRODUCTION ................................................................................................... 1
CHAPTER 2: SVO IN CROWDSOURCED IDEA GENERATION ............................................ 5
Characterizing Crowdsourced Idea Generation .......................................................................... 5
SVO in Crowdsourced Idea Generation ..................................................................................... 7
Research Question .................................................................................................................... 11
CHAPTER 3: THEORETICAL FRAMEWORK ......................................................................... 14
The Motivated Information Processing Perspective on Creativity ........................................... 14
The Theory of SVO .................................................................................................................. 17
Research on Group Creativity ................................................................................................... 21
CHAPTER 4: PILOT STUDIES .................................................................................................. 27
Pilot Study 1: A Case Study of Crowdsourced Idea Generation Platforms .............................. 28
Pilot Study 2: SVO, Task Type, Task Interest and Creativity .................................................. 34
CHAPTER 5: SVO, TASK STRUCTURE AND CREATIVITY IN CROWDSOURCED IDEA
GENERATION ............................................................................................................................. 50
Theory and Hypotheses ............................................................................................................. 50
Study 1 ...................................................................................................................................... 54
Results ....................................................................................................................................... 61
Discussion ................................................................................................................................. 64
CHAPTER 6: SVO, TASK STRUCTURE, IDEA EXPOSURE AND CREATIVITY IN
CROWDSOURCED IDEA GENERATION ................................................................................ 66
Theory and Hypotheses ............................................................................................................. 66
Study 2 ...................................................................................................................................... 70
Results ....................................................................................................................................... 74
Discussion ................................................................................................................................. 77
CHAPTER 7: SVO, TASK STRUCTURE, OTHER’S PERFORMANCE AND CREATIVITY
IN CROWDSOURCED IDEA GENERATION .......................................................................... 79
Theory and Hypotheses ............................................................................................................. 79
Study 3 ...................................................................................................................................... 82
Results ....................................................................................................................................... 86
Discussion ................................................................................................................................. 90
CHAPTER 8: DISCUSSION AND CONCLUSION ................................................................... 93
References ................................................................................................................................... 103
Appendices .................................................................................................................................. 126
vi
List of Tables
Table 1: An Example of the Triple-Dominance Measure 36
Table 2: Study 1 Descriptive Statistics and Correlations 60
Table 3: Study 2 Descriptive Statistics and Correlations 73
Table 4: Study 3 Descriptive Statistics and Correlations 85
Table 5: Comparing Participants’ Creativity across Studies 95

vii
List of Figures
Figure 1. An example of the SVO Slider Measure 37
Figure 2. The interaction effect of SVO and task structure on idea productivity 62
Figure 3. The three-way interaction effect of SVO, task structure and idea exposure on idea
productivity 76
Figure 4. The interaction effect of SVO and others’ performance on idea originality 88
Figure 5. The interaction effect of SVO and others’ performance on idea usefulness 90

viii
Abstract
As more organizations crowdsource creative ideas online, understanding participant
creativity in crowdsourcing becomes critical. Although research has established that creativity is
an outcome of personality and situation, much remains unknown about how personal disposition
and contextual factors in crowdsourcing impact creativity. Integrating the motivated information
processing in groups (MIP-G) model with research on social value orientation and group
creativity, the current dissertation examines creativity in crowdsourced idea generation as a
consequence of social value orientation and social influence from task environment. Social value
orientation is a dispositional attribute indicating individual preference for outcome distributions
between oneself and others. Through a series of online experiments, the dissertation reveals that
prosocial participants’ creativity in crowdsourcing is less susceptible to influence from others or
the task environment. By contrast, the creativity of proselves is more sensitive to contextual
influences. Either a competitive task structure or the presence of highly productive others was
sufficient to arouse the competitive orientation of proselves and influence their creativity.
Nonetheless, whereas competition in crowdsourcing stimulates the productivity and originality
of proselves, it can be damaging for their idea usefulness. The dissertation is among the first
studies that disentangle the relations among social value orientation, social influence and
creativity in crowdsourcing.

Keywords: social value orientation, social influence, crowdsourcing, creativity, idea generation
1
CHAPTER 1: INTRODUCTION
Organizations of various types are turning online to utilize the distributed intelligence of
the crowd to generate new product ideas, designs, and solutions to social issues (Fuge, Tee,
Agogino, & Maton, 2014). Crowdsourced idea generation is a new production model that
outsources an idea generation task to a distributed network of individuals (the crowd) through an
open call (Afuah & Tucci, 2012; Brabham, 2008; Howe, 2008; Malone, Laubacher, &
Dellarocas, 2010). Participants volunteer to contribute their ideas based on their own value
orientation, interest, and expertise (Howe, 2008; Jeppesen & Lakhani, 2010) and they may or
may not get monetary rewards for their contribution.
Crowdsourced idea generation is not a new phenomenon in the digital age (Surowiecki,
2005), but it has become significantly easier to implement in recent years due to information
technologies. Thus far, it has been applied to a variety of problem domains, including research
on complex scientific problems (Majchrzak & Malhotra, 2013). Prominent examples of
successful application include InnoCentive, Quirky, and OpenIdeo. Apart from these new,
Internet-based organizations, traditional organizations in various industries are harnessing the
wisdom of crowds through crowdsourced idea generation as well. For example, Starbucks had an
online community, MyStarbucksIdea.com, for about 10 years to solicit consumer ideas regarding
its products and service
1
. The U.S. Department of Energy ran the SunShot Catalyst initiative for
at least two years to crowdsource software solutions from developers and data scientists to
improve solar energy efficiency in the United States
2
.

1
MyStarbucksIdea.com is no longer in operation as of 2017, but Starbucks keeps a feature on its website
to crowdsource ideas: https://ideas.starbucks.com/
2
For details, visit: https://www.energy.gov/eere/articles/sunshot-catalyst-innovators-take-software-
challenges-deploy-solar-technology-across
2
In crowdsourced idea generation, participant creativity determines the quality of ideas the
organizations can harness, and is therefore of critical importance to the sponsoring organizations.
Whereas early work on employee creativity has established that individual creativity in
traditional organizations is an outcome of personality and situation (Hackman & Oldham, 1976;
Oldham & Cummings, 1996; Shalley, Zhou, & Oldham, 2004), little research has examined how
personal disposition and task contexts impact participant creativity in crowdsourced idea
generation.
Integrating the motivated information processing in groups (MIP-G) model, theory of
social value orientation, or SVO (McClintock, Messick, Kuhlman, & Campos, 1973), and
research on group creativity (Diehl & Stroebe, 1987), the current dissertation explicates how
social value orientation affects individual creativity in crowdsourced idea generation. Through a
series of online experiments, it also examines how task structure and communication (i.e.
exposure to others’ ideas and performance) moderate the relation between SVO and participant
creativity.
Social value orientation, is a dispositional attribute indicating individual preference for
outcome distributions between oneself and others (McClintock, 1972; McClintock & Liebrand,
1988; McClintock et al., 1973). Rather than assuming individuals to be uniformly self-interested,
the theory of SVO argues that individuals vary in terms of their dispositional preferences to
distribute values between themselves and others in interdependent situations. Whereas proselves
maximize their own gains as traditional economic theory assumes, prosocials care about others’
gains and try to benefit others, sometimes even at their own expense.
The open, public nature of crowdsourced idea generation, along with high efforts and low
compensation, has led scholars to argue that concern for others’ benefits is a critical factor that
3
drives many participants (Benkler, 2006; Dholakia, Bagozzi, & Pearo, 2004; Wasko, Faraj, &
Teigland, 2004). Despite research evidence demonstrating the heterogeneity in participants’ care
for others, prior studies of crowdsourcing have mainly focused on the effects of crowd
composition, incentives, and other design features on participant behavior (Boudreau & Lakhani,
2015; Terwiesch & Xu, 2008; Yu & Nickerson, 2011). Little, if any, research has investigated
how variation in social motives impacts participant creativity in crowdsourced idea generation.
The current dissertation project is among the first studies to examine how variation in
participant SVO interacts with contextual factors, specifically task structure and communication,
to impact creativity in crowdsourcing. Since previous SVO research has been mainly conducted
in offline labs or in exploration of cooperation in behavioral games (Bechtoldt, de Dreu, Nijstad,
& Choi, 2010; Liebrand, Wilke, Vogel, & Wolters, 1986; van Lange, Klapwijk, & Munster,
2011), this project also develops the theory by expanding and exploring its mechanism in online
creative tasks.
The dissertation also produces design implications for crowdsourcing platforms and
organizations that plan to crowdsource their internal tasks. Specifically, understanding how SVO
influences creative collaboration in crowdsourcing will help project and platform designers to
create better task structures, communication mechanisms, and reward systems to promote
creativity in participating crowds. The investigation of group communication on creativity can
also offer practical implications about how to cultivate a social environment that facilitates
collective creativity in crowdsourcing.
The content of the dissertation is organized as follows. In Chapter 2, I will start with
discussions about the features of crowdsourced idea generation. Chapter 2 argues that the open,
public feature of crowdsourcing reveals the heterogeneity in participants’ SVOs and that
4
participant creativity in a given context cannot be understood without examining the impact of
the disposition. The research question of the dissertation is then proposed at the end of this
chapter. Chapter 3 reviews the theoretical foundation of the dissertation project, namely the MIP-
G model, the theory of SVO, and group creativity. The integrated insights from the three streams
of research suggest that the relation between participant SVO and creativity cannot be explicated
without examining the moderating effects of two contextual factors: task structure and
communication (i.e. exposure to others’ ideas and information about others’ performance).
Chapter 4 presents a set of pilot studies, designed to explore the empirical context of
crowdsourced idea generation and to inform the design of the main experiments. Chapters 5-7
formulate hypotheses regarding the effects of SVO, task structure and communication on
participant creativity, and also report the results of the three studies conducted to test the
hypotheses. Study 1, to be discussed in Chapter 5, examined the influence of SVO and task
structure on participant creativity. Chapter 6 covers the theories and findings of Study 2, which
investigated how SVO, task structure, and idea exposure interact to affect creativity. In Chapter
7, Study 3 is reported. This study explored the effects of SVO, task structure and information
about others’ performance on participant creativity in crowdsourcing. To conclude the
dissertation, Chapter 8 provides a general discussion of the findings as well as the theoretical and
practical implications. Limitations and directions for future research are also identified.
5
CHAPTER 2: SVO IN CROWDSOURCED IDEA GENERATION
Characterizing Crowdsourced Idea Generation
Crowdsourced idea generation is a production model that outsources an idea generation
task to a distributed network of individuals (the crowd) through an open call (Afuah & Tucci,
2012; Brabham, 2008; Howe, 2008; Malone et al., 2010). An example can be:
MyStarbucksIdeas.com, the online crowdsourcing idea generation community of Starbucks,
which publicly solicited opinions and new ideas about the company’s products and services.
Rather than designating a specialized crew to come up with solutions (e.g. product design) to an
issue or problem (e.g. new product development), crowdsourcing publicly discloses the details of
the problem to invite anyone who may be interested in the topic.
The topics of crowdsourced idea generation projects vary greatly, depending on the need
of the project sponsors. For example, whereas MyStarbucksIdeas.com crowdsourced ideas about
drinks, OpenIdeo.com publicly calls for solutions to social issues such as health care and
poverty. Regardless of topics or issues, however, idea contributors are often not the direct
beneficiaries of their own ideas. Nor are the benefits of a proposed idea or solution exclusive to
idea contributors. Instead, the proposed ideas are commonly harnessed by the project sponsors.
For example, consumer ideas proposed on MyStarbucksIdeas.com were utilized by Starbucks to
develop better products and advance its profits. Although, arguably, idea contributors could
enjoy products that cater to their preference once their product ideas were implemented, this
advantage was not exclusive to themselves, but also benefited all other consumers who preferred
similar products yet did not make contributions.
In this sense, crowdsourced idea generation projects at least partially mimic the structure
of a public goods dilemma: since the outcome of contribution is non-excludable from individuals
6
regardless of their contribution, rational, self-interested individuals should not contribute to the
collective project as they can simply free-ride (Marwell & Oliver, 1993; Olson, 1965). As a
result, no one should contribute to public goods, and public goods cannot be sustained (Hardin,
1968).
A common solution proposed by research on social dilemmas to cultivate contributions to
public goods is to set up a central governance mechanism and provide private incentives for
contribution (Olson, 1965; Ostrom, 1990). Yet crowdsourced idea generation projects typically
do not have a top-down organizational structure. Participants volunteer to contribute their ideas
based on their own interest and expertise without managerial enforcement from an organizational
hierarchy (Howe, 2008; Jeppesen & Lakhani, 2010).
Crowdsourced idea generation projects sometimes do provide certain incentives for
participation (Boudreau & Lakhani, 2015; Malone et al., 2010). However, the incentives may not
guarantee substantial private gains and can be hard to obtain, given that crowdsourcing projects
usually attract hundreds or even thousands of participants to compete for the rewards. Among the
crowdsourcing projects (e.g. Innocentive, some projects on Kaggle, Amazon Mechanical Turk)
that do reward contributors financially, some (e.g. Amazon Mechanical Turk) provide relatively
low compensation, with workers being paid cents for short tasks (Litman, Robinson, &
Rosenzweig, 2015). In other cases, however, the pay can be substantial. For example, some
projects on Kaggle, an online crowdsourcing community for machine learning algorithms,
reward winning teams with a prize worth a few thousand dollars.
Nonetheless, the prize does not come easily. The Kaggle community has more than
500,000 registered users
3
and each competition normally contains hundreds or thousands of

3
The data was updated in May, 2016; for more detailed information, visit:
https://en.wikipedia.org/wiki/Kaggle
7
participating teams. Since the problems posted on Kaggle are intellectually challenging (e.g.
building machine learning algorithms to segment images captured by vehicles), winning
solutions are likely to come from a team consisting of multiple members. Consequently, most
participants’ efforts will be in vain, and the prize for winning teams also needs to be further
divided among team members.
Despite the efforts that contributors need to put forth, many crowdsourcing projects (e.g.
MyStarbucksIdea.com) do not provide monetary rewards, but only social appreciation or
recognition. For example, although most Kaggle projects require considerable time and
intellectual input, many competitions on the platform do not provide monetary rewards other
than “knowledge” or “kudos.”
Overall, crowdsourced idea generation projects are open, public systems that share some
features of a public goods dilemma. They typically lack managerial controls and depend on
voluntary contributions. Although some projects require considerable input from participants,
crowdsourced idea generation projects may or may not provide monetary incentives for
contribution.
SVO in Crowdsourced Idea Generation
The open, public nature of crowdsourced idea generation, coupled with low
compensation, high effort, and lack of managerial enforcement, therefore challenges the premise
of traditional collective action theories. Traditional theories assume that individuals are
homogeneously self-interested agents who try to maximize their own private gains (Olson,
1965). Given the public goods feature of many crowdsourcing projects and their variation in
contribution incentives, it is hard to assume that contributions to crowdsourced idea generation
are solely driven by private value maximization goals (Benkler, 2002, 2006). It is likely that
8
while some individuals participate in crowdsourcing for private gains, others contribute their
time and effort for the sake of helping their community.
Indeed, scholars have argued that social and other-oriented motivations are critical factors
that propel crowd-led innovation (Adler, 2001; Benkler, 2002, 2006; Powell, 1990). Apart from
private gains (Lerner & Triole, 2000), empirical evidence suggests that participants in online
crowd communities are also driven by altruism (Quinn & Bederson, 2011), reciprocity (Faraj &
Johnson, 2011; Lakhani & von Hippel, 2003), community reputation (Lampel & Bhalla, 2007),
social capital and relationships (Butler, Sproull, Kiesler, & Kraut, 2007; Dholakia et al., 2004;
Wasko & Faraj, 2005), and group commitment (Bateman, Gray, & Butler, 2010).
The variation in participants’ social motives in crowdsourcing is best summarized by
theory and research on social value orientation. Social value orientation, or SVO in short, is a
dispositional attribute indicating individual preference for outcome distributions between oneself
and others (McClintock, 1972; McClintock & Liebrand, 1988; McClintock et al., 1973). The
concept is also known as social motives or social motivation (Bechtoldt et al., 2010;
MacCrimmon & Messick, 1976). In social systems, the choices of one person often impact the
outcomes of others due to their interdependence (MacCrimmon & Messick, 1976). In early
experimental research utilizing behavioral games, researchers found that some individuals tended
to cooperate more than others in various types of social dilemmas (Messick & McClintock,
1968). Consequently, rather than assuming individuals are homogeneously self-interested, the
theory of SVO argues that individuals vary in terms of their dispositional preferences to
distribute values between themselves and others in interdependent situations. The theory goes
beyond the egoistic, self-interested assumption of human behavior. It argues that concern for
others, just like concern for self-interest, is also part of rationality (de Dreu, 2006).
9
In the early conceptualizations, researchers proposed several different types of SVO: 1)
individualistic (i.e. self-interest oriented) participants who tend to maximize their own gain in
interdependent situations, regardless of others’ gains; 2) competitive individuals who try to
maximize their relative gain against others; 3) cooperative individuals who are concerned about
the collective and strive to maximize the joint gain; 4) altruists who work to maximize others’
gains; and 5) aggressive individuals who try to minimize others’ gains, even at the expense of
their own (MacCrimmon & Messick, 1976, 1976; McClintock, 1972; McClintock et al., 1973).
Among them, altruistic and aggressive individuals, particularly aggressive individuals, are rare in
empirical settings (McClintock et al., 1973). Some scholars also have proposed equality (i.e.
preference for equal gains between self and the other) and maximin (i.e. maximizing the gain for
the individual who received the lowest benefits) as additional types of SVO (Eek & Gärling,
2006, 2008; Iedema & Poppe, 1994; Knight & Dubro, 1984; van Lange, 2000).
In general, these classifications of SVOs fall into two basic categories: prosocial
orientation, which includes cooperation and altruism; and proself orientation, which generally
includes individualism and competition. This general distinction has been widely adopted in
prior research (Balliet, Parks, & Joireman, 2009; Bogaert, Boone, & Declerck, 2008; de Dreu,
Nijstad, Bechtoldt, & Baas, 2011; van Lange & Liebrand, 1991) and is the classification applied
in the current project. Individuals are considered to manifest prosocial orientation if they care
about others’ in the collective and try to benefit others’ outcomes when they are making choices
in social situations (MacCrimmon & Messick, 1976). Otherwise, they are considered as proself-
oriented if they only consider their own outcome as with a traditional economic man assumption,
or try to increase their relative outcome compared to others.
10
Prior research has shown that SVO consistently predicts cooperative behaviors of
participants in different behavioral games: prosocial participants (i.e. cooperators, altruists)
cooperate more than individualists than competitors in different game situations (Balliet et al.,
2009; Bogaert et al., 2008; Brucks & van Lange, 2007; Hilbig & Zettler, 2009; Kuhlman &
Marshello, 1975a; Kuhlman & Wimberley, 1976; Liebrand, 1984; Liebrand & van Run, 1985;
Roch & Samuelson, 1997; Simpson, 2004; van Dijk & de Cremer, 2006). A recent simulation
study shows that crowd collaboration can be successful as long as prosocial members exist
(Levine & Prietula, 2014).
SVO is also relatively stable across time, contexts, and culture (Bogaert et al., 2008;
Eisenberg et al., 2002; Liebrand & van Run, 1985; McClintock & Allison, 1989; Yamagishi,
1995), although it is not completely free from temporal development or contextual influence
(Liebrand, Wilke, et al., 1986; van Lange, Otten, de Bruin, & Joireman, 1997). In a longitudinal
study of young adults, Eisenberg and colleagues (2002) found that prosocial disposition in
individuals’ early adulthood was significantly related to reported prosocial behaviors in their
childhood and the dispositional orientations remained stable across five years of research.
Participants SVOs have also been shown to match their friends’ descriptions of them (Bem &
Lord, 1979).
Taken together, prior research has established that SVO is a dispositional attribute that
differs among individuals across social contexts: whereas some individuals (proselves) mainly
care about their own gains, others (prosocials) act with concern for the collective and for others.
This variation in SVO is clearly manifested in participants of crowdsourced idea generation,
since these online projects commonly generate non-excludable benefits for others, but may or
may not provide private incentives for contributors. Based on research on SVO, prosocials are
11
perhaps more likely to participate in crowdsourcing projects aimed at helping the community or
others, even without private incentives. Proself individuals, however, are likely to participate in
crowdsourcing projects if the issues concern their own interests or if the project offers substantial
private incentives (de Kwaadsteniet, van Dijk, Wit, & de Cremer, 2006; Galletta, Marks, Polak,
& McCoy, 2003; Karau & Williams, 1993).
Research Question
Although the research and practice of crowdsourcing both provide considerable support
for the variation in participant SVO, little, if any, previous research has investigated how
participants’ SVOs impact their behavioral outcomes in crowdsourcing. Since crowdsourced idea
generation projects harness crowd ideas to help solve their issues of concern, the creativity of
participants’ contributions is of primary interest to project organizers. Following prior research, I
define creativity in crowdsourced idea generation as individuals’ ability to generate ideas and
solutions that are both novel and appropriate (Amabile, 1996; de Dreu et al., 2011; George,
2007).
Creativity in crowdsourced idea generation is a complex social process which requires
both convergence and divergence among participating individuals (Goncalo & Staw, 2006;
Milliken, Bartel, & Kurtzberg, 2003; Nijstad & de Dreu, 2012): on one hand, it requires
individuals to be motivated toward the collective goal and to be attentive to others’ ideas and
actively contribute their own; on the other hand, it needs the participants to be self-determining
and propose original and unique ideas. Since SVO dictates the type of information participants
process, it is likely to impact participants’ susceptibility to contextual influences in social
interaction and affect their creativity in crowdsourcing (Kuhlman & Marshello, 1975a; Liebrand,
Wilke, et al., 1986).
12
Thus far, only a small amount of research has explored the creative process of the crowd.
Extant research on creativity in crowdsourcing mainly tackled the following factors:
Diversity. Research has been consistent in showing that having a more diverse pool of
participants in terms of experience and expertise in crowdsourcing leads to better performance of
the crowd in creative problem solving (Armisen & Majchrzak, 2014; Boudreau, 2012; Jeppesen
& Lakhani, 2010; Malhotra & Majchrzak, 2014; Terwiesch & Xu, 2008). Yet although the
increased number of participants may lead to higher diversity, having too many participants
working on the same issue may negatively affect participant innovation (Boudreau, 2012).
Incentives. A few researchers have investigated how incentive structures employed by
crowdsourcing project sponsors may influence crowd creativity. Terwiesch and Xu (2008)
mathematically proved that performance-based rewards are more beneficial than fixed rewards in
crowdsourcing contests. Moreover, although increasing monetary rewards does not significantly
increase average crowd creativity, it does improve the likelihood of obtaining highly creative
solution (Wu, Corney, & Grant, 2014).
Task Structure. Boudreau and colleagues (Boudreau, Lakhani, & Menietti, 2014)
showed that in crowdsourcing contests, most participants performed worse as the competition
intensified, but the most capable contestants performed better. The relation between competition
and participant performance may also depend on task uncertainty: whereas competition reduces
contestants’ effort in more certain tasks, it increases their effort on highly uncertain problems
(Boudreau, Lacetera, & Lakhani, 2011).
Communication and interaction. Research on crowdsourced idea generation has also
demonstrated that the crowd are capable of integrating and combining others’ ideas (Yu &
Nickerson, 2011, 2013). When they build on each other’s ideas rather than arguing with each
13
other, they generate more creative ideas (Armisen & Majchrzak, 2014; Yu & Nickerson, 2011,
2013). In addition, while disclosing solutions and progress in crowdsourcing contests helps
participants to more effectively improve their solutions, it may also result in convergence among
final solutions within the crowd (Boudreau & Lakhani, 2015).
Despite the variation in participant SVO in crowdsourcing, the existing research on
crowd creativity has not explored how different SVOs may impact the creative performance of
participants in a given context. Implicitly, crowds have been studied as if they consisted of
participants who are homogeneous in their social motivations.
Moreover, although prior research has investigated the effect of crowd diversity,
incentives, and task competition on creativity, much remains unknown about the ideation
processes through which the diverse knowledge of crowds becomes integrated, or how incentives
and competition elicit better or worse performance. Whereas investigations on group
communication have shed light on the cognitive stimulation and social influence processes in
crowdsourcing, more needs to be investigated regarding the extent of the influence. For example,
which crowdsourcing participants are more likely to be impacted by others’ ideas and behavior
in interaction? Under what conditions are participants more likely to build on each other’s ideas?
These creative processes cannot be understood without considering the fundamental value
orientation of the participants in social situations. Therefore, the overarching research question of
the current dissertation project is:
RQ1: How does SVO impact individual creativity in crowdsourced idea generation?

14
CHAPTER 3: THEORETICAL FRAMEWORK
In this chapter, I discuss the theoretical framework of the dissertation project. I start by
reviewing the three streams of research that serve as the theoretical foundation of the
dissertation, namely, the Motivated Information Processing in Groups model (MIP-G), theory of
SVO and research on group creativity (particularly in idea generation). Each stream of research
contributes a particular set of insights to understand the relation between SVO and creativity in
crowdsourced idea generation, but is inadequate on its own to fully explicate the creative process
in crowdsourcing. Whereas the motivated information processing perspective bridges the
research on SVO and group creativity by theorizing the effect of social motivation on creativity,
it does not incorporate insights from the other two research paradigms regarding the moderating
effects of task structure and others’ behavior in social interaction. Therefore, these approaches
need to be integrated and need to complement each other, in order to explicate the creative
process in crowdsourced idea generation.
The Motivated Information Processing Perspective on Creativity
The Motivated Information Processing in Groups model (MIP-G) (de Dreu, Nijstad, &
van Knippenberg, 2008; Nijstad & de Dreu, 2012) posits that groups are information processing
systems. The systems function through two dynamics processes: 1) individual group members
search and process information, and 2) member information processing becomes integrated
through communication. Since individuals are motivated information processors, group-level
information processing is driven by the motivation of group members. One of the key
motivations that drives individuals, according to MIP-G, is the SVO, or social value orientation,
of group members.
15
According to MIP-G, individual SVO determines the type of information processed by
group members when generating ideas in groups (Nijstad & de Dreu, 2012). Prosocial
individuals are more likely to process information in a way that benefits fairness, integration and
overall quality of collective outcomes. Prosocial value orientation can thus benefit creativity in
interdependent settings, since the social environment needed for creative behavior – high levels
of trust, cohesion and psychological safety, is more likely to appear in cooperation (Carnevale &
Probst, 1998; Nijstad & de Dreu, 2012).
By contrast, individuals with a proself-orientation tend to process information to boost
their individual outcomes or try to outperform others in groups. Proself individuals may be less
concerned about producing ideas for the collective good. However, compared to prosocials, they
may be more likely to pursue different personal goals and generate original and unique ideas
(Goncalo & Duguid, 2012; Goncalo & Kim, 2010; Goncalo & Staw, 2006; Janssen & Huang,
2008).
The MIP-G model thus argues that the specific effects of SVO on group creativity are
contingent upon another motivation, epistemic motivation (de Dreu et al., 2011). Epistemic
motivation is often shaped by contextual factors such as task environment (e.g. high time
pressure). Prosocials are more creative in situations that stimulate in-depth information
processing, or high epistemic motivation (Beersma & de Dreu, 2005). However, if the tasks they
perform lead to heuristic information processing (low epistemic motivation), prosocial group
members are likely to commit to premature concession and convergence, which in turn results in
low levels of creativity (de Dreu et al., 2011). On the other hand, proself members are likely to
be more creative in collective settings if their personal goals are aligned with collective goals
(Nijstad & de Dreu, 2012).
16
However, one limitation in research applying the MIP-G perspective is that SVO was
often primed with a contextual situation in the experiments – i.e. prosocial orientation was
manipulated by cooperative tasks in which collective performance was rewarded, and proself
orientation was manipulated as competitive tasks in which individual performance was rewarded
(Beersma & de Dreu, 2005). Since SVO has been established as a dispositional motivation that is
stable across contexts (Bogaert et al., 2008; Eisenberg et al., 2002; Liebrand & van Run, 1985;
McClintock & Allison, 1989; Yamagishi, 1995), some of the results can be misleading, because
what they claimed to be the consequences of SVO could be the effects of task structure. The
potential interaction effect between SVO and task structure also remains under-researched.
Furthermore, manipulating SVO via task structure assumes that group members are
homogeneous in terms of SVO on a given task, which is incorrect according to research on SVO
(Carnevale & Probst, 1998; Liebrand, Wilke, et al., 1986).
In a few other cases, SVO was not directly examined, but studied as potentially related
factors such as agreeableness, narcissism and self-construal (in fact, self-construal was primed
rather than measured as well) (Bechtoldt, Choi, & Nijstad, 2012; Bechtoldt et al., 2010; Goncalo,
Flynn, & Kim, 2010; Goncalo & Staw, 2006). These measures may not capture the effect of
SVO since there is limited direct evidence supporting their association with SVO.
More importantly, the main contribution of the MIP-G model to research on group
creativity is that it proposes a perspective to understand group-level processes based on
motivations of group members. However, the model is almost exclusively focused on group
member motivation, and does not integrate insights from previous research regarding how the
effects of SVO may be moderated by other contextual factors (e.g. task structure, social
interaction) in group ideation. Consequently, the perspective is less informative about the
17
creative process in crowdsourced idea generation, where task structure and communication
mechanisms vary. It therefore needs to be combined with research on SVO and group creativity
to elucidate participant creativity in crowdsourcing.
The Theory of SVO
Previous research on SVO has demonstrated that individuals with different SVOs may
react differently to contextual factors, such as task structure and others’ behavior, in social
interaction. In one of the early conceptual framework of SVO, McClintock (1972) proposed that
the consequence of SVO can be influenced by task structure (e.g. cooperative vs. competitive
task structure). The proposition was later echoed by other scholars (Lozano, 2018; van Lange,
2000), and proved by a series of empirical studies (Bekkers, 2004; Brucks & van Lange, 2007;
Cornelissen, Dewitte, & Warlop, 2011; Liebrand, Wilke, et al., 1986; McClintock & Liebrand,
1988; McClintock et al., 1973; Utz, 2004). In a study exploring cooperation across three different
behavioral games, Liebrand and colleagues (Liebrand, Wilke, et al., 1986) found that although
SVO consistently predicted cooperative behaviors across three different games, the level of
cooperation was influenced by the task structure of the games.
Task structure may also interact with SVO to influence participants’ task effort.
Researchers have shown that after being rewarded for high efforts, cooperators demonstrated
higher persistency on cooperative tasks, whereas individualists were more persistent in
competitive tasks. Competitors, however, were equally persistent across tasks (Eisenberger,
Kuhlman, & Cotterell, 1992).
In interdependent situations such as crowdsourced idea generation, participants often
interact with each other and are exposed to others’ ideas and behavior. The communication and
interaction processes may also impact participants’ behavior, and interact with SVO to influence
18
participants’ performance and collective outcomes. Among the few studies that directly
investigated the effects of group communication on cooperation in social dilemmas, Liebrand
(1984) found that when communication was allowed within groups in multi-stage games, there
was significant cooperation (i.e. less is taken for self) for both prosocial and proself participants.
Many other studies have examined how others’ behavior (i.e. direct exposure or indirect
information) in group interaction impacts participants’ cooperation in games. McClintock and
colleagues (1973) showed that cooperative behavior within groups varied less compared to the
amount of variation between groups, indicating behavioral assimilation within groups when
group members were exposed to each other’s behavior.
In general, research has shown that cooperation leads to more cooperation and
competitive behavior leads to the reverse, particularly for prosocial individuals (Kelley &
Stahelski, 1970; Kuhlman & Marshello, 1975a; Kuhlman & Wimberley, 1976; McClintock &
Liebrand, 1988). One of the earliest studies tackling this question was conducted by Kelley and
Stahelski (1970). They found that prosocial participants (i.e. cooperators), compared to proselves
(i.e. competitors), were more likely to assimilate to competitive behaviors in the prisoner’s
dilemma. They also demonstrated that while the cooperators were aware of the heterogeneity in
people’s social orientation, competitors believed that others were uniformly competitive. This
finding was later labelled as the “triangle hypothesis” by other scholars (Liebrand, Wilke, et al.,
1986).
However, contrary to the triangle hypothesis, Kuhlman and Wimberley (1976) found that
people of all SVO types assumed that most others would be similar to themselves. They argued
that Kelley and Stahelski (1970)'s hypothesis was more of a result of the game structure. Their
argument was corroborated by other studies showing that when cooperators participated in a
19
series of games in which defection would lead to further loss for the collective, participants did
not adapt to the selfish behavior of competitors (Liebrand, 1984). The study also shows that both
prosocials and proselves have relatively accurate expectations about others, with competitors
expecting others to take less than themselves, and altruists expecting others to take more than
themselves.
Proself participants can also be responsive to the behavior of others under certain
conditions. Liebrand and colleagues (Liebrand, Wilke, et al., 1986) showed that in larger groups,
prosocials were not significantly affected by feedback about the majority behavior in the group,
yet proselves were more likely to defect if they were informed that the majority of the group had
defected than when they were informed that the majority chose to cooperate. Knowing that
others had defected seemed to have strengthened proselves’ inclination to defect. Furthermore,
several studies have revealed that prosocials are generally more likely to reciprocate others’
cooperation, with no discrimination concerning others’ traits (de Cremer & van Lange, 2001;
Parks & Rumble, 2001; Utz, 2004; van Lange, 1999; van Lange & Semin-Goossens, 1998). By
contrast, proselves reciprocate more when they perceive others as honest or after being retaliated
against (Parks & Rumble, 2001; van Lange & Semin-Goossens, 1998).
Overall, evidence suggests that both prosocials and proselves are responsive to the
behavior of others in social systems. Yet prosocials are concerned about the equality and
collective benefits of their social groups, and thus are more likely to reciprocate the cooperative
actions of others (Bogaert et al., 2008; Kanagaretnam, Mestelman, Nainar, & Shehata, 2009). On
the contrary, proselves care about maximizing their absolute or relative gain and tend to take
advantage of others to benefit themselves when possible.
20
In summary, considerable evidence in previous research has demonstrated that people
with distinct SVOs react differently to task structure and the behavior of others while interacting.
It is therefore logical to predict that in crowdsourced idea generation, while individuals’ SVO
impacts their level of cooperation and creativity, its effect should be understood in relation to the
task structure and communication process.
One limitation in prior research on SVO is that the findings concerning the impact of
SVOs on cooperative behavior can be somewhat tautological. Since SVO is defined as people’s
preference to consider the benefits of others, it may not be particularly surprising to know that a
person who is concerned about others is more likely to take less in public goods dilemmas or to
reciprocate the cooperation of others. More research is needed to study how SVO may impact
individual cognition and behavior in different social circumstances and while performing
different tasks.
Research on SVO and the motivated information processing perspective of group
creativity can therefore complement each other to explicate the creative process in crowdsourced
idea generation: while the MIP-G model advances research of SVO by bringing the effects of
social orientation and creativity, research on SVO provides further insights into how SVO and
contextual factors in crowdsourcing (i.e. task structure and group communication) interact to
impact an individual’s creative process. Other research on group creativity, while implicitly
characterizing groups as consisting of members with homogeneous SVO, focuses on how social
interaction and communication in groups impact collective creativity. It therefore adds to the two
streams of research by providing rich knowledge regarding the impact of group interaction on
creativity in idea generation.
21
Research on Group Creativity
One reason people believe in collective creativity is the potential benefit of cognitive
stimulation and its subsequent synergy when individuals with diverse information and
perspectives communicate with each other (Baruah & Paulus, 2009; Paulus, 2000; Sawyer,
2007). Contrary to the general belief, however, previous research on group creativity has
generated mixed findings regarding whether communication may benefit or impede group
creativity.
Early research in group creativity consistently demonstrated a significant negative effect
of group interaction on creativity, both in labs and in organizations (Diehl & Stroebe, 1987;
Larey & Paulus, 1999; Paulus, Larey, & Ortega, 1995; Van de Ven & Delbecq, 1974). In more
recent studies of electronic brainstorming groups, however, interactive groups performed better
than nominal groups when the group size became large enough (Dennis & Valacich, 1993;
Gallupe et al., 1992; Gallupe, Bastianutti, & Cooper, 1991; Valacich, Dennis, & Connolly, 1994;
Ziegler, Diehl, & Zijlstra, 2000). A meta-analysis of research on team innovation has also
established that both internal and external communication of teams are positively related to team
innovation (Hülsheger, Anderson, & Salgado, 2009).
When examining the effects of communication on creativity, prior research generally has
focused on two types of information that participants often receive access to in interaction: ideas
generated by other participants, and information regarding the performance of others (i.e. idea
productivity).
Exposure to others’ ideas. Exposure to the ideas of others while brainstorming may
trigger a number of cognitive processes. The process that negatively influences creativity is
cognitive fixation. Cognitive fixation refers to “something that blocks or impedes the successful
22
completion of various types of cognitive operations, such as those involved in remembering,
solving problems, and generating creative ideas” (Smith, 2003, p. 16). When fixation happens,
people’s cognition tends to converge on other’s opinions, and people are less likely to search for
different solutions to a problem. Both simulation and empirical studies have suggested that
members of interactive groups show a strong tendency to converge on each other’s ideas
(Brown, Tumeo, Larey, & Paulus, 1998; Coskun & Yilmaz, 2009; Larey & Paulus, 1999).
Furthermore, this tendency is not only restricted to groups in which members interact face-to-
face; it also occurs in virtual groups in which group members only have mediated interactions
with each other (Yu & Nickerson, 2011; Ziegler et al., 2000).
On the contrary, exposure to the ideas of others may generate cognitive stimulation,
which improves creativity in idea generation. A simulation modeling groups’ brainstorming
process has demonstrated that although participants gradually converge on fewer categories as
attention to each other goes up, group interaction is beneficial to group creativity if group
members can prime each other to generate ideas that they would not when working alone (Brown
et al., 1998). Research on crowdsourced idea generation has also shown that participants in
crowdsourcing can integrate their ideas with others’, and consequently produce more creative
ideas (Armisen & Majchrzak, 2014; Yu & Nickerson, 2011).
Whether exposure to other’s ideas may benefit or impede creativity seems to be
contingent on the quality of ideas one is exposed to. Research on crowdsourcing has been
consistent in showing that a crowd’s diversity is positively related to its creativity (Armisen &
Majchrzak, 2014; Boudreau, 2012; Jeppesen & Lakhani, 2010; Malhotra & Majchrzak, 2014;
Terwiesch & Xu, 2008), likely because exposure to a variety of opinions can further inspire
creativity among crowd participants.
23
Nijstad and colleagues (Nijstad, Stroebe & Lodewijkx, 2002) showed that participants
generated more diverse types of ideas after being exposed to heterogeneous ideas, whereas they
produced more ideas in a limited number of categories if exposed to homogeneous ideas.
Somewhat contradictory to this finding, evidence also suggests that exposure to a large number
of common, rather than original ideas, leads to higher productivity and idea originality
(Connolly, Routhieaux, & Schneider, 1993; Dugosh & Paulus, 2005). Since the number of idea
recall after idea exposure was found to be positively related to participants’ productivity (Dugosh
& Paulus, 2005; Dugosh, Paulus, Roland, & Yang, 2000; Paulus & Yang, 2000), this effect may
be explained by the fact that participants are able to remember more ideas when exposed to a
large number of common ideas (Dugosh & Paulus, 2005). When exposure to a significant
number of original and rare ideas in brainstorming, however, participants may be cognitively
overwhelmed and be unable to remember many of them. The potential cognitive stimulation
effect of original ideas is thus not realized.
Information about others’ performance. In general, research on group creativity shows
that participants’ productivity in idea generation is significantly boosted if they see a large
amount of ideas in brainstorming (Dugosh & Paulus, 2005; Dugosh et al., 2000; Nijstad et al.,
2002; Paulus, Kohn, Arditti, & Korde, 2013; Shepherd, Briggs, Reinig, Yen, & Nunamaker,
1995). Furthermore, participants tend to be more persistent in the idea generation task as group
size increases (Nijstad, Stroebe, & Lodewijkx, 1999). Taken together, these findings suggest that
information about others’ performance may stimulate productivity by enhancing social
comparison and competition within groups (Paulus & Dzindolet, 1993; Paulus et al., 2013;
Zajonc, 1965).
24
On the flip side, the social comparison process generated by information about others’
performance may also lead to downward performance matching among group members
(Camacho & Paulus, 1995; Paulus & Dzindolet, 1993): research suggests that individual
productivity in interactive brainstorming groups is more similar than that of nominal groups, and
that participants are more likely to match their productivity with low-performing members.
However, the creativity of interactive groups is significantly increased if they are given a
performance standard, in both face-to-face groups and computer-mediated groups (Paulus &
Dzindolet, 1993; Paulus, Larey, Putman, Leggett, & Roland, 1996; Shepherd et al., 1995).
Above all, prior research on group creativity has demonstrated that exposure to others’
ideas and behavior in group interactions leads to social and cognitive influence on group member
creativity in idea generation. However, the influence process still remains an under-investigated
black box. It is largely unclear under which conditions interaction with others will lead to
cognitive stimulation and social facilitation (thus improving creativity) and under what
conditions it will do the reverse.
There are two major limitations in this body of research, which at least partially lead to
the issue. The first is that groups have been studied as if they consisted of members with
homogeneous motivations. The second is that groups have been studied as if they are devoid of
context. Except for creativity research adopting the MIP-G perspective (Bechtoldt et al., 2010;
Beersma & de Dreu, 2005), little group creativity research has examined how variance in
member motivation or change in task structure may impact group creativity in brainstorming.
Nevertheless, the MIP-G model and theory of SVO suggest that group members differ in
their social motivation and have diverse preference regarding outcome distributions between
themselves and others. Their difference in SVO channels them to process distinct types of
25
information in social situations, which in turn impacts their creativity. The creative tasks
performed by groups may also diverge on task structure, which alters members’ cooperative
orientation and epistemic motivation in the social context. These factors, according to prior
research, may interact with SVO to influence group member cognition and behavior
(Eisenberger et al., 1992; Nijstad & de Dreu, 2012). Since SVO and task structure shape the
content and depth of individual information processing, they may impact how others’ ideas and
behavior influence member creativity in brainstorming. Combining research on group creativity
with the MIP-G model and theory of SVO can therefore shed light on the social and cognitive
influence processes in group creativity.
Summary
Taken together, the MIP-G model conceptualizes the effect of dispositional SVO on
group creativity. However, it is inadequate to fully explicate the social cognitive process in
ideation, which involves varying task structure and communication in the social context. The
theory of SVO has produced considerable insights regarding how task structure and
communication moderate the relation between SVO and cooperative behavior in behavioral
games. Yet its consequence on creativity still remains under-studied. Whereas prior research on
group creativity has extensively explored the influence of social interaction on creativity, it
generally treats groups as if they were made up of members with homogeneous SVOs and
devoid of task contexts.
The integrated insights of the MIP-G model, the theory of SVO and research on group
creativity suggest that the relation between SVO and creativity in crowdsourced idea generation
cannot be fully elucidated without taking into account the task context (i.e. task structure) and
26
the social influence process (i.e. impacts generated by others’ ideas and behavior) in interaction.
Building on prior research, additional research questions for the dissertation are proposed:
RQ2: How does task structure moderate the relation between SVO and creativity in
crowdsourced idea generation?
RQ3: How does communication moderate the relation between SVO and creativity in
crowdsourced idea generation?
Building on the three streams of research, the hypotheses regarding the effects of SVO,
task structure and communication are formulated and tested in the next four chapters (Chapters 4
to 7). Chapter 4 concerns a set of pilot studies conducted to explore the empirical context of
crowdsourced idea generation and to inform the design of the main experiments. Chapter 5
discusses the influence of SVO and task structure on creativity in crowdsourced idea generation
and reports the empirical results of an online experiment (Study 1) conducted to test the
proposed hypotheses. Chapters 6 and 7 explore the social influence process in social interaction.
Chapter 6 studies the influence of others’ ideas on creativity and how it may moderate the effects
of SVO and task structure. The results of the corresponding empirical study, Study 2, are
reported and discussed. Chapter 7 examines the effects of SVO, task structure and information
about others’ performance on creativity in crowdsourced idea generation. The findings of the
final online experiment (Study 3) are reported in this chapter.

27
CHAPTER 4: PILOT STUDIES
The empirical studies of the dissertation project consist of a series of online experiments.
Before designing the main experiments, I conducted a set of pilot studies. The purposes of the
pilot studies were to 1) inform the experimental design, particularly the choice of task and the
manipulation of task structure and social interaction; 2) determine the measure of SVO; and 3)
select an online participant recruitment platform. In this chapter, I will report the findings of the
pilot studies in detail and explain how they contribute to the designs of the main studies.
Pilot 1 is a case study of online crowdsourced idea generation projects, aiming to gain
empirical insights about the task topic, incentives, task structure and communication features in
online crowdsourced idea generation. Paulus and colleagues (Paulus, Brown, & Ortega, 1999)
have criticized prior research on group brainstorming as using tasks that are often not relevant
participants. Yet in organizations, brainstorming is often used when employees need new ideas
to help solve issues they encounter at work. One of the main purposes of Pilot Study 1 is
therefore to help create an idea generation task that mimics the topic, structure, and
communication pattern in crowdsourced idea generation and would be relevant and interesting to
the online participants.
Building on findings of Pilot Study 1, Pilot Study 2 consistes of four online experiments
designed to test if different idea generation tasks, SVO measures (i.e. the Triple-Dominance
Measure and the Slider Measure), and online participant recruit platforms (i.e. Amazon
Mechanical Turk and Prolific Academic) have an influence on participant behavior. The findings
of the four pilot experiments informed the final decision regarding the task topic, SVO measure
and recruitment platform in the main experiments.
28
Pilot Study 1: A Case Study of Crowdsourced Idea Generation Platforms
The purpose of Pilot Study 1 was to empirically probe the task topic, task structure and
communication features in online crowdsourced idea generation projects in order to inform the
design of the following pilot studies and the main experiments. Since the purpose of the first
pilot study was exploratory, I started with a qualitative case study by researching the idea
generation projects hosted by the online idea generation platform IdeaScale.
IdeaScale is a leading online crowdsourcing platform that hosts idea generation projects
for sponsors. Typically, a sponsor posts a question or problem on the platform to solicit ideas
from relevant participants. While many other crowdsourcing platforms only focus on one
particular type of topic (e.g. OpenIdeo focuses on social issues, Kaggle crowdsources machine
learning algorithms), the task topics vary greatly on IdeaScale, depending on the needs of the
sponsors. The sponsors on the platform differ in type and industry, ranging from government and
public services such as the Department of Labor of the United States, the City of Minneapolis
and the Ottawa Public Library, to private non-profit and for-profit organizations such as Magneti
Marelli, NBC and Columbia University.
The platform also offers considerable freedom for sponsors to decide on the incentives,
task structure and communication of the projects. For example, sponsors can decide what they
would like to offer for contributors and whether they would like their project to be a
collaborative community or to host a contest. While the platform provides a number of
communication features such as viewing and commenting on each other’s ideas in project
forums, sponsors can turn them on or off to allow or prohibit participant interaction. IdeaScale
therefore constitutes an idea platform in which users can scan a variety of task topics, structures
and communication mechanisms. I searched and accessed all the publicly available cases of
29
crowdsourced idea generation projects archived by IdeaScale in August 2016, resulting in a total
of 34 projects. For each project, I recorded its task topic, incentives, task structure and
communication features. Appendix A provides a summary of all the idea generation projects in
the case study.
Task Topic
The purpose of exploring the task topics of the crowdsourced idea generation projects
was to create a task that would be relevant and interesting for online crowdsourcing participants.
For task topics, I did not presume any categories before the research. Rather, I surveyed all the
project tasks to see if common themes would emerge. The sponsors of crowdsourced idea
generations have posted a variety of problems to solicit crowd ideas. For example, the
Department of Labor crowdsourced ideas from the public to help improve the employment and
careers of people with disabilities, whereas Magneti Marelli, an automotive component
manufacturer, was interested in getting solutions to improve engineer efficiency.
Despite the variety in task topics, three general types of tasks emerged, depending on
who would benefit from the crowdsourced ideas. The first type of tasks can be considered as
self-oriented, since they are problems that directly relate to contributors’ own usage or gains. A
typical example is a for-profit company (e.g. Electric Arts Inc.) asking its users to provide
opinions and ideas to improve its products or services. Although the contributed ideas are likely
to be harnessed by the company for profit-earning purposes, this type of task is categorized as
being more self-oriented because the participants do not need to be concerned about a collective
or community they belong to, but seem to be more driven by their preference for or unpleasant
experience with their own usage of a particular product.
30
The second type of task concerns community-oriented issues. This type of task usually
calls for solutions to problems that are related to a collective or a community. For instance, the
City of Huntsville, AL asked local residents to contribute their ideas regarding the future
development of the city. Of course, the ideas residents contribute would benefit themselves, but
the positive returns are not only limited to themselves, but also would benefit other residents in
the community, regardless of their contributions. This type of task therefore has a stronger public
goods nature than the self-oriented tasks discussed above.
Lastly, other-oriented tasks are problems or issues concerning a group of people who are
not immediately related to the idea contributors. Using the example mentioned earlier, the
Department of Labor crowdsourced ideas from the public to help improve the employment and
careers of people with disabilities. While individuals with disabilities and their families directly
benefit from the ideas, other contributors may also contribute without immediate gains, but
mainly for the sake of others.
Incentives
To investigate the incentives of the crowdsourced idea generation projects, I looked at
how contributors were rewarded by the task sponsors. In general, projects do not provide rewards
for participation, however some participants may get rewarded if their ideas are highly evaluated
or chosen to be implemented by the sponsor. I categorized the project incentives into either
financial or social rewards. Financial rewards were money or things with monetary values (e.g.
gift cards), whereas social rewards were incentives that involve public recognition or
appreciation. Implementation of ideas was considered as belonging to the social category, since
it provides social recognition of the ideas’ value. Some sponsors, such as Princess Cruises, chose
to reward the best idea contributors financially, while many others, such as NBC and Columbia
31
University, decided to reward contributors by social means. Of the 34 idea generation projects on
IdeaScale, 26 appreciated contributors through social rewards, 8 provided both social and
monetary incentives, and none used only monetary rewards.
Task Structure
In order to inform the task design, I also examined the task structure of the idea
generation projects on IdeaScale. I define task structure as the interdependence structure among
the project contributors. It can be either cooperative or competitive. Cooperative tasks are tasks
in which participants are encouraged to collaborate by task instructions and/or rewards and build
on each other’s ideas, instead of competing with each other to win a reward. In competitive
tasks, however, participants are explicitly instructed to behave as though they are in a contest in
which they should outperform others to be the winner.
On IdeaScale, the sponsors of crowdsourced idea generation projects can decide whether
their idea generation project is more cooperative, competitive, or both, by altering the
community structure and task scenario. 21 of the projects that I surveyed, including Princes
Cruises and the Department of Labor, mainly operated as a collaborative community. Although
some of the projects ended up rewarding the top contributors of ideas, competition is generally
not encouraged in these collaborative projects. Instead, participants are provided with a social
space where they can publically post ideas, talk to each other, and comment on each other’s
ideas. Eight projects, such as the Department of Energy, tried to motivate participants through
competition or contests. Another five of the projects employed both task structures. For instance,
the Laptime Club project sponsored by Magneti Marelli occasionally hosts ideation contests,
while primarily it remains a collaborative community in which people can comment and build on
32
each other’s ideas. These findings empirically prove the prevalence of cooperative and
competitive tasks in crowdsourced idea generation.
Communication
For each idea generation project, I also checked if communication was explicitly allowed
among participants on the IdeaScale platform. Particularly, I looked at whether the ideas
submitted were visible to all participants, whether participants could talk to each other, and
whether they could evaluate others’ ideas. Since IdeaScale enables community-like features that
allow participants to comment and vote on each other’s ideas, most idea generation projects on
the platform (27 out of 34) followed this design. In these projects, contributors publically post
their ideas and others can evaluate their ideas through comments or votes. However, some other
sponsors (e.g. the City of Atlanta) decided not to allow participant discussion and only solicited
ideas through private submissions. Idea contributors in the projects thus could not see what
others had submitted.
Summary
In short, Pilot Study 1 has revealed that task topics in crowdsourced idea generation can
be categorized into three types: 1) self-oriented (i.e. regarding private experience of a product or
service); 2) community-oriented (i.e. concerning issues relating to a collective that one belongs
to); and 3) other-oriented (i.e. aiming at helping others who are not directly related to the idea
contributor). Since participants’ SVO defines their dispositional care for others in the collective,
I believe that these three types of tasks may arouse differential interest from individuals with
distinct SVOs. Specifically, proself participants may be more driven by their private usage of
products but less interested in helping the community or others, whereas prosocials are more
likely to volunteer to contribute their ideas for the sake of their groups or others. Therefore, in
33
the subsequent pilot experiments, I examined whether prosocial and proself participants differ in
their interest and performance in these three types of tasks.
Second, the examination of project incentives showed no project employed only
monetary rewards, although many of them only rewarded top idea generation ideas socially
through public recognition and/or idea implementation. This provides further empirical support
for the argument that contributors of online crowdsourcing projects are not only driven by
pursuit of private gains, but also socially-oriented purposes. It again demonstrates the importance
of considering the variation in participant SVO in crowdsourced idea generation. The findings of
task structure also empirically prove the prevalence of cooperative and competitive tasks in
crowdsourced idea generation, and shows the importance of investigating participants’
performance in these two types of task structure. While it is not empirically uncommon for
crowdsourced idea generation projects to employ the elements of both cooperative and
competitive tasks, conceptually differentiating the two task structures and comparing participant
behaviors between them can potentially help crowdsourcing project sponsors and designers to
make more informed decisions.
Finally, the fact that the majority of the projects allowed communication indicates that
participants are highly likely to be exposed to and influenced by others’ ideas and behavior in
crowdsourced idea generation. It is thus essential for crowdsourcing project designers and
sponsors to understand the effects of communication on participant creativity in context.
Appendix A summarizes the task type, task structure, incentives and communication mechanism
of the crowdsourced idea generation projects surveyed in Pilot Study 1.
34
Pilot Study 2: SVO, Task Type, Task Interest and Creativity
Pilot Study 2 consisted of four experiments designed to examine whether individuals with
distinct SVOs differ in their interest and performance in different task types. In the experiments,
I also compared two commonly used measures of SVO, namely, the SVO Slider Measure
(Murphy & Ackermann, 2014; Murphy, Ackermann, & Handgraaf, 2011) and the Triple-
Dominance Measure (Kuhlman & Marshello, 1975b; Messick & McClintock, 1968). I tested
whether the distribution of SVO and its relation with task interest and performance would change
depending on the use of SVO measure. Moreover, the experiments were conducted recruiting
participants from two leading online crowdsourcing platforms, Amazon Mechanical Turk and
Prolific Academic. The results based on the two samples were compared to explore whether the
choice of platforms would influence the findings.
The first experiment investigated the effect of SVO on participants’ task interest,
applying the Slider Measure of SVO and recruiting participants on Amazon Mechanical Turk.
Experiment 2 examined the same question but employed the Triple-Dominance Measure to
gauge participant SVO. Experiment 3 followed the same design and research question, but
recruited participants on Prolific Academic. In Experiment 4, participants performed three
different idea generation tasks and the relation between their SVO and creativity was studied.
Participants in the final experiment were recruited on Amazon Mechanical Turk and their SVO
was measured using the Slider Measure. In the following the sections, I will start by introducing
the two measures of SVO as well as the two online crowdsourcing platforms. The method and
results of each experiment will then be reported and discussed.

35
Measures of SVO
Research on SVO typically measures the disposition employing the decomposed game
(Messick & McClintock, 1968). The decomposed game is a social dilemma in which participants
need to make a set of choices regarding resource distribution between themselves and a
hypothetical other. The participants’ SVO were classified based on their choices regarding the
resource gains for themselves and their assumed partners. For example, the participants who
consistently choose to maximize the joint resource gain of the dyad were classified as a
cooperative or prosocial individual. By contrast, individuals who choose to maximize their own
gains while neglecting the gains of others were classified as being individualistic- or proself-
oriented. Early research has demonstrated that in 2x2 matrix games such as the prisoner’s
dilemma, only cooperation and defection could be differentiated, and that even prosocially
oriented individuals could defect under some conditions (Kelley & Stahelski, 1970). As a result,
the decomposed game was created to delineate more diverse SVOs and to distinguish social
orientation from strategies applied in games.
In empirical operationalization, there are several variations of the decomposed game
measure. The most commonly used decomposed game measure is the Triple-Dominance
Measure (Kuhlman & Marshello, 1975b). The Triple-Dominance Measure consists of 12
multiple-choice questions. Every question contains three choices of value distributions, each
indicating a distinct SVO. Based on their choices, participants can be classified as having
prosocial orientations (cooperative, altruistic) or having proself orientations (individualistic,
competitive). Only participants who make consistent choices (i.e. demonstrating the same SVO
in 9 or more questions) can be classified. Table 1 provides a sample of the Triple-Dominance
Measure. In Table 1, choice A maximizes the resource difference between a participant and the
36
other. Choice B maximizes a participant’s own gain, whereas choice C maximizes the dyad’s
joint gain. A participant who consistently chooses choice C or its variation in the Triple-
Dominance Measure will be classified as being cooperative, or prosocial. Favoring choice A or B
suggests one’s proself orientations (A=competitive, B=individualistic). A complete version of
the Triple-Dominance Measure is included in Appendix B.
Table 1
An Example of the Triple-Dominance Measure
A B C
You get 480 540 480
Other gets 80 280 480

In recent years, Murphy, Ackermann and Handgraaf (2011) proposed another version of
the decomposed measure, which they call the Slider Measure. Similarly, participants need to
answer a set of questions regarding resource distribution between themselves and a hypothetical
other. However, in each question, they have more choices regarding the resource distribution,
and the choices are more closely clustered with each other compared to those in the Triple-
Dominance Measure. Murphy and colleagues claim that the Triple-Dominance Measure forces
the respondents to make a set of highly discrete choices. They argue that providing respondents
with an assembly of tightly connected choices can more accurately reflect the continuation in
individuals’ social preferences. SVO can also be calculated as a continuous score based on the
amount of resources each respondent allocates to themselves and others:

37
In the equation, A
s
is the total amount of resources that respondents allocate to
themselves, while A
o
is the sum of resources one allocates to the other. The score 22.45◦ is the
boundary between prosocial and proself orientation. Participants who score higher than 22.45◦
are classified as prosocials whereas participants who score lower than or equal to 22.45◦ are
classified as proselves. The measure was shown to have higher reliability and validity compared
to the Triple-Dominance Measure (Murphy et al., 2011). The researchers also proposed a web-
based form of the Slider Measure, which can be easily applied to online experimental platforms
such as Qualtrics. The web-based measure has two versions, which differ in sequence of choices.
Figure 1 shows an example of the web-based SVO Slider Measure. The full version of the
measure can be found in Appendix B. Apart from the Triple-Dominance Measure and the Slider
Measure, there are also other variations of the decomposed game, such as the ring measure
(Liebrand, 1984; Liebrand & McClintock, 1988), which have been used less often in prior
research and are therefore not discussed or tested in this dissertation.

Figure 1. An example of the SVO Slider Measure
Online Crowdsourcing Platforms for Participant Recruitment
There are multiple online crowdsourcing platforms through which researchers can recruit
volunteers to participate in their studies. Among them, the most frequently used platform is
Amazon Mechanical Turk (MTurk). MTurk is an online crowdsourcing platform on which
38
researchers (known as requesters on MTurk) can post relatively short tasks, or Human
Intelligence Tasks (HITs), for interested volunteers (workers) to participate. The task requesters
can set their own criteria for task participation. For example, a commonly used criterion by
requesters is approval rate above 95%, which means that the workers need to be approved by
more than 95% of the requesters for whom they previously have completed tasks. Interested
workers who meet the participation criteria can choose to accept and perform the HITs. They
will get paid only if the HIT requesters approve their submission. Apart from providing an
efficient method for data collection, recent research has suggested that the participants recruited
via MTurk are more representative than traditional data collection methods which are typically
based on college participant pools (Berinsky, Huber, & Lenz, 2012; Buhrmester, Kwang, &
Gosling, 2011). The studies conducted via MTurk are also at least as reliable as traditional
recruiting methods.
Another platform that has been gaining popularity in recent years is Prolific Academic.
Prolific Academic is similar to MTurk in the basic recruiting mechanism. However, it provides
more choices for researchers to screen participants. For instance, researchers can screen
participants based on demographic variables such as political affiliation or smoking habits in
addition to participant performance. It also provides features that make it easier to conduct
longitudinal studies and prevent participants from repeatedly participating in the same study.
Unlike MTurk, Prolific Academic mandates a minimum payment from task requesters for
participants (£5 an hour). To decide on the platform to be used in the main experiments, Pilot
Study 2 compared the findings of two experiments, each recruiting participants from one of these
two online platforms.

39
Experiment 1: SVO and Task Interest with the Slider Measure on MTurk
Method. The first experiment in Pilot Study 2 examined whether participant SVO
influences their interest in the three different task types in crowdsourced idea generation. The
experiment followed a 2 (SVO: prosocial vs proself) X 3 (Task type: self-, community-, or other-
oriented task) mixed factorial design. The first factor, SVO, is a between-subjects factor. It was
measured using the Slider Measure. The participants of Experiment 1 were recruited via Amazon
MTurk.
Task type is a within-subject factor. Based on the findings of Pilot Study 1, I created
three types of idea generation tasks: a self-oriented task, a community-oriented task and an other-
oriented task. There are three specific problems under each task type. The self-oriented problem
is one that is directly related to the workers in MTurk. An example is “please suggest changes to
Amazon Mechanical Turk to improve the efficiency of task completion in the system.” The
community-oriented task is an issue that is related to the broader community of the participant.
An example of the task is “what should the public park in our local neighborhood be like?”
Lastly, the other-oriented problem concerns a remote other who is unlikely to be directly related
to the participants. An example would be “how to use mobile technology to facilitate health care
access in developing countries?” Each participant in the experiment reported their interest in
contributing ideas to help resolve the three types of problems in randomized orders.
The participation requirement was set to above 95% approval rate and more than 100
HITs completed. After accepting the HIT on MTurk, the participants were directed to a Qualtrics
survey to perform the task. Each approved participant was paid $0.25. Appendix C discloses the
materials used in Pilot Experiment 1, including the MTurk recruitment ad, a complete list of the
idea generation tasks, and the experimental measures.
40
Results. A total of 160 valid responses were received. 4 respondents were classified as
competitors, 63 were individualists, 93 were cooperators, and none were altruists. Since the
number of competitors was small, competitors and individualists were combined as pro-selves
(67). 62.9% of the participants were between 25-44 years old. 51.3% of them were female and
91.8% of them had had some college education or higher.
The mixed ANOVA analysis showed no effects of SVO on participants’ general interest
in participating in these idea generation tasks, F(1,158)=0.04, p =.85. There was also no
interaction effect of SVO and task type, F(1,158)=1.00, p =.32. However, there was a significant
main effect of task type, F(1,158)=14.05, p < .01, h
p
2
= .08. Simple effect analysis revealed that
in general, participants had significantly higher interest in the self-oriented (M=4.79, SD=1.38)
and community-oriented tasks (M=4.73, SD=1.43) than the other-oriented tasks (M=4.37,
SD=1.57), t(159)=3.63, p < .01 and t(159)=3.01, p < .01. There was no significant difference
between the interest in the self-oriented and community-oriented tasks, t(159)=0.54, p =.59.
Discussion. Applying the Slider Measure to gauge the SVO of participants recruited via
MTurk, Experiment 1 did not find any effect of SVO on participants’ interest in online idea
generation tasks. However, participants were generally more interested in the self-oriented and
community-oriented tasks than the other-oriented tasks. This effect was consistent for both
proself and prosocial participants.
Experiment 2: SVO and Task Interest with the Triple-Dominance Measure on MTurk
Method. The second pilot experiment examined the same research question as
Experiment 1: whether participant SVO influences their interest in the three different task types
in crowdsourced idea generation. The experiment followed the same 2 (SVO: prosocial vs
proself) X 3 (Task type: self-, community-, or other-oriented task) mixed factorial design. Each
41
participant in the experiment reported their interest in contributing ideas to help resolve the three
types of problems in a randomized order. The tasks, measures, and participant recruitment
platform used in Experiment 2 were also identical to those in Experiment 1. The only difference
was that participant SVO was measured by applying the Triple-Dominance Measure rather than
the Slider Measure.
Results. A total of 186 valid responses were received, however 29 participants’ SVO
choices were inconsistent and thus unclassifiable by the Triple-Dominance Measure. Of the
remaining 157 participants, 15 were classified as competitors, 85 were cooperators, and 57 were
individualists. The distribution of SVO types was similar to Experiment 1, except that slightly
more competitors were identified. The competitors and individuals were combined as proselves
(72). 65% of the participants were between 25-44 years old. 52.9% of them were male and
89.8% of them had had some college education or higher.
Consistent with Pilot Experiment 1, a mixed ANOVA analysis did not support the effects
of SVO on participants’ interest in participating in idea generation tasks, F(1,155)=0.03, p =.87.
There was also no interaction effect of SVO and task types, F(1,155)=0.01, p =.91. Again, there
was a significant main effect of task types, F(1,155)=47.36, p <.01, h
p
2
= .23. Consistent with
the first pilot experiment, participants reported having significantly higher interest in the self-
oriented (M=5.76, SD=1.23) and community-oriented tasks (M=5.42, SD=1.05) than the other-
oriented tasks (M=4.79, SD=1.55), t(156)=6.76, p < .01 and t(156)=5.72, p < .01. They were also
significantly more interested in self-oriented tasks than community-oriented tasks, t(156)=3.13, p
< .01.
Discussion. Although Experiment 2 employed a different measure, the Triple Dominance
Measure, to assess the SVO of participants recruited via MTurk, its results are generally
42
consistent with those of Experiment 1. In addition to the similar distribution of participant SVOs,
Experiment 2 also did not reveal any main effect of SVO or interaction effect of SVO and task
type on participants’ interest in online idea generation. The main effect of task type was found
again: participants were more interested in the self-oriented and community-oriented tasks than
the other-oriented tasks. However, whereas Experiment 1 did not show any difference between
participant interest in self-oriented or community-oriented tasks, Experiment 2 demonstrated that
participants were more interested in generating ideas to solve self-oriented issues.
Experiment 3: SVO and Task Interest with the Slider Measure on Prolific Academic
Method. Consistent with Experiments 1 and 2, the third pilot experiment tested the effect
of SVO on participant interest in the three different task types in crowdsourced idea generation.
It thus followed the same design as the first two experiments: a 2 (SVO: prosocial vs proself) X 3
(Task type: self-, community-, or other-oriented task) mixed factorial design. It applied the same
Slider Measure of SVO as Experiment 1. The main difference of Experiment 3, compared to the
previous experiments, was that it recruited participants through Prolific Academic rather than
MTurk. After accepting the task on Prolific Academic, the participants were directed to a
Qualtrics survey to perform the task. Each approved participant was paid $1.
Results. A total of 130 valid responses were received. Among them, 41 were
individualists (proselves) and 89 were cooperators (prosocials). No competitors or altruists were
identified. 85.4% of the participants were between 18-35 years old. 59.2% of them were male
and 81.5% of them had had some college education or higher.
Similar to the two prior pilot experiments, the mixed ANOVA analysis did not suggest
any effects of SVO on participants’ interest in participating in idea generation tasks,
F(1,128)=2.10, p =.15. There was also no interaction effect of SVO and task types,
43
F(1,128)=0.07, p =.79. Yet again, there was a significant main effect of task types,
F(1,128)=11.40, p < .01, h
p
2
= .08. Consistent with previous findings, participants reported being
significantly more interested in the self-oriented (M=5.58, SD=1.20) and community-oriented
tasks (M=5.47, SD=1.09) than the other-oriented tasks (M=5.14, SD=1.35, t(129)=3.54, p < .01
and t(129)=2.88, p < .01). There was no significant difference between their interest in the self-
oriented and community-oriented tasks, t(129)=0.93, p =.36, which is the same as in Pilot
Experiment 1 but not Pilot Experiment 2.
Discussion. Recruiting participants via Prolific Academic and assessing their SVO with
the Slider Measure, Pilot Experiment 3 identified significantly more prosocial individuals than
the two prior experiments. Participants recruited on Prolific Academic were also slightly younger
than those recruited through in MTurk. Despite these differences, the experiment produced
similar findings: whereas there was no main effect of SVO or interaction effect of SVO and task
type on participant interest, participants varied in their interest in different task types. The main
effect of task type was identical to that found in Experiment 1: although participants were more
interested in the self-oriented and community-oriented tasks than the other-oriented tasks, they
were not significantly different in their interest in self-oriented or community-oriented tasks.
Nevertheless, Experiment 2 did discover a significant difference between participant interest in
the two tasks. This minor inconsistency should not be due to the change in SVO measures or
recruiting platforms, since there was no main or interaction effect of SVO on task interest, and
the fact that participants recruited through two different online platforms produced the same
results in Pilot Experiments 1 and 3.

44
Experiment 4: SVO, Creativity and Crowdsourced Evaluation of Idea Usefulness
Method. Instead of studying the effect of SVO on task interest, the last pilot experiment
tested the impact of SVO and task type on participant creativity in online idea generation.
Similar to the designs of prior pilot experiments, this study followed a 2 (SVO: prosocial vs
proself) X 3 task type (self-, community-, or other-oriented task) mixed factorial design. The first
factor, SVO, is a between-subjects factor. It was measured using the Slider Measure.
Task type is a within-subject factor. However, rather than reporting their interest in the
tasks, participants generated ideas to help solve three types of problems: self-oriented,
community-oriented and-other oriented. There was one question for each task type, and the
sequence of the three questions was randomized. The tasks were chosen from the list of tasks
used in the previous pilot experiments. These were the most popular tasks rated by participants in
each task type. The self-oriented problem was: “How might Amazon Mechanical Turk improve
the efficiency of task completion in the system?” The community-oriented task was “What
features should the public park in our local neighborhood have?”. Lastly, the other-oriented
problem was “How can we use mobile technology to facilitate health care access in developing
countries?”
The participants were recruited on MTurk. The participation requirement was set to
above 95% approval rate and more than 100 HITs completed. After accepting the HIT on
MTurk, the participants were directed to a Qualtrics survey to perform the task. Each approved
participant was paid $0.75. The MTurk Ad and tasks used in Experiment 4 are listed in Appendix
D of the dissertation.
Since prior research has suggested that being creative involves generating ideas that are
both original and useful (Amabile, 1983, 1996), the creativity of the participants in the idea
45
generation task was measured based on three dimensions: 1) idea productivity, 2) originality of
their ideas, and 3) usefulness of their ideas.
Productivity. Productivity was measured as the number of ideas generated by each
participant. The respondents generated 1065 ideas in total. On average, each participant
produced 7.66 total ideas (SD=4.14), with 2.13 (SD=1.17) ideas in the self-oriented task, 3.8
ideas (SD=2.91) in the community-oriented task and 1.73 ideas (SD=0.85) in the other-oriented
task.
Originality. An original idea is an idea that is unique. I judged the originality of an idea
based on how frequently it was generated by the participants of the study (Thompson &
Brajkovich, 2003). Statistically, an idea generated by less than five percent of the participants in
the study is rare, and was given an originality score of 1. Other ideas were given a score of 0. To
illustrate, the current study had 139 participants. 5% of 139 is 6.95. An idea that was generated
by fewer than or equal to six people was given an originality score of 1. Participants’ originality
was the average originality score of all the ideas they generated, multiplying by 7. The
multiplication was conducted to make the originality score equivalent in scale with the
usefulness score. On average, the originality score of each participant was 1.62 (SD=1.46). For
the self-oriented, community-oriented and other-oriented tasks, the mean originality scores of
participants were 2.70 (SD=2.74), 0.96 (SD=1.77) and 1.42 (SD=2.46), respectively.
Usefulness. Since prior research on MTurk has demonstrated that workers can evaluate
the quality of others’ ideas in a way that strongly correlates with experts (Green, Seepersad, &
Hölttä-Otto, 2014; Heilman & Smith, 2011; Kittur, Chi, & Suh, 2008), I crowdsourced the
evaluation of idea usefulness on MTurk. An individual HIT was created for the idea evaluation
task. After accepting the HIT, the participants were directed to a Qualtrics survey, in which they
46
evaluated the usefulness of six ideas based on a 7-point scale ranging from “not useful at all” to
“extremely useful”. Each approved participant was paid $0.50. The materials used in the
crowdsourced idea evaluation can be found in Appendix E.
A total of 631 participants successfully completed the evaluation task. Each idea was
rated by an average of 3 raters. To ensure that the ratings given by workers were of high quality,
I followed the recommendation of prior literature (Mason & Suri, 2012), and checked to see
whether each rating disagreed with other ratings of the same idea. For each idea, I removed the
ratings that were extremely deviant from others (more than 1 standard deviation away from the
mean rating), and calculated the usefulness of an idea as the mean of all the ratings left.
Averaging multiple ratings solicited from the crowd was also a typical method used in previous
work utilizing crowd evaluations (Kittur, Smus, Khamkar, & Kraut, 2011). To illustrate, if the
three usefulness ratings for Idea 1 were 2, 6 and 7, I removed the rating 2, since it is more than 1
standard deviation (2.65) away from the mean 5. The usefulness of Idea 1 was then averaged to
6.5 (the mean of 6 and 7).
I randomly selected 100 ideas (9.4%) from all the ideas generated by the participants and
rated the usefulness of the ideas. The intra-class correlation (ICC) coefficient of my rating and
the MTurkers’ rating was 0.71. I concluded that the MTurkers’ evaluation of idea usefulness was
fairly reliable. Participants’ idea usefulness was then calculated as the mean usefulness score
received by all the ideas they generated. The average idea usefulness for participants was 3.58
(SD=0.46). For the self-oriented, community-oriented and other-oriented tasks, the mean
usefulness scores were 3.63 (SD=0.76), 3.52 (SD=0.82) and 3.58 (SD=0.73).
Results. A total of 139 valid responses were received. 1 participant was classified as a
competitor, 63 were individualists, and 75 cooperatives/prosocials. No altruists were identified.
47
The competitor and the individualists were combined as proselves (64). 84.1% of the participants
were between 25-44 years old. 37.4% of them were female and 93.4% of them had had some
college education or higher.
The mixed ANOVA analysis revealed no effects of SVO on participants’ productivity in
idea generation tasks, F(1,137)=0.85, p =.36. There was also no interaction effect of SVO and
task types either, F(1,137)=0.06, p =.81. However, there was a significant main effect of task
types, F(1,137)=14.87, p < .01, h
p
2
= .10. Participants generated the most ideas in the
community-oriented task (M=3.80, SD=2.91), followed by the self-oriented task (M=2.13,
SD=1.17). The difference between participants’ productivity in the two task types was
significant, t(138)=8.07, p < .01. Participants generated significantly fewer ideas in the other-
oriented tasks (M=1.73, SD=0.85) than in the community- and self-oriented tasks, t(138)=-9.12,
p < .01 and t(138)=-3.86, p < .01, respectively.
The mixed ANOVA analysis of idea usefulness suggested no effects of SVO,
F(1,103)=2.60, p = .11, task type, F(1,103)=0.09, p = .77, or their interaction, F(1,103)=0.59, p
= .44. SVO did not significantly influence idea originality either, F(1,137) = 0.42, p = .52. There
was also no interaction effect of SVO and task type on idea originality F(1,137) = 0.00, p = .97.
However, task type had a significant main effect on idea originality, F(1,137)=15.50, p < .01, h
p
2

= .10. Ideas generated in the self-oriented task (M=2.70, SD=2.74) were significantly more
original than those in the community- (M=0.96, SD=1.77) or other-oriented tasks (M=1.42,
SD=2.46), t(138)=7.13, p < .01 and t(138) =3.97, p < .01. Idea originality of community-oriented
tasks was slightly lower than that in other-oriented tasks, but the difference was not statistically
significant, t(138)=-1.72, p = .09.
48
Discussion. Experiment 4 investigated the effect of SVO and task type on participant
creativity in crowdsourced idea generation. The results of the experiment did not reveal any main
or interaction effect of SVO on participant creativity, measured as idea productivity, originality
and usefulness. But participant creativity differed across task types: while participants were more
productive in community-oriented tasks than in self-oriented tasks, and more so than in other-
oriented tasks, their idea was most original when they performed a self-oriented task. The
findings suggest that self-oriented tasks may promote participant originality and enhance their
creativity compared to other-oriented tasks. Yet although participants produced more ideas in the
community-oriented tasks, their originality was not stimulated by the task.
General Discussion
Based on the findings of the pilot studies, I decided to adopt the self-oriented task in the
main experiments because participants were interested in the task and performed well in terms of
both idea quantity and quality. Pilot Study 2 revealed that self- and community-oriented tasks
were more interesting to participants compared to other-oriented tasks. Although participants
generated the most ideas in community-oriented tasks, their ideas were of the lowest originality.
In addition, the self-oriented task (asking MTurk workers to generate suggestions to help
improve requesters’ task designs) resembles real organizations’ crowdsourcing projects, which
solicit ideas from users to develop their products or service. It is a commonly used type of task to
harness distributed information through digital media (von Hippel, 1994, 2006).
To measure participant SVO in the main experiments, I decided to apply the Slider
Measure. The two different measures of SVO, namely, the Slider Measure and the Triple-
Dominance Measure, did not lead to significant difference in the distribution of SVOs or the
findings regarding the effect of SVO on individual interest in idea generation tasks. However, the
49
Slider Measure was reported to have higher reliability and validity compared to the Triple-
Dominance Measure (Murphy et al., 2011). Unlike the Triple-Dominance Measure, which
usually leaves out a small number of unclassifiable participants, the Slider Measure is also able
to identify the SVO of all participants, since it provides a closely clustered set of choices. It also
has a web-based version that can be easily employed in online platforms such as Qualtrics.
Furthermore, the pilot experiments suggested that competitors and altruists are relatively
rare in the online population on MTurk or Prolific Academic. Combining the empirical findings
with prior research, I classified participants in the main studies into two types: proself
(competitive and individualistic) and prosocial (cooperative and altruistic) participants.
In order to access participants with more equally distributed SVOs, I selected Amazon
MTurk as the participant recruitment platform. Although the pilot studies did not reveal any
effect of platform choice on study findings, participants recruited through Prolific Academic
tended to be younger and more prosocial compared to those on MTurk. MTurk also has a larger
participant pool and has been more often used in prior research.
Lastly, I decided to rely on MTurk workers to evaluate the usefulness of the ideas
generated by their peers. Crowdsourcing idea evaluation through MTurk has been applied in
prior research (Green et al., 2014). Experiment 4 crowdsourced the evaluation of idea usefulness
on MTurk. The crowdsourced evaluations displayed high level agreement with my own
evaluation, suggesting good reliability.
50
CHAPTER 5: SVO, TASK STRUCTURE AND CREATIVITY IN CROWDSOURCED
IDEA GENERATION
Building on the MIP-G perspective on creativity and research on SVO, the current
chapter postulates a set of hypotheses regarding how SVO may influence individual creativity in
online idea generation, and how the relation may be moderated by task structure in context. An
online experiment was conducted to test the hypotheses. Its results are reported and discussed.
Theory and Hypotheses
Since crowdsourced idea generation tasks are typically open, public, and provide
relatively low private incentives, participants’ contribution cannot be fully explained by pursuit
of self-oriented goals. Researchers have argued that socially-oriented motivations drive the
participation, cognition and behavior of many participants in open collaborations such as
crowdsourced idea generation (Adler, 2001; Benkler, 2002; Powell, 1990). Variation in
participant SVOs is thus a fundamental feature that characterizes crowdsourced idea generation.
In cognitive tasks such as idea generation, SVO biases participants’ information processing
(Nijstad & de Dreu, 2012): whereas prosocial individuals are more likely to process information
in a way that promotes integration and collective outcome, proselves tend to process information
to boost their individual gains.
The MIP-G model posits that compared to proself orientation, prosocial orientation is
more beneficial for participant creativity in crowdsourcing (de Dreu et al., 2011; Nijstad & de
Dreu, 2012). To begin with, prosocial orientation it is likely to cultivate trust, cohesion and
psychological safety among participants in social settings. Since being creative involves risk
taking, a social climate that is trusting and safe encourages participants to think more flexibly
and innovatively (Carnevale & Probst, 1998; Nijstad & de Dreu, 2012; Taggar, 2002).
51
Prosocially-oriented individuals are also more likely to be driven by intrinsic motivation
in crowdsourced idea generation. Prosocial individuals are dispositionally concerned about
others and the collective. Prior research on SVO shows that proself individuals are driven by
their pursuit of self-interest and are more likely to be influenced by external changes that affect
their benefits (van Kleef & van Lange, 2008; van Vugt, Meertens Ree M., & van Lange, 2006).
By contrast, the behavior and cognition of prosocials is more guided by their internal morality,
sense of fairness and belief in equality (Bem & Lord, 1979; Liebrand, Jansen, Rijken, & Suhre,
1986).
In crowdsourced idea generation tasks, the monetary rewards for participants are often
trivial. Participants typically do not enjoy direct or exclusive benefits from their idea
contribution, either. Compared to proselves, prosocial individuals are more likely to be
intrinsically interested in contributing their ideas to the collective task when no extra private
incentives are provided (Karau & Williams, 1993). Research on SVO has consistently
demonstrated that prosocial participants are more likely to contribute to the collective good or
the good of others regardless of rewards or task structure (Balliet et al., 2009; Bogaert et al.,
2008; Kuhlman & Marshello, 1975a; Liebrand, 1984; Liebrand & van Run, 1985).
Since research on creativity has established that intrinsic motivation enhances creativity
(Amabile, 1996), prosocial orientation is therefore more likely to lead to higher participant
creativity in crowdsourcing. Research has revealed that prosocial orientation promotes creativity
more than proself orientation in idea generation (Bechtoldt et al., 2012, 2010). Prosocial
orientation also strengthens the positive effect of intrinsic motivation on creativity, because
interested prosocial participants are more likely to take others’ perspectives and in turn generate
ideas that are both original and useful (Grant & Berry, 2011). Therefore:
52

H1: Prosocials are more creative than proselves in crowdsourced idea generation.

Nevertheless, a number of studies have produced contradictory evidence, showing that
proself value orientation can positively influence individual creativity (Goncalo & Duguid, 2012;
Goncalo & Kim, 2010; Goncalo & Staw, 2006; Janssen & Huang, 2008). Although proself
individuals are less interested in working to help others, they seem to be more likely to pursue
different personal goals and generate original and unique ideas. Interactive groups instructed to
pursue different goals or think of themselves as different were also found to be more creative in
idea generation (Goncalo & Staw, 2006; Nemeth, Personnaz, Personnaz, & Goncalo, 2004).
Even a moderate level of narcissism can be beneficial to group creativity in idea generation tasks
(Goncalo et al., 2010). Furthermore, in intact teams, researchers have shown that whereas team
identification (indicating high prosocial orientation) was associated with prosocial behavior
within teams, individual differentiation (indicating high proself orientation) was positively
related to creative behavior (Janssen & Huang, 2008).
To resolve the inconsistency in the existing literature, I propose task structure as a
moderating factor between SVO and creativity in crowdsourced idea generation. Task structure
has been identified by existing research on SVO as a critical contextual factor that interacts with
SVO to impact individual behavior in groups because it dictates the interdependence structure
among participants (Liebrand, Wilke, et al., 1986; McClintock & Liebrand, 1988). As suggested
by Pilot Study 1, there are generally two types of task structure in crowdsourced idea generation:
cooperative tasks and competitive tasks (Boudreau & Lakhani, 2009, 2013; Malone et al., 2010;
Yu, Nickerson, & Sakamoto, 2012). On cooperative tasks, participants are encouraged to work
collaboratively through reward structure (group reward) or instructions. In competitive tasks,
53
participants try to perform better than others to win a contest and the best performer usually gets
some type of reward (e.g. reputation, recognition, or money).
Considerable research indicates that competition can promote participant creativity in
crowdsourcing. In his original conceptualization, Osborn (1953) outlined rivalry/competition as a
major driver for creativity in collective brainstorming. Scholars have shown that groups
instructed to be competitive tend to be more creative in idea generation (Beersma & de Dreu,
2005; Nemeth et al., 2004). Competition may further stimulate individual creativity when the
task topic is related to competition and conflict (de Dreu & Nijstad, 2008). In crowdsourcing
contests, research has revealed that competition increased participants’ effort if the task was
uncertain (Boudreau et al., 2011). Since creative tasks are highly uncertain, participants’ efforts
are likely to be increased in competitive crowdsourced idea generation.
Specifically, evidence suggests that the positive impact of competitive tasks on creativity
is mainly attributed to its stimulating effect for proself participants but not prosocials. The
cognition and behavior of prosocials are mainly guided by their concern for collective good and
their internal morality (Bem & Lord, 1979; Liebrand, Jansen, et al., 1986). For prosocials, cues
of trust and belief in equality are necessary to induce high cooperation (Bogaert et al., 2008).
They are consistently more likely to contribute to collective good regardless of rewards or task
structure (Balliet et al., 2009; Bogaert et al., 2008; Kuhlman & Marshello, 1975a; Liebrand,
1984; Liebrand & van Run, 1985).
By contrast, proselves need extra external incentives or control to help align their own
goals with the group goals. Otherwise, they are more likely to free-ride or take advantage of
others (de Kwaadsteniet et al., 2006; Galletta et al., 2003; Karau & Williams, 1993). Proselves
also report having higher motivation for achievement than prosocials (Platow & Shave, 1995).
54
Competition should act as an extra private incentive for proselves since it cultivates rivalry and
rewards individual achievements.
Indeed, research has revealed that proselves, compared to prosocials, are more likely to
be motivated by competition in behavioral games (Pulford, Krockow, Colman, & Lawrence,
2016). In group brainstorming, proself groups also generate more ideas than prosocial groups
when they are primed with equity rather than equality-based resource distribution (Goncalo &
Kim, 2010). This relation is mediated by the competitive behavior of group members in idea
generation. Taken together, the following hypothesis is proposed:

H2: The creativity of proselves in crowdsourced idea generation depends on the task
structure, such that proselves are more creative on competitive tasks than on cooperative tasks.

Study 1
Participants
Participants of Study 1 were recruited via Amazon Mechanical Turk (MTurk). I created a
Human Intelligence Task (HIT) to recruit and pay participants on MTurk. To ensure that I
recruited high quality MTurk workers, I set their approval rate to above 95% (more than 95% of
their previous work was approved by requesters). I also requested the workers to have completed
at least 100 HITs on MTurk. I did not set the number too high (e.g. above 1000) because I did
not want to include only experienced workers or exclude new workers.
A total of 176 participants successfully completed the task and each was paid $0.75. As
promised, one of the participants was randomly drawn to win a $15 lottery in each task
condition. 39.2% of the participants were male. 64.2% of the participants were between 25-44
years old and 90.4% had had some college education or higher.
55
Materials
Study 1 followed a 2 (SVO: prosocial vs proself) X 2 (Task structure: cooperative vs
competitive) between-subjects factorial design. The experiment was conducted online using
Qualtrics.
SVO. Participants were classified as having prosocial- or proself-orientation based on
their SVO. SVO was measured using the SVO Slider Measure (Murphy et al., 2011). I used the
web-based version of the Slider Measure provided by the measure developers. The web-based
measure has two versions, which differ in the sequences of choices. I randomly assigned the
participants to one of the two versions in Qualtrics. A complete version of the measure can be
found in Appendix B.
When completing the measure, participants had to make a set of choices regarding
resource distribution between themselves and a hypothetical other. A SVO score was then
calculated based on their choices. The SVO score 22.45◦ was used as the boundary between
prosocial and proself orientation (Murphy et al., 2011). Participants who scored higher than
22.45◦ were classified as prosocials, whereas participants who scored lower than or equal to
22.45◦ were classified as proselves. Of the 176 participants, 69 were classified as proselves,
while 107 participants were prosocials.
Task structure. Participants were randomly assigned to either a cooperative task or
competitive task. On the cooperative task, participants were told that they were collaborating
with a group of workers on MTurk to win a lottery for a task bonus:
“You will collaborate in groups to generate ideas to answer a question posted by the
requester. The HIT you are completing are grouped into batches of 9 assignments. You will
collaborate with 8 other workers who accept the same batch to get the chance to earn a $15
56
bonus. You will enter a lottery for the bonus after the HIT is completed. For each non-redundant
idea your batch generates, you get one chance in the lottery for the bonus. The more non-
redundant ideas your batch collectively generates, the more likely you and other workers in your
batch are to receive the bonus.”
On the competitive task, participants were told that they were competing in groups to win
a lottery for a task bonus:
“You will compete in groups to generate ideas to answer a question posted by the
requester. The HIT you are completing are grouped into batches of 9 assignments. You will
compete with 8 other workers who accept the same batch to get the chance to earn a $15 bonus.
If you generate the most non-redundant ideas in your batch, you will enter a lottery to win the
bonus after the HIT is competed. For each non-redundant idea you generate, you get one chance
in the lottery for the bonus. The more non-redundant ideas you generate, the more likely you are
to receive the bonus.”
A similar manipulation of task structure has been used in prior research (Bechtoldt et al.,
2012, 2010).
Manipulation Check. To make sure that the participants were paying attention to the
task manipulation, respondents answered a multiple-choice question after reading the task
instruction: “In this task, I will: a) generate non-redundant ideas to win a $5 bonus; b)
collaborate with my group of workers to generate non-redundant ideas and win a $15 bonus; c)
compete with my group of workers to generate non-redundant ideas and win a $15 bonus; d)
evaluate others’ ideas to win a $10 bonus.” Only the participants who correctly answered the
question were included in the study.
57
In the post-experiment survey, participants also answered two questions that gauged their
collaborative orientation during task performance. They rated two statements on a 7-point Likert
scale ranging from “strongly disagree” to “strongly agree”: “When I was generating the ideas, I
tried to contribute to my batch of workers” (cooperative orientation) and “When I was generating
the ideas, I tried to perform better than other workers in my batch” (competitive orientation).
Task. The idea generation task performed by the participants was: “What can MTurk
requesters do to ensure workers perform their HITs with good quality?” The task was adapted
from a self-oriented task in the pilot studies. The task was chosen because participants in the
pilot studies were highly interested in the self-oriented tasks and performed well on the task type
in terms of both idea quantity and quality. The task also resembles real organizations’
crowdsourcing projects, which typically solicit ideas from their users to develop their products or
service. It was expected to be a task that MTurkers would find relevant and would be
knowledgeable about.
Measures. Since prior research has suggested that being creative involves generating
ideas that are both original and useful (Amabile, 1983, 1996), the creativity of the participants in
the idea generation task was measured based on three dimensions: 1) idea productivity, 2)
originality of their ideas, and 3) usefulness of their ideas.
Productivity. A total of 1096 ideas were generated by the participants. Productivity was
measured as the number of ideas generated by each participant. On average, each participant
produced 6.23 ideas (SD=3.59).
Originality. An original idea is an idea that is unique. As in the pilot study, I judged the
originality of an idea based on how frequently it was generated by the participants of the study
(Thompson & Brajkovich, 2003). An idea that was generated by less than five percent of the
58
participants in the study was given an originality score of 1, whereas others were given a score of
0. The current study had 176 participants. 5% of 176 is 8.8. Thus, an idea that was generated by
fewer than or equal to 8 people was given an originality score of 1. Participants’ originality was
the average originality score of all the ideas they generated, multiplied by 7. The multiplication
was conducted to make the originality score equivalent in scale with the usefulness score. On
average, the originality score of each participant was 2.07 (SD=1.66).
Usefulness. Based on Pilot Study 2, the usefulness of the ideas was crowdsourced. I
relied on MTurkers to judge the usefulness of the ideas since they have the best knowledge
regarding to what extent the ideas can lead to higher quality of their work. An individual HIT
was created for the idea evaluation task. After accepting the HIT, the respondents were directed
to a Qualtrics survey in which they evaluated the usefulness of 20 ideas based on a 7-point scale
ranging from “not useful at all” to “extremely useful”. The recruitment ad and measures used in
the evaluation HIT are the same as in the Pilot Study (Appendix E).
460 valid responses were received. Each participant was paid $0.50. Since each idea
received multiple usefulness ratings, ratings that were extremely deviant from others (more than
1 standard deviation away from the mean usefulness rating) were removed. After the adjustment,
on average, each idea was rated by 4.5 raters.
An idea’s usefulness was then calculated as the average score given by all the evaluators
of the idea. 183 ideas (16.7%) were randomly selected from the ideas and rated by the
researcher. The intra-class correlation (ICC) between the researcher’s evaluation and the mean
usefulness score given by the MTurk workers was 0.85, indicating good reliability of the MTurk
workers’ ratings. Participants’ idea usefulness was then calculated as the mean usefulness score
59
received by all the ideas they generated. The average idea usefulness for participants was 4.78
(SD=0.98).
Covariates. Since the current study was an online experiment, participants did not take
part in the study in a controlled environment. Although this situation can potentially improve the
external validity of the study, it may also introduce exogenous factors that may impact
participants’ effort in the study. In their review of the analysis of covariance (ANCOVA), Miller
& Chapman (2001) argued that ANCOVA can be a useful tool to remove noise in the dependent
variable(s) and increase the power of statistical analysis. But it is under-used in social science
research.
To control for noise in the task environment, I conducted an ANCOVA analysis,
including measures of task efforts as covariates. Participants’ task effort in the study was
measured in two ways. The first was a self-reported measure using the statements “I was
motivated to exert efforts when I was generating the ideas” and “I tried hard to generate ideas to
answer the question”. Participants rated their agreement with the two statements on a 7-point
Likert scale, ranging from “strongly disagree” to “strongly agree”. The two scales (a=0.67)
were averaged as the final measure of participants’ task effort (M=5.86, SD=1.06). The other
measure of participants’ effort was their time spent on the task (i.e. system recorded length of
time in minutes). This can be considered as a more objective behavioral indicator of effort
compared to the self-reported task effort. On average, participants spent 12.74 minutes on the
task (SD=7.69). Table 2 summarizes the descriptive statistics and correlations of the variables.

60
Table 2
Study 1 Descriptive Statistics and Correlations
Mean SD 1 2 3 4 5 6 7
1. SVO 0.61 .49
-
2. Task Structure 0.46 .50
-.01
-
3. Idea Productivity 6.23 3.59 .05 .22
**

-
4. Idea Originality 2.07 1.65
.02 .29
**
.38
**

-
5. Idea Usefulness 4.78 .98
.04 -.17
*
-.05 -.49
**

-
6. Task Effort 5.86 1.06 .15 .03 .37
**
.12 -.02
-
7. Time Spent 12.74 7.69
.18
*
.27
**
.33
**
.14 .02 .15
-
8. Time Spent (Residual) 0.00 7.28 .00 .00 .28
**
.06 .06 .12
.95
**

Note. *p<.05, **p< .01.

A critical premise of ANCOVA, however, is that the covariates included in the analysis
are not significantly correlated with the experimental conditions (Elashoff, 1969; Miller &
Chapman, 2001). This is because when significant correlation exists, the effects of the
experimental conditions can be hard to interpret. In some cases, the analysis may even remove
some part of the experimental effects and produce spurious results. However, according to Table
2, time spent on the task was significantly correlated with SVO and task condition at low to
medium levels. To remove the correlation between the covariate and the experimental
conditions, I regressed participants’ task time on SVO and task condition and computed the
residual of the regression. The residual represented the variance in task time that could not be
accounted for by SVO or task condition. As shown in Table 2, the procedure removed the
correlation between SVO, task condition and time spent, but did not significantly influence the
relation between time spent and other variables. The residual of time spent, rather than time
spent itself, was included as a covariate in the ANCOVA analysis. The materials used in Study 1
(i.e. the recruitment Ad on MTurk, manipulations and measures) can be found in Appendix F.

61
Procedure
Participants who accepted the HIT on MTurk was directed to Qualtrics to take part in the
online experiment, where they first signed an informed consent. They then took the Slider
Measure of SVO. After that, they read the task instruction, which manipulated the task structure.
Participants then proceeded to the idea generation task. After generating ideas, participants took
the post-experimental survey asking about their cooperative orientation, task effort and
demographic information. The participants were then debriefed and got the code they need to
claim their payments in MTurk.
Results
Manipulation Check
The manipulation of task structure was effective. Participants in the cooperative task
condition (M=5.84, SD=1.14) were significantly more cooperative-oriented than participants in
the competitive task condition (M=5.05, SD=1.61), t(174) = 3.80, p < .01. Participants in the
competitive task condition (M=5.44, SD=1.48) were significantly more competitive-oriented
than participants in the cooperative task condition (M=4.31, SD=1.73), t(174) = 4.64, p < .01.
Productivity
To test the hypotheses, an analysis of covariance (ANCOVA) test was conducted to
examine the effects of the experimental conditions on participant productivity. Participants’ self-
reported effort and time spent (residuals) on the task were included in the analysis as covariates.
The two covariates, although not strongly correlated with the experimental conditions,
significantly predicted participant productivity. This shows that noise in the task environment
had a significant impact on participants’ behavior in the online experiment. It is thus necessary to
control for these factors using the ANCOVA analysis.
62
Although prosocial participants (M=6.37, SD=3.57) appeared to be slightly more
productive than proselves (M=6.00, SD=3.62), the main effect of SVO was not significant,
F(1,170) = 0.02, p = .89. Therefore, H1, which predicted prosocials to be more creative than
proselves, was supported. However, the analysis revealed a significant main effect of task
condition, F(1,170) = 12.08, p < .01, h
p
2
= .07. Participants were more productive in the
competitive task condition (M=7.06, SD=3.94) than in the cooperative condition (M=5.52,
SD=3.10).
In support of H2, there was a significant interaction effect of SVO and task condition on
participant productivity, F(1, 170) = 4.18, p = .04, h
p
2
= .02. Simple effect analysis suggested
that whereas the productivity of prosocial participants did not differ significantly on cooperative
(M=6.96, SD=4.11) or competitive tasks (M=5.88, SD=3.00), t(105) = 1.57, p = .12, proself
participants were significantly more productive on competitive tasks (M=7.22, SD=3.71) than on
cooperative tasks (M=4.95, SD= 3.22), t(69) = 2.72, p < .01. The interaction effect is illustrated
in Figure 2.

Figure 2. The interaction effect of SVO and task structure on idea productivity
63
Originality
Another ANCOVA analysis was conducted to examine the effects of the experimental
conditions on participant originality, controlling for self-reported task effort and time spent
(residual) on the task. There was only a significant main effect of task condition, F(1, 170) =
15.17, p < .01, h
p
2
= .08. Participants on competitive tasks (M=2.59, SD=1.63) generated more
original ideas than those on cooperative tasks (M=1.58, SD=1.74). However, there were no main
interaction effect of SVO and task structure on idea originality F(1,170)=0.02, p = .89. H2 was
not supported in terms of idea originality. In addition, although ideas generated by prosocial
participants (M=2.10, SD=1.64) were slightly more original than those of proselves (M=2.03,
SD=1.69), the main effect of SVO was not significant, F(1,170) = 0.01, p = .92, rejecting H1.
Usefulness
An ANCOVA test was conducted to examine the effects of the experimental conditions
on idea usefulness, including self-reported effort and time spent (residual) on the task as
covariates. Again, while prosocial participants (M=4.81, SD=0.95) generated more useful ideas
than proselves (M=4.73, SD=1.03), the difference was not statistically significant,
F(1,170)=0.21, p = .65. Thus, so far, no support had been found for H1. The hypothesized
interaction effect of SVO and task condition (H2) on creativity was not significant either,
F(1,170) = 2.30, p = .13. Moreover, the significant main effect of task structure found on
productivity and originality was not significant for idea usefulness, F(1,170) = 3.54, p = .06. In
fact, rather contrarily, while competition enhanced participants’ productivity and originality,
individuals appeared to have produced more useful ideas in cooperation (M=4.93, SD=1.01) than
in competition (M=4.60, SD=0.92).
64
Discussion
In this chapter, I investigated how SVO impacts participant creativity in crowdsourced
idea generation. To reconcile the inconsistency in existing literature regarding the relation
between SVO and creativity, I proposed task structure as a moderating factor between SVO and
creativity. Since crowdsourced idea generation projects are generally open and provide relatively
low compensation for participants, it was predicted that prosocials should be more creative in
general because their cognition and behavior are driven by their intrinsic morality and interest.
The creativity of proself participants, however, was posited to be contingent upon task
conditions: whereas proselves might be less interested in cooperative idea generation, their
creativity could be significantly stimulated by competition.
Overall, the results provide mixed support for the postulated hypotheses. To begin with,
although prosocials were more productive and generated more original and useful ideas than
proselves, the difference in their creativity was not significant. Nevertheless, consistent with the
argument of Beersma and de Dreu (2005), competition led to higher productivity and originality
of ideas in crowdsourced idea generation. As hypothesized, the results reveal that the
productivity of proselves, but not prosocials, is dependent on task structure: while prosocials
were equally productive across task conditions, proself participants were significantly more
productive on competitive tasks than on cooperative tasks. Contrary to the triangle hypothesis,
which posits that prosocials are more responsive to contextual factors (Kelley & Stahelski,
1970), this finding shows that proselves’ creative thinking is more sensitive to task structures
than prosocials in crowdsourcing. Instead, the creativity of prosocials seems to be more
propelled by their intrinsic value orientation and interest, and remain relatively stable across
contexts.
65
In short, whereas Study 1 had elucidated the effects of SVO on participant creativity in
crowdsourcing by including task structure as a moderator, it did not include any investigations of
the social influence processes that may happen in crowdsourcing. As revealed in the pilot
studies, participants in crowdsourced idea generation are likely to interact with other participants
in the project, thereby gaining exposure to the ideas and behavior of others. The information they
obtain in interaction is likely to generate a significant influence on their creativity (Brown et al.,
1998; Coskun & Yilmaz, 2009; Larey & Paulus, 1999).
The lack of support for the positive effect of prosocial orientation in Study 1 could be due
to the omission of social influence in crowdsourcing. Since prosocial participants may be more
likely to search for integration and collaboration with others (de Dreu, Weingart, & Kwon, 2000;
Nauta, Dreu, & Vaart, 2002), they may be more likely to pay attention to others’ ideas and
produce more creative ideas by building novel combinations. The effect of SVO on creativity
therefore cannot be fully explicated without considering the influence of others in interaction.
In the following chapters, the impacts of others’ ideas and performance on creativity are
added to the discussion and investigation. Integrating insights of the MIP-G model, the theory of
SVO, and research on group creativity, Study 2 was proposed to examine how exposure to
others’ ideas may interact with SVO and task structure to affect participant creativity in
crowdsourcing. Its theoretical hypotheses and results are presented in the next chapter.

66
CHAPTER 6: SVO, TASK STRUCTURE, IDEA EXPOSURE AND CREATIVITY IN
CROWDSOURCED IDEA GENERATION
In crowdsourced idea generation, participants are often exposed to each other’s ideas in
the interaction process. Exposure to others’ ideas in brainstorming has been recognized by group
creativity researchers as a critical factor that stimulates social-cognitive process (Brown et al.,
1998; Dugosh et al., 2000; Nijstad et al., 2002). However, limited research has explored the
boundary conditions regarding the effects of idea exposure on creativity. Since SVO determines
the bias in information processing and task structure defines the interdependence level among
participants, these two factors can define whether and to what extent participants pay attention to
others’ ideas in interaction. Consequently, ideas that participants are exposed to may induce
differential degrees of influence on their creativity, depending on participants’ SVO and the task
structure. This chapter examines how SVO, task structure, and exposure to others’ ideas interact
to affect participant creativity in crowdsourcing.
Theory and Hypotheses
Exposure to others’ ideas cognitively stimulates creativity by reducing the cognitive load
in idea generation through a two-stage cognitive process (Nijstad et al., 2002). The first stage is
the knowledge activation stage. In this stage, ideas that one is exposed to serve as external
stimuli, providing cues to search for related information in one’s memory. In the second, the idea
generation stage, participants produce ideas drawing on information they probed from their
memory.
Therefore, ideas generated by participants tend to converge with ideas participants have
been exposed to in brainstorming (Boudreau & Lakhani, 2015; Brown et al., 1998; Coskun &
Yilmaz, 2009; Larey & Paulus, 1999; Yu & Nickerson, 2011; Ziegler et al., 2000). In a
67
crowdsourced design task, participants were more likely to create new designs based on others’
designs displayed to them (Yu & Nickerson, 2011). Disclosing contestant solutions in
crowdsourcing contests also leads to convergence in final solutions among the participants
(Boudreau & Lakhani, 2015).
However, this convergence can benefit participants’ creativity by providing memory
clues that reduce cognitive effort. Research has shown that interactive groups who exchanged
written ideas are more creative than groups that are required to discuss and exchange ideas orally
and then memorize the exchanged ideas (Paulus & Yang, 2000). However, when group members
generate ideas individually after the exchange of ideas, they are more productive than individuals
who did not exchange ideas. In crowdsourcing contests, although exposure to others’ solutions
leads to convergence in crowd solutions, it also helped contestants to improve their solutions
more efficiently (Boudreau & Lakhani, 2015). Participants in crowdsourced idea generation are
also shown to produce more creative ideas by integrating their ideas with others’ (Armisen &
Majchrzak, 2014; Yu & Nickerson, 2011, 2013).
The stimulating effect of others’ ideas on creativity can be more significant if the ideas
are diverse. In a simulation, Brown and colleagues (Brown et al., 1998) showed that while
participants tended to converge on fewer categories as they were exposed to each other’s
opinions, they were more creative if the ideas of others could prime them to generate ideas that
they could not have done while working independently. Later, Nijstad and colleagues (2002)
corroborated this finding by demonstrating that participants came up with more diverse ideas
after being exposed to heterogeneous ideas, but generated more homogenous ideas if exposed to
similar ideas. Research on crowdsourcing has also demonstrated that diversity of the crowd is
positively associated with its creativity (Armisen & Majchrzak, 2014; Boudreau, 2012; Jeppesen
68
& Lakhani, 2010; Malhotra & Majchrzak, 2014; Terwiesch & Xu, 2008), presumably because
exposure to a variety of opinions can inspire novel combinations and developments, which in
turn leads to higher creativity in crowd solutions.
However, while prior research on group creativity has extensively investigated the
stimulating effect of idea exposure in brainstorming, groups have been studied as if they
consisted of homogenous members who were impacted by others’ ideas in identical ways.
Despite prior research, the question of which group members are more prone to the influence of
others’ ideas remains unanswered.
In crowdsourced idea generation, participants’ SVO can affect their likelihood to be
affected by others’ ideas because it shapes the content of individual information processing
(Nijstad & de Dreu, 2012). Research has shown that in negotiations, prosocial participants use
more cooperative heuristics than proselves do (de Dreu & Boles, 1998; Olekalns & Smith, 1999),
and search for more integrative agreements that lead to better joint outcomes (de Dreu et al.,
2000; Nauta et al., 2002). Prosocial group members are more likely to cooperate if they can
communicate with others in tasks such as the prisoner’s dilemma (Liebrand, 1984). Research on
SVO and creativity also demonstrates that prosocial groups converge with imposed social norms
more than proself groups (Bechtoldt et al., 2010).
The trusting, cohesive and collaborative social environment cultivated by prosocial
orientation is also likely to enhance the stimulating effect of others’ ideas on creativity. In studies
of field teams, minority dissent and diversity in team member backgrounds are found to be
positively associated with team creativity and innovation only when the team members are
cooperative (de Dreu & West, 2001; Nijstad, Berger-Selman, & Dreu, 2014; Shin & Zhou, 2007;
Taggar, 2002). Research also suggests that while newcomers can be a source of creativity for
69
work teams, their creative potential can only be utilized in teams that share the same social
identity (Kane, Argote, & Levine, 2005).
Consequently, the stimulating effect of others’ ideas should be stronger for prosocials.
Since prosocials tend to process information with concerns for others and the collective (de Dreu
et al., 2011; Nijstad & de Dreu, 2012), they should be more attentive to others’ ideas in
interaction, and more likely to find common ground to build novel combinations. By contrast,
since proselves care more about their own benefits, they are perhaps less likely to pay attention
to others’ ideas or to look for integration between themselves and others’ opinions if no further
incentives are provided on the task. Therefore, I hypothesize that in general, prosocial
participants are more likely to be influenced by others’ ideas than proselves.

H3: The creativity of prosocials is more likely to be influenced by others’ ideas than
proselves in crowdsourced idea generation.

Furthermore, since prior literature and results from Study 1 suggest that the cognition and
behavior of prosocials in crowdsourcing are more driven by their internal values and interest
rather than the task conditions (Bem & Lord, 1979; Bogaert et al., 2008), the impact of others’
ideas on prosocials should also be consistent across different task structures. On the contrary,
proselves are sensitive to changes in task structure in crowdsourced brainstorming. Their
creativity tends to be stimulated by competition rather than cooperation because the former
offers additional private incentives to them.
In addition to cognitive stimulation, competition may also make proselves more attentive
to others’ ideas. Because proself individuals are more likely to process information that benefits
themselves, they can be more responsive to others if they know that the behavior of others can
70
impact their outcomes (van Kleef & van Lange, 2008). As proselves strive to outperform others
in competition, they should be more likely to cognitively process others’ ideas, which
consequently results in stronger influence of others’ ideas on their memory retrieval and idea
generation.

H4: The influence of others’ ideas on proselves’ creativity in crowdsourced idea
generation depends on the task structure, such that proselves are more likely to be impacted by
others’ ideas on competitive tasks than on cooperative tasks.

Study 2
Participants
The participant recruitment platform and criteria of Study 2 remained the same as Study 1
(i.e. MTurk workers with an approval rate above 95% and experience of more than 100 HITs). A
total of 287 participants successfully completed the task, each of which were paid $0.75. As
promised, one of the participants was randomly drawn to win a $15 lottery in each task
condition. 38.5% of the participants were male. 65.8% of the participants were between 25-44
years old and 92.7% had had some college education or higher.
Materials
To test the postulated hypotheses regarding the effects of SVO, task structure and idea
exposure on creativity in crowdsourcing, an online experiment was conducted through Qualtrics.
The experiment followed a 2 (SVO: prosocial vs proself) X 2 (Task structure: cooperative vs
competitive) X 2 (Idea Exposure: common vs original) between-subjects factorial design.
Consistent with Study 1, participants’ SVO was measured and classified applying the Slider
71
Measure. Of the 287 participants, 88 were classified as proselves, whereas 199 participants were
prosocials.
The manipulation of task structure was the same too. As a manipulation check of the task
condition, participants had to answer a multiple choice question regarding the task structure, and
their cooperative orientation was measured. Only participants who answered the multiple choice
question correctly were included in the study. For details of the manipulations, please refer to
Appendix F.
Ideas Exposure. Exposure to others’ ideas was manipulated by providing an example
idea to participants in the idea generation activity’s instructions. The two ideas used in the
manipulation were the actual ideas generated by prior participants in Study 1. They were equal in
terms of rated usefulness (6.5) but had distinctive originality scores. In the common idea
condition, participants were provided with the example idea: “offer bonuses for coherent and
intelligent answers”. The bonus idea was frequently generated by participants in Study 1 and
therefore scored 0 in terms of originality. In the original idea condition, participants were
provided with the example idea “improve the HIT's website access and provide instructions on
what to do if there's a glitch”. The idea was rarely generated by participants in Study 1 and thus
was highly original. Similar manipulation of idea originality has been applied by previous
research (Connolly et al., 1993; Dugosh & Paulus, 2005). To make sure that the manipulation
was effective, participants were asked to rephrase the example ideas in their own words. Only
the participants who correctly answered the question were included in the study.
Task. The idea generation task the participants performed was the same as in Study 1.
Participants generated ideas to answer the question: “What can MTurk requesters do to ensure
workers perform their HITs with good quality?”
72
Measures. Consistent with prior research (Amabile, 1983, 1996; Beersma & de Dreu,
2005; Diehl & Stroebe, 1987) and Study 1, the creativity of the participants in the idea
generation task was measured based on: 1) idea productivity, 2) originality of their ideas, and 3)
usefulness of their ideas.
Productivity. A total of 2035 ideas were generated by the participants. Productivity was
measured as the number of ideas generated by each participant. On average, each participant
produced 7.09 ideas (SD=4.78).
Originality. An idea that was generated by less than five percent of the participants in the
study (N=287) was given an originality score of 1, whereas others were given a score of 0.
Participants’ originality was the average originality score of all the ideas they generated,
multiplied by 7. The multiplication was conducted to make the originality score equivalent in
scale with the usefulness score. The average originality score of participants’ ideas in Study 2
was 2.98 (SD=1.84).
Usefulness. As with Study 1, the usefulness of the ideas was crowdsourced on MTurk.
689 valid evaluations were received. Each respondent received $0.50. After removing usefulness
ratings that were inconsistent with each other, on average, each idea was rated by 4.3 raters. An
idea’s usefulness was then calculated as the average score given by all the evaluators of the idea.
120 ideas (6.9%) were randomly selected from the full set of ideas and rated by the researcher.
The intra-class correlation (ICC) between the researcher’s evaluation and the mean usefulness
score given by the MTurk workers was 0.80, indicating good reliability of the MTurk workers’
ratings. Participants’ idea usefulness was then calculated as the mean usefulness score received
by all the ideas they generated. The average idea usefulness for participants was 4.92 (SD=0.73).
The materials and measures used in the evaluation HIT are provided in Appendix E.
73
Covariates. As in Study 1, two measures of task effort were included in the ANCOVA
analysis as covariates to control for the noise in the task environment that might impact
participants’ performance. Task effort was measured as self-reported effort and time spent on the
task, as in Study 1 (Appendix F). The two scales measuring the self-reported task effort
(a =0.82) were averaged (M=6.10, SD=0.97). On average, participants spent 19.18 minutes on
the task (SD=15.96). The descriptive statistics and correlation matrix of the variables are
presented in Table 3.
Table 3
Study 2 Descriptive Statistics and Correlations
Mean SD 1 2 3 4 5 6 7 8
1. SVO 0.69 .46
-
2. Task Structure 0.50 .50
-.06
-
3. Idea Exposure 0.47 .50 .00
-.08 -
4. Idea Productivity 7.09 4.78 .14
*
.06
.14
*
-
5. Idea Originality 2.98 1.84 .12
*
-.02 .23
**

.31
**
-
6. Idea Usefulness 4.92 .73
.06 -.02 -.10 -.05
-.35
**
-
7. Task Effort 6.10 .97
.14
*
-.00 -.02 .34
**
.12
*

-.07 -
8. Task Effort (Residual) 0.00 .96 .00 .00 -.02 .32
**
.10
-.08 .99
**
-
9. Time Spent 19.18 15.96 .09 .05 .05 .34
**
.23
**
-.00 .23
**

.22
**

Note. *p<.05, **p< .01.

As illustrated in Table 3, the self-reported task effort had a small but significant
correlation with SVO. Since ANCOVA analysis can produce spurious results if the covariates
are correlated with the experimental conditions (Elashoff, 1969; Miller & Chapman, 2001), the
self-reported task effort was regressed on SVO. The residual of the regression was saved as an
adjusted measure of task effort, with its correlation with SVO removed (Table 3). The residual of
the self-reported task effort, rather than the original measure, was included as a covariate in the
ANCOVA analysis.

74
Procedure
Participants who accepted the HIT on MTurk were directed to Qualtrics to participate in
the online experiment, where they would first sign an informed consent. They then took the
Slider SVO measure. After that, they read the task scenario, which manipulated the task
structure. Participants then proceeded to the idea generation task instructions and were randomly
assigned to one of the idea conditions. They then generated ideas for the posted question. After
generating ideas, participants took the post-experimental survey asking about their task effort,
cooperative orientation, and demographic information. The participants were then debriefed and
received the code they needed to claim their payments in MTurk.
Results
Manipulation Check
The manipulation of task structure was effective. Participants in the cooperative task
condition (M=5.81, SD=1.24) were significantly more cooperative-oriented than participants in
the competitive task condition (M=5.21, SD=1.64), t(285)=3.51, p < .01. Participants in the
competitive task condition (M=5.48, SD=1.36) were significantly more competitive-oriented
than participants in the cooperative task condition (M=4.61, SD=1.69), t(285)=4.82, p < .01.
Productivity
To test the hypotheses, an analysis of covariance (ANCOVA) test was conducted to
examine the effects of the experimental conditions on participant productivity, controlling for
two measures of task effort, i.e. self-reported task effort (residual) and time spent on task. Again,
the two covariates significantly predicted participant productivity in the experiment, supporting
the necessity to control for these two factors.
75
The analysis suggested a significant main effect of idea condition, F(1, 277)=5.65, p
= .02, h
p
2
= .02. Participants produced significantly more ideas after being exposed to highly
original ideas (M=7.80, SD=5.83) than after being exposed to common ideas (M=6.47, SD=3.52).
Although, consistent with Study 1, participants tended to be more productive under the
competitive condition (M=7.40, SD=5.25) than the cooperative condition (M=6.78, SD=4.26),
the difference was not statistically significant at the 0.05 level, F(1, 277)=3.59, p= .06. The
significant interaction effect of SVO and task condition revealed in Study 1 (H2), is not
statistically significant either, F(1, 277) =3.15, p = .08. However, consistent with Study 1, simple
effect analysis showed that proselves were more motivated by competition (M=7.33, SD=6.14)
than cooperation (M=4.60, SD=2.10), t(86)=2.69, p < .01, whereas prosocials did not
significantly differ in productivity across task conditions (M=7.62, SD=4.57 and M=7.43,
SD=4.77), t(197) = 0.29, p = .77.
The results also revealed a significant main effect of SVO, F(1,277)=5.31, p = .02, h
p
2

= .02. Prosocials (M=7.53, SD=4.66) produced more ideas than proselves (M=6.09, SD=4.92).
This finding provides support for H1 proposed in Chapter 5, but is contrary to the results of
Study 1. Since participants in Study 1 did not see any ideas produced by others, this
inconsistency could be due to the stimulating effect of idea exposure during brainstorming.
However, the hypothesized two-way interactions of SVO and idea originality (H3) was
not found, F(1, 277) = 0.41, p = .52. Prosocials did not seem to be more influenced by others’
original ideas than proselves. Yet in support of H4, there was a significant three-way interaction
effect of SVO, task structure, and idea exposure, F(1, 277) = 5.08, p = .03, h
p
2
= .02. Further
analysis revealed that, as expected, the productivity of prosocials were not significantly different
across tasks F(1, 193) = 0.04, p = .84 or idea conditions F(1, 193) = 2.47, p = .12. Nonetheless,
76
for proselves, there was a significant interaction effect of task structure and idea exposure,
F(1,82)=4.09, p < .05: while the productivity of proselves improved significantly after being
exposed to highly original ideas in the competitive condition (M=9.09, SD=8.13 vs M=4.56,
SD=2.26), t(39) = 2.29, p= .03, idea originality did not have a significant effect on proselves on
cooperative tasks (M=5.72, SD=2.75 vs M=4.64, SD=2.01), t(45) = 1.52, p = .13. The three-way
interaction effect is visually illustrated in Figure 3.

Figure 3. The three-way interaction effect of SVO, task structure and idea exposure on idea
productivity
Originality
Another ANCOVA analysis was conducted, examining the effects of the experimental
conditions on participant originality, which controlled for self-reported task effort (residual) and
time spent on task. This revealed a significant main effect of idea originality. Participants who
were exposed to the original idea (M=3.42, SD=1.82) produced significantly more original ideas
than those who saw the common idea (M=2.59, SD=1.77), F(1,277) = 15.33, p < .01, h
p
2
= .05.
77
However, no support was found for the hypothesized two-way interaction effect of SVO and idea
exposure (H3), F(1,277) = 1.08, p = .30, or the three-way interaction effect of SVO, task
structure and idea exposure (H4), F(1,277) = 0.11, p = .74.
Usefulness
An ANCOVA test was conducted to examine the effects of the experimental conditions
on the usefulness of ideas generated by the participants, controlling for effort (residual), and time
spent on the task. Somewhat contrary to the findings on productivity and originality, participants
who were exposed to the common idea (M=4.99, SD=0.67) produced more useful ideas than
those who saw the original idea (M=4.85, SD=0.79), despite the fact that the two ideas were
similar in usefulness. However, the effect was not statistically significant at the 0.05 level,
F(1,277)=3.06, p= .08. Neither the hypothesized two-way interaction effect (H3), F(1,277) =
0.11, p = .74, nor the three-way interaction effect (H4), F(1,277) = 0.09, p = .77, were supported
for idea usefulness, either.
Discussion
To summarize, the results of Study 2 demonstrate that exposure to others’ ideas in
crowdsourced idea generation can generate considerable influence on participant creativity.
Seeing original ideas promoted participants’ productivity and idea originality. However,
exposure to highly original ideas did not affect participants’ idea usefulness. Instead, it seemed to
be a bit detrimental to the usefulness of ideas generated by participants. It may be because
participants in the experiment were instructed to produce more non-redundant ideas rather than
useful ideas. Cognitively stimulated by an original idea, participants might have focused on
generating more different ideas but neglecting whether their ideas were appropriate.
78
The hypothesized interaction effects of SVO and idea originality were not supported.
Prosocials were not more likely to be stimulated by others’ ideas than proselves. Instead, the
productivity of prosocial participants was not significantly different across idea or task
conditions. Nonetheless, the creativity of proself participants was more likely to be influenced by
others’ original ideas on competitive tasks than on cooperative tasks.
The finding suggests that the creativity of prosocials is more driven by their internal
interest and morality rather than contextual factors. Proselves, by contrast, are more sensitive to
task structure in crowdsourced idea generation. This is in line with the results of Study 1. It could
be that proselves in competition tried to generate more ideas to win, so they built more on others’
ideas to reduce their cognitive load and to be more productive. Consequently, original ideas in
competitive tasks led to higher creativity of proselves, since they were more likely to inspire
proselves to retrieve a broader variety of information. On the contrary, common ideas did the
reverse and blocked the idea production of proselves as they paid attention to the ideas.
To conclude, following the idea exposure paradigm, Study 2 examined how the relation
between SVO and creativity is moderated by idea exposure and task structure in crowdsourced
idea generation. Nevertheless, participants in crowdsourcing may not only be exposed to others’
ideas, but also others’ performance. Previous research has suggested that others’ behavior (i.e.
level of cooperation, idea productivity) can generate significant impact on individuals in
interaction, depending on their SVO and on the task structure (Kelley & Stahelski, 1970;
Liebrand, Wilke, et al., 1986; Paulus & Dzindolet, 1993; Shepherd, et al., 1995). Following this
research, the next chapter postulates hypotheses regarding the interaction effect of SVO, task
structure, and information about others’ performance on creativity in crowdsourcing. Study 3,
designed to test the hypotheses, will be reported and discussed.
79
CHAPTER 7: SVO, TASK STRUCTURE, OTHER’S PERFORMANCE AND
CREATIVITY IN CROWDSOURCED IDEA GENERATION
Apart from others’ ideas, the performance of others in brainstorming has also been
established as a social factor that can influence participant creativity (Nijstad et al., 2002; Paulus
et al., 2013). Likewise, it may interact with SVO and task structure to influence individual
behavior in social settings (Kelley & Stahelski, 1970; Liebrand, Wilke, et al., 1986; van Lange,
1999). The last empirical study of the dissertation project, Study 3, explores how information
about others’ performance influences participant creativity in crowdsourced idea generation, as
well as whether and how it moderates the relation between SVO, task structure and creativity.
Theory and Hypotheses
Literature on group creativity has revealed that information about others’ performance
enhances social comparison and competition among participants in brainstorming and leads
participants to compare or match their productivity with others (Dugosh & Paulus, 2005; Paulus,
2000; Paulus & Dzindolet, 1993). Scholars have shown that when group members generate ideas
in interactive groups and know about others’ productivity, the number of ideas that they generate
is more similar to each other than that of nominal groups, in which group members do not
interact and therefore are not informed about others’ productivity (Camacho & Paulus, 1995;
Paulus & Dzindolet, 1993). This finding corresponds to the result of another study in SVO
research, which suggests that within group variation of cooperative behavior is smaller than its
between group variation, indicating convergence among group members when they are exposed
to each other’s behavior (McClintock et al., 1973).
Other evidence suggests that participants’ productivity in idea generation is significantly
boosted if they are exposed to a large amount of ideas (Dugosh & Paulus, 2005; Dugosh et al.,
80
2000; Nijstad et al., 2002; Paulus et al., 2013; Shepherd et al., 1995). Participants are also more
persistent in idea generation when more people participate and more ideas are generated (Nijstad
et al., 1999). Moreover, in both face-to-face and computer-mediated groups, when participants
are given a performance standard, their creativity is significantly improved (Paulus & Dzindolet,
1993; Paulus et al., 1996; Shepherd et al., 1995).
Although neglected by research on group creativity, evidence from SVO research
indicates that participants’ likelihood to be influenced by others’ performance is contingent on
their SVO. As discussed before, SVO biases the information processing of individuals in social
settings (Nijstad & de Dreu, 2012). The triangle hypothesis proposed by early research on SVO
posits that prosocial participants, compared to proselves, are more responsive to others’ behavior
in interdependent situations (Kelley & Stahelski, 1970). However, evidence from the two early
studies of the dissertation showed that the creativity demonstrated by prosocials seems to be
more guided by their intrinsic interest and internal sense of morality (Bem & Lord, 1979;
Liebrand, Jansen, et al., 1986). As a result, their creativity in crowdsourced idea generation is
less sensitive to social influence and variance in task structure.
Conversely, proselves care about maximizing their absolute or relative gain (de Dreu et
al., 2011; McClintock, 1972). They need additional external incentives to become interested in a
collective goal (Karau & Williams, 1993). They are thus more likely to be cognitively stimulated
and think more creatively in competition, as competition offers private rewards that boost their
attentiveness. Since exposure to others’ performance in interaction has been shown to stir a sense
of competition and comparison among group members in brainstorming (Dugosh & Paulus,
2005; Paulus, 2000; Paulus & Dzindolet, 1993), proselves may be more enthusiastic about
generating ideas to outperform others when they know about others’ performance level.
81
Accordingly, proself participants may be more likely to be influenced by others’ performance
than prosocials.

H5: The creativity of proselves is more likely to be influenced by others’ performance
than prosocials in crowdsourced idea generation.

Furthermore, task structure may moderate the impact of others’ performance on
proselves. Proself individuals’ sensitivity to task structure may inspire them to adapt their
thinking when they are informed about others’ performance in order to maximize their gain
(Kanagaretnam et al., 2009; Liebrand, Wilke, et al., 1986). Evidence suggests that, in behavioral
games, proselves are more likely to defect if they are informed that the majority of the group has
defected than when they are informed that the majority has cooperated, whereas prosocials do
not differ significantly across the two conditions (Liebrand, Wilke, et al., 1986). In public
dilemmas, proselves are also more likely to take more for themselves when the size of the
resource is uncertain compared to when the resource is certain, while prosocials do not differ
significantly between these two conditions (de Kwaadsteniet et al., 2006; Roch & Samuelson,
1997).
Therefore, proself participants should be less interested in generating ideas when they
know that others are generating ideas on cooperative tasks, given that they are unconcerned
about collective goals and tend to free-ride when they know that others are contributing (Karau
& Williams, 1993). On the contrary, in competitive tasks, proselves will be more interested in
producing ideas when they know how productive others are, in order to out-perform other
participants and earn higher personal rewards.

82
H6: The influence of others’ performance on proselves’ creativity in crowdsourced idea
generation depends on the task structure, such that proselves are more likely to be impacted by
others’ performance on competitive tasks than on cooperative tasks.

Study 3
Participants
The participant recruitment platform and criteria remains the same as the two prior
studies (i.e. MTurk workers with an approval rate above 95% and experience of more than 100
HITs). A total of 499 participants successfully completed the task, and each was paid $0.75. As
promised, one of the participants was randomly selected to win a $15 lottery in each task
condition. 47.3% of the participants were male. 63.9% of the participants were between 25-44
years old and 89.4 % had some college education or higher.
Materials
The experiment designed to test the above hypotheses followed a 2 (SVO: prosocial vs
proself) X 2 (Task structure: cooperative vs competitive) X 3 (Information about others’
performance: no/average/high) between-subjects factorial design. The experiment was conducted
online using Qualtrics. As in Study 1 and Study 2, participants’ SVO was measured and
classified based on the Slider Measure. Of the 499 participants, 175 were classified as proselves,
whereas 324 participants were prosocials.
The manipulation of task structure was consistent with the two prior studies as well. As a
manipulation check of the task condition, participants had to answer a multiple choice question
regarding the task structure. Their cooperative orientation was also measured. Only participants
who answered the multiple choice question correctly were included in the study. The complete
details regarding the manipulations can be found in Appendix F.
83
Information about others’ performance. Information about others’ performance was
manipulated by telling participants how many ideas others generated in previous idea generation
tasks. In the average productivity condition, participants were told that, on average, participants
generated 7 ideas in previous activities (Study 1 and Study 2). This was the actual average
productivity of individuals in prior studies. In the high productivity condition, participants were
told that, on average, participants generated 14 ideas in previous activities. In the no information
condition, nothing was told about individual productivity. To make sure that the participants
were paying attention to the information manipulation, participants in the two information
conditions were asked “how many ideas on average did workers generate in our prior
activities?”. Only the participants who responded with the number corresponding to their
information condition were included in the study.
Task. The idea generation task the participants performed was identical to the two prior
studies. Participants generated ideas to answer the question: “What can MTurk requesters do to
ensure workers perform their HITs with good quality?”
Measures. Consistent with Studies 1 and 2, the creativity of the participants in the idea
generation task was measured based on: 1) idea productivity, 2) originality of their ideas, and 3)
usefulness of their ideas (Amabile, 1983, 1996).
Productivity. A total of 3520 ideas were generated by the participants. Productivity was
measured as the number of ideas generated by each participant. On average, each participant
produced 7.05 ideas (SD=3.60).
Originality. An idea that was generated by less than five percent of the study participants
(N=499) was given an originality score of 1, whereas others were given a score of 0.
Participants’ originality was the average originality score of all the ideas they generated
84
multiplied by 7. Again, the multiplication was conducted to make the originality score equivalent
in scale with the idea usefulness score. The average originality score of the participants’ ideas in
Study 3 was 2.92 (SD=1.69).
Usefulness. Consistent with the two prior studies, the usefulness of the ideas was
crowdsourced. The evaluation HIT posted on MTurk received 960 valid responses. Each
participant was paid $0.50. After removing usefulness ratings that were inconsistent with each
other, on average each idea was rated by 4 raters. An idea’s usefulness was then calculated as the
average score given by all the evaluators of the idea. 100 ideas (3%) were randomly selected
from the full set of ideas and rated by the researcher. The intra-class correlation (ICC) between
the researcher’s evaluation and the mean usefulness score given by the MTurk workers was 0.81,
indicating good reliability of the MTurk workers’ ratings. Participants’ idea usefulness was then
calculated as the mean usefulness score received by all the ideas they generated. The average
idea usefulness for participants was 4.83 (SD=0.84). The materials and measures used in the
evaluation HIT are included in Appendix E.
Covariates. Consistent with Studies 1 and 2, two measures of task effort were included in
the ANCOVA analysis as covariates to control for the noise in the task environment. Again, task
effort was measured as self-reported effort and time spent on the task (Appendix F). The two
scales measuring the self-reported task effort (a =0.78) were averaged (M=5.98, SD=0.62). On
average, participants spent 14.57 minutes on the task (SD=13.51). The descriptive statistics and
correlation matrix of the variables are presented in Table 4.

85
Table 4
Study 3 Descriptive Statistics and Correlations

Mean SD 1 2 3 4 5 6 7 8
1. SVO 0.65 .48 -
2. Task Structure 0.52 .50
.01 -
3. Others’ Performance 0.76 .75 -.02 -.01 -
4. Idea Productivity 6.88 3.36
.05 -.13
**
.28
**
-
5. Idea Originality 2.89 1.69 .08 -.10
*
.03 .25
**
-
6. Idea Usefulness 4.83 .84
.02 .07 -.10
*
-.07 -.31
**
-
7. Task Effort 5.97 .61 .04 .02 .00 .25
**
.03 .05 -
8. Time Spent 14.54 13.62
.15
**
-.02 -.02 .05 .07 .10
*
.12
**
-
9. Time Spent (Residual) -.49 13.45 .00 -.02 -.02 .04 .06 .10
*
.11
*
0.99
**

Note. *p<.05, **p< .01.

As illustrated in Table 4, participants’ time spent on task had a small but significant
correlation with SVO. Since ANCOVA analysis can produce spurious results if the covariates
are correlated with the experimental conditions (Elashoff, 1969; Miller & Chapman, 2001),
participants’ task time was regressed on SVO. The residual of the regression was saved as an
adjusted measure of task time, with its correlation with SVO removed (Table 4). The residual of
participants’ time spent on task, rather than task time, was included as a covariate in the
ANCOVA analysis.
Procedure
Participants who accepted the HIT on MTurk were directed to Qualtrics to participate in
the online experiment, where they would first sign an informed consent. They then took the
Slider SVO measure. After that, they read the task scenario, which manipulated the task
structure. Participants then proceeded to the idea generation task instructions and were randomly
assigned into one of the information conditions. Next, they generated ideas for the posted
86
question. After generating ideas, participants took the post-experimental survey asking about
their effort and cooperative orientation on the task, as well as their demographic information. At
the end, the participants were debriefed and received the code they needed to claim their
payments in MTurk.
Results
Manipulation Check
The manipulation of task condition was effective. Participants in the cooperative task
condition (M=5.86, SD=0.94) were significantly more cooperative-oriented than participants in
the competitive task condition (M=5.06, SD=1.40), t(497)=7.57, p < .01. Participants in the
competitive task condition (M=5.24, SD=1.29) were significantly more competitive-oriented
than participants in the cooperative task condition (M=4.31, SD=1.59), t(497)=7.05, p < .01.
Productivity
To test the hypotheses, an analysis of covariance (ANCOVA) test was conducted to
examine the effects of the experimental conditions on participant productivity, including self-
reported task effort and time spent on the task (residual) as covariates. The analysis revealed a
significant main effect of task condition, F(1,485)=5.67, p = .02, h
p
2
= .01. Consistent with prior
studies, participants were more productive on the competitive task (M=7.53, SD=3.63) than on
the cooperative task (M=6.62, SD=3.52).
There was also a significant main effect of information about others’ performance, F(2,
485)= 22.21, p< .01, h
p
2
= .08. Post hoc comparisons using the Tukey HSD revealed that
information about others’ high performance improved participant productivity (M=8.75,
SD=3.92) significantly more than information about average productivity (M=7.31, SD=3.40),
and even more than no information (M=6.06, SD=3.29). Consistent with findings in Study 1, but
87
not Study 2, although prosocials (M=7.19, SD=3.62) were slightly more productive than
proselves (M=6.81, SD=3.55), SVO had no significant influence on participant productivity,
F(1 ,485)=0.41, p = .52. Thus, H1 is not supported by Study 3.
The hypothesized interaction effect of SVO and information about others’ performance
(H5), was not supported either, F(2,485) =0.98, p = .38. Moreover, no significant interaction
effect of SVO and task structure (H2) was found on participant productivity in Study 3,
F(1,485)=0.05, p = .82. The hypothesized three-way interaction (H6) was not significant, either,
F(2,485) =0.05, p = .95.
Originality
Another ANCOVA analysis examining the effects of the experimental conditions on
participant task effort and time (residual), revealed no main effects of SVO, F(1,485) = 0.82, p
= .37, task structure, F(1,485) = 2.44, p = .12, or information about others’ performance,
F(2,485)=0.77, p = .46, on idea originality. Again, although prosocial participants (M=3.01,
SD=1.68) produced more original ideas than proselves (M=2.74, SD=1.70), the difference was
not statistically significant.
In support of H5, there was a significant interaction effect of SVO and information about
others’ performance, F(2,485) = 3.03, p < .05, h
p
2
= .01. Simple effect analysis showed that
knowing that others are highly productive significantly increased the idea originality of proself
individuals, F(2,172)=3.10, p < .05. A post hoc comparison applying Tukey HSD showed that
participants’ idea originality in the high performance information condition (M=3.26, SD=1.53)
was significantly higher than that of the no information condition (M=2.44, SD=1.65), but not
significantly different from the average information condition (M=2.79, SD=1.80). The idea
originality between the average performance information condition and the no information
88
condition was not significantly different. For prosocials, there was no significant difference
among their idea originality across the information conditions, F(2,321)=0.66, p = .52. The
interaction effect is illustrated in Figure 4. Lastly, the three-way interaction effect of SVO, task
structure, and information about others’ performance, predicted by H6, was not supported for
idea originality F(2,485) = 0.03, p = .97.

Figure 4. The interaction effect of SVO and others’ performance on idea originality
Usefulness
An ANCOVA test was conducted to examine the effects of the experimental conditions
on participants’ idea usefulness, controlling for effort and time spent (residual) on the task. The
analysis suggested a significant main effect of information about others’ performance on the
usefulness of participants’ ideas, F(2,485) = 3.93, p = .02, h
p
2
= .02. Yet contrary to its effect on
productivity, a post hoc comparison (Tukey HSD) revealed that participants in the high
productivity information condition produced significantly less useful ideas (M=4.64, SD=0.82)
89
compared to those in the no information condition (M=4.89, SD=0.83), but not the average
productivity information condition (M=4.86, SD=0.84). There was no significant difference in
idea usefulness between participants in the no information condition or average productivity
information condition. While the ideas generated by prosocials (M=4.84, SD=0.78) were slightly
more useful than those of proselves (M=4.80, SD=0.95), the difference was again nonsignificant,
F(2,485) = 1.04, p = .31. H1 was again not supported.
Supporting H5, there was also a significant interaction effect of SVO and performance
information on idea usefulness, F(2,485) = 3.79, p = .02, h
p
2
= .02. For prosocials, although there
was a slight drop in idea usefulness when they were informed that others were highly productive,
information about others’ performance did not have a significant effect on their idea usefulness
overall, F(2,321) = 0.74, p = .48. However, the idea usefulness of proselves was significantly
impacted by information about others’ performance, F(2,172) = 5.58, p < .01. Contrary to the
pattern of participant productivity, a post hoc analysis (Tukey HSD) demonstrated that ideas
generated by proselves were significantly less useful when proselves were informed that others
were highly productive (M=4.43, SD=0.82), compared to when they had no information about
others’ performance (M=5.03, SD=0.87). Proselves’ idea usefulness was not significantly
different between the no information condition and the average productivity information
condition (M=4.76, SD=1.05), or between the average information condition and the high
productivity information condition. Figure 5 visually illustrates the interaction effect.
90

Figure 5. The interaction effect of SVO and others’ performance on idea usefulness
Discussion
In Study 3, I investigated the effect of SVO, task structure, and information about others’
performance on creativity in crowdsourced idea generation. In general, information about others’
performance did promote participants’ productivity. However, knowing that others were highly
productive significantly decreased the usefulness of their ideas. Perhaps because when
participants were focused on generating more ideas to match or beat others’ performance, they
neglected the quality of their ideas.
As hypothesized, the effect of information about others’ performance is also contingent
upon participants’ SVO. Overall, the originality and usefulness of prosocials’ ideas were not
significantly influenced by information about others’ performance. Nevertheless, if others’
performance was average, proself participants’ idea originality was not significantly influenced.
However, the presence of highly productive others seems to have enhanced proselves’
91
competitive orientation and boosted their idea originality, regardless of task conditions.
Moreover, as proselves paid attention to others’ performance, the usefulness of their ideas
suffered. They generated significantly fewer useful ideas when they knew about others’
performance.
These findings suggest that the influence of others’ performance on participants’
creativity is more salient for proselves than for prosocials. This is consistent with the results of
Studies 1 and 2, which have revealed that the creativity of proselves is more sensitive to task
contexts than prosocials. It seems that in addition to a competitive task structure, the
performance of others can also act as a contextual factor that stirs competition and stimulates the
creativity of proselves.
Finally, the hypothesized three way-interaction of SVO, task structure, and information
about others’ performance was not supported. The relationship between participant SVO, others’
performance, and creativity in crowdsourced idea generation was relatively consistent across
different task structures. Although this was anticipated for prosocials, it was rather unexpected
for proself participants. One possible explanation is that, on both competitive and cooperative
tasks, producing more ideas than others is more beneficial to participants’ private gains. On the
cooperative task, trying to produce more ideas as a group also increases participants’ individual
likelihood to win the bonus. Additionally, producing more ideas in competition brings direct
benefit to the participants as well. As a consequence, the best strategy for proself participants to
maximize their own gains after knowing how many ideas others produce is to always outperform
others.
Another alternative explanation is that, in the cooperative task, participants were
instructed to collaborate with people in their batch to increase their chances of winning the
92
bonus. Essentially, they were competing together with other groups on the cooperative task. The
task condition might have introduced inter-group competition. Prior research has suggested that
in situations in which inter-group competition exists, in-group cooperation may be promoted
(Bornstein & Ben-Yossef, 1994). This stimulating effect is particularly strong for proselves
(Carnevale & Probst, 1998). It could be that when proselves knew that others were highly
productive, their value maximizing orientation was equally aggravated by within-group
competition on competitive tasks and by inter-group competition on cooperative tasks.
Consequently, they produced significantly more ideas. When information about others’
performance was unavailable, however, the stimulating effect of inter-group competition did not
seem to be prominent, since proselves on cooperative tasks did not have a concrete target to
compete against.

93
CHAPTER 8: DISCUSSION AND CONCLUSION
Integrating the motivated information processing model in groups (MIP-G), the theory of
SVO, and research on group creativity, the current dissertation project investigated how SVO
impacts participant creativity in crowdsourced idea generation, and how task structure and
others’ ideas and performance in context moderate the relationship. Whereas research on
employee creativity has established that individual creativity in traditional organizations is a
result of personality and context (Hackman & Oldham, 1976; Oldham & Cummings, 1996;
Shalley, Zhou, & Oldham, 2004), much less is known about how personal disposition and task
contexts impact participant creativity in crowdsourced idea generation.
Through a series of online experiments, the dissertation demonstrates that compared to
prosocials, the creativity of proselves in crowdsourced idea generation is more sensitive to
contextual influences such as task structure and others’ performance. Overall, the creativity of
prosocials seems to be more driven by their intrinsic interest in the task and their internal sense
of morality, and is less likely to be influenced by others or the task environment. On the contrary,
whereas proself participants’ creativity in crowdsourced idea generation may be relatively lower
on cooperative tasks, they can be stimulated to be more creative by task contexts that induce
competition. Either a competitive task structure or highly productive others can be enough to
arouse the competitive orientation of proselves, and consequently influence their creativity in
crowdsourced idea generation. However, whereas competition in tasks seems to stimulate
productivity and originality, it can be damaging for the usefulness of ideas generated by
proselves. Moreover, the stimulating effect of others’ performance and competitive task structure
on proselves seems to be a cognitive one rather than a motivational one, since the task effort of
participants was controlled for in the studies.
94
The dissertation also shows that participants with either prosocial or proself orientations
can be creative in crowdsourced idea generation. Previous literature on SVO and creativity has
produced mixed evidence regarding whether prosocial or proself orientation is more beneficial
for creativity in group brainstorming. The proponents of prosocial orientation posit that
compared to proself participants, prosocial participants should be more creative in group idea
generation. They claim that prosocial value orientation may cultivate trust, cohesion, and
psychological safety in interdependent situations, which in turn encourages participants to
behave more creatively (Bechtoldt et al., 2012; Carnevale & Probst, 1998; de Dreu et al., 2011;
Taggar, 2002). Others argue that proself participants should be more creative since they are less
concerned about being different and tend to pursue distinct personal goals (Beersma & de Dreu,
2005; Goncalo, Flynn, & Kim, 2010; Goncalo & Staw, 2006; Janssen & Huang, 2008).
The findings of this dissertation, however, suggest that the dichotomy between prosocial
and proself orientation on creativity may be a false one, and that both social orientations can
positively influence individual creativity, depending on other contextual factors. Whereas
prosocial individuals can be creative in idea generation across task conditions, proselves can also
be inspired to generate original ideas when stimulated by task competition and information about
others’ ideas and performance.
The results also reveal that social influence occurs in crowdsourcing platforms when
participants are exposed to the ideas and performance of others. In general, the creativity of
prosocials and proselves is both highly responsive to others’ ideas and behavior in interaction.
To further examine the social influence of others’ ideas and behavior on participant creativity in
crowdsourced idea generation, I pooled the data collected from Studies 1, 2 and 3 together. I
coded the data in Study 1 and the no information condition in Study 3 as the control condition.
95
Data collected in Study 2 was classified as being in the idea exposure condition and the data of
Study 3 was coded as the performance information condition. Three one-way ANOVAs were
conducted to compare participants’ productivity, idea originality and idea usefulness across the
three conditions. Table 5 summarizes the results.
Table 5
Comparing Participants’ Creativity across Studies
Control
Study 1 & Study 3
Control Condition
Idea Exposure
Study 2
Others’
Performance
Study 3 Information
Conditions
ANOVA Result
M SD M SD M SD
Productivity 6.13 3.42 7.09 4.78 7.80 3.64 F(2, 959) = 15.14
**

Originality 2.50 1.75 2.98 1.84 2.96 1.64 F(2, 959) = 8.39
**

Usefulness 4.84 0.90 4.92 0.73 4.78 0.84 F(2, 959) = 2.02
Note. *p<.05, **p< .01.

The results reveal that others’ ideas and performance generate comparable influence on
participants’ productivity and originality: while participants were more productive and original
after being exposed to others’ ideas or being informed about others’ performance, their
productivity and originality were similar in the two information conditions. However, neither
others’ ideas nor their performance appeared to impact participants’ idea usefulness.
Theoretical Contributions
Overall, this dissertation project demonstrates that the creativity of participants in
crowdsourced idea generation is a function of individual disposition (SVO) and contextual
factors (task structure and social influence). Despite the variation in participant SVO in
crowdsourcing, little prior research has investigated its effect on participant creativity in
crowdsourced idea generation. Crowds have been mainly studied as if they consisted of members
with homogeneous social motivations (Jeppesen & Lakhani, 2010; Yu & Nickerson, 2011). This
96
dissertation project is among the first studies that examine the impact of SVO, task structure, and
social influence on creativity in crowdsourced idea generation.
Second, though previous research on crowdsourcing has investigated the effect of crowd
diversity, task structure, and communication on creativity, these factors were mostly studied
independently from each other (Boudreau & Lakhani, 2015; Jeppesen & Lakhani, 2010;
Terwiesch & Xu, 2008). The consequence of this is that the creative process in context, with
complex influence from participant SVO, task structure, and others’ behavior, remains under-
investigated. For example, whereas prior research on crowdsourcing has consistently
demonstrated that diversity promotes creativity in crowdsourcing (Armisen & Majchrzak, 2014;
Jeppesen & Lakhani, 2010), much remains unknown about the boundary conditions of
diversity’s effects. The current dissertation reveals that diversity in the crowd is more likely to
stimulate proselves in competition.
Moreover, this dissertation project builds on and integrates the motivated information
processing perspective of group creativity, the theory of SVO, and research on group creativity.
The motivated information processing perspective, or the MIP-G model, focuses on the
motivational aspect of group members but does not incorporate existing insights regarding how
group interaction may influence group creativity (de Dreu et al., 2011; Nijstad & de Dreu, 2012).
Conversely, prior research on group creativity, although extensively investigating how
interaction and exposure to others’ ideas impact creativity in brainstorming, generally neglects
the heterogeneity in participants’ social motivation in interdependent situations. The current
project complements both of these theoretical traditions, and helps to reconcile the previously
contradictory findings regarding SVO and creativity by clarifying the boundary conditions for
participants’ creativity.
97
In addition, whereas researchers of SVO have produced considerable insights regarding
the consequences of SVO on cooperative behavior in behavioral games and social dilemmas
(Kuhlman & Marshello, 1975b; McClintock & Liebrand, 1988; van Lange & Liebrand, 1991),
much less research has been devoted to exploring the impact of SVO on creativity (Bechtoldt et
al., 2010; de Dreu et al., 2011) in online social settings. This project, on the one hand, builds on
the fruitful effort of previous SVO research to elucidate the creative process in crowdsourced
idea generation. On the other hand, it develops the theory and research on SVO by explicating its
effect on creativity in crowdsourcing environments. Apart from validating some of the prior
findings on SVO in the crowdsourcing context, this dissertation has also produced significant
new insights regarding the interaction effects of SVO, task structure, and social influence in
online creativity.
Practical Implications
The findings of this dissertation project also suggest design implications for online
crowdsourcing platforms and crowdsourcing project sponsors. Since prosocial and proself
participants have been shown to be creative in crowdsourcing, idea generation projects can be
designed to appeal to both types of participants. This can be done by emphasizing both the social
benefits and private benefits of idea contribution. For example, if a crowdsourcing project calls
for ideas to improve local public transportation, it should state how a reformed public transit
system may benefit everyone by shifting people to a more environmentally friendly way of
commuting in order to appeal to prosocial participants (van Lange, van Vugt, Meertens, &
Ruiter, 1998). To attract proself participants, the organizers may also want to stress how
inefficient the current public transit is and how the participants’ contributions can make the
system more convenient for themselves. They can also offer extra incentives for contribution,
98
such as recognition of contributors on the website, fun games and pictures on the platform,
and/or monetary rewards.
The results also demonstrate that competitive and cooperative task structures in
crowdsourced idea generation both have pros and cons. While competition promotes productivity
and originality, it can be somewhat detrimental for the usefulness of ideas. Therefore, online
crowdsourcing projects may want to incorporate both cooperative and competitive elements into
the idea generation process. Many platforms are already doing this, by establishing a
collaborative ideation community while holding periodical idea generation contests (e.g. the
Laptime club discussed in Chapter 4).
Another, and perhaps better, approach can be designing the idea generation project as a
multi-phase activity and employing the most appropriate task structure for each phase. The first
phase focuses on attracting a large number of diverse ideas from the crowd. Competition should
be induced in this phase to encourage participants’ productivity and originality. In the next
phase, more original ideas are shortlisted from those generated in the first phase and then
provided to the participants in a cooperative task. Participants in this stage are instructed to
collaborate with each other to further develop the ideas they see. This stage can be repeated
several times depending on the project sponsors’ needs. Also, competition can be induced in
some of the repeated stages as well. For example, after one round of cooperative idea
development and integration, project managers can have a round of competitive idea
development to inspire more original combinations of others’ ideas, select the most innovative
combinations, and start a new phase of collaborative refining of the integrated solution.
Moreover, participants can be informed about their average performance in the
crowdsourced idea generation to maintain a collective standard and inspire creativity. However,
99
displaying top performers on a leaderboard does not seem to be a good idea based on the
findings of this dissertation, as highly productive others may induce too much competition and
distract participants, leading them to focus on the quantity rather than the applicability of their
ideas, especially for proself participants.
Limitations and Future Directions
This dissertation project categorized individual SVOs into two types. This categorization
was based on extensive prior research (Bechtoldt et al., 2010; Bogaert et al., 2008; van Lange &
Liebrand, 1991) and was supported by empirical SVO distributions revealed in the pilot studies
of the project. However, some researchers have argued that SVO is a continuous rather than
categorical variable, and that individuals may be concerned about themselves and others
simultaneously (de Dreu, 2006; de Dreu & Nauta, 2009; MacCrimmon & Messick, 1976;
McClintock et al., 1973). The Slider Measure (Murphy & Ackermann, 2014; Murphy et al.,
2011) also provides a way to measure SVO as a relatively continuous disposition, with higher
numbers indicating stronger prosocial orientation. Instead of classifying individuals into distinct
SVOs, as most prior research has done, future studies can research SVO as a continuous variable
and examine its relation to creative behavior in crowdsourcing.
Another limitation of the dissertation is that the distribution of participants’ SVOs is not
equal. Since the empirical studies recruited participants through a crowdsourcing platform
(MTurk) in which participants are typically rewarded with low monetary compensation, the
sample contained significantly more prosocially-oriented participants. Overall, of the 965 total
participants in the sample, 624 (64.67%) of them were classified as prosocials. To include
equivalent numbers of proselves and prosocials, and to maximize the statistical power, future
research can measure participants’ SVO before assigning them to experimental conditions.
100
In this study, when originality of ideas was improved by competition, idea usefulness
decreased. It may appear that the two qualities of ideas – originality and usefulness, are mutually
exclusive, or that growth in idea originality may result in decline in usefulness. However, the
pattern revealed in the studies could be an artifact of the task incentive and instructions
employed. Since the task instructions mainly focused on the quantity, rather than quality of the
ideas, participants might have been directed to generate more diverse, rather than useful ideas.
Prior research on creativity has used originality and usefulness as two orthogonal dimensions to
measure creativity, and has found evidence supporting the integration of both in creative
solutions (Amabile, 1983, 1996). Future studies can further explore the conditions and task
designs that can encourage idea originality and usefulness simultaneously.
Moreover, participants in these studies did not really participate in interactive
brainstorming as they would in most crowdsourced idea generation projects. In the current
studies, I decided to not allow participants to interact in order to detangle the effect of social
influence (i.e. exposure to others’ ideas and information about others’ performance) on
creativity, because interactive groups may generate other factors (i.e. trust, conflict) that would
interfere with the mechanism of the two variables of interest. However, future studies can
analyze creative behavior of individuals in interactive settings to examine if the same effects
persist and investigate other boundary conditions that may moderate the social influence of
others.
In addition, Study 3 did not identify any difference in participants’ reaction to others’
performance across tasks. However, this could be due to the manipulation of task structure
employed by the experiments. In the experimental studies, the cooperative task instructed
participants to collaborate with their peers in the same batch to generate more ideas. It might
101
have stirred inter-group competition among different batches, which in turn aroused the
competitive orientation of proselves and inspired them to be productive on cooperative tasks as
well. Future studies should manipulate task structure in such a way that within-group cooperation
is imposed without implying inter-group competition, and observe whether proself individuals
behave differently in the cooperation-only condition. Researchers can also compare the influence
of the three types of task structure, namely, tasks with within-group competition, within-group
cooperation, and inter-group competition on participant creativity in crowdsourcing. Since inter-
group competition is likely to trigger higher levels of group identification (Tajfel & Turner,
1979), it may result in more in-group cooperation and out-group competition, but only for proself
participants. Prior research investigating SVO and inter-group competition showed that although
prosocial participants contributed more to their own groups in intergroup competitions, they did
not behave in a way that would hurt the outgroups (de Dreu, 2010).
The manipulation of information about others’ performance in Study 3 employed three
conditions: no information, information about average performance, and information about high
performance. However, previous literature on group creativity has shown that participants tend to
converge with others’ performance in interaction, and that they are more likely to match their
productivity with lower performing members in groups (Camacho & Paulus, 1995; Paulus &
Dzindolet, 1993). Future research should further investigate this possibility by adding an
information condition in which participants are informed about a below average performance
level. It is likely that informing the participants about a performance standard that is lower than
the actual average performance may decrease participant performance due to downward
matching. This detrimental effect of low performance information may be more significant for
proself participants and in cooperation.
102
Concluding Remarks
The development of information technology has enabled organizations to crowdsource
creative ideas and solutions publically. Crowdsourced idea generation projects initiated online
are usually large in scale, with hundreds or even thousands of participants working on the same
task. Participants often volunteer to contribute without monetary rewards, and collaborate with
each other via communication technologies online. While previous work on employee creativity
has established that individual creativity is a function of both personal and contextual factors,
much remains unknown about how disposition and situation impact participant creativity in
crowdsourcing.
Acknowledging the heterogeneity in participant SVO in crowdsourced idea generation,
the current dissertation integrates the motivated information processing model in groups (MIP-
G), the theory of social value orientation (SVO), and research on group creativity. It investigates
participant creativity in crowdsourced idea generation as a consequence of participants’
dispositional SVO, task structure (i.e. cooperation or competition), and social influence (i.e.
exposure to others’ ideas and information about others’ performance). Through a series of online
experiments, the dissertation project reveals that the creativity of proselves in crowdsourced idea
generation is more sensitive to contextual influences such as task structure and others’
performance. On the contrary, prosocials’ creativity seems to be mostly driven by their intrinsic
interest and internal sense of morality, and is less susceptible to influence from others or the task
environment. This dissertation is among the first studies to examine the relation between SVO,
task context (i.e. task structure and social influence), and participant creativity in crowdsourcing.

103
References
Adler, P. S. (2001). Market, hierarchy, and trust: The knowledge economy and the future of
capitalism. Organization Science, 12(2), 215–234.
Afuah, A., & Tucci, C. L. (2012). Crowdsourcing as a solution to distant search. Academy of
Management Review, 37(3), 355–375.
Amabile, T. M. (1983). The social psychology of creativity: A componential conceptualization.
Journal of Personality and Social Psychology, 45(2), 357–376.
http://dx.doi.org.libproxy.usc.edu/10.1037/0022-3514.45.2.357
Amabile, T. M. (1996). Creativity in context: Update to the social psychology of creativity.
Boulder, Colo: Westview Press.
Armisen, A., & Majchrzak, A. (2014). Tapping the innovative business potential of innovation
contests (SSRN Scholarly Paper No. ID 2535944). Rochester, NY: Social Science
Research Network. Retrieved from http://papers.ssrn.com/abstract=2535944
Balliet, D., Parks, C., & Joireman, J. (2009). Social value orientation and cooperation in social
dilemmas: A meta-analysis. Group Processes & Intergroup Relations, 12(4), 533–547.
https://doi.org/10.1177/1368430209105040
Baruah, J., & Paulus, P. B. (2009). Enhancing group creativity: The search for synergy. In
Creativity in Groups (Vol. 12, pp. 29–56). Emerald Group Publishing Limited. Retrieved
from http://www.emeraldinsight.com.libproxy1.usc.edu/doi/full/10.1108/S1534-
0856%282009%290000012005
Bateman, P. J., Gray, P. H., & Butler, B. S. (2010). The impact of community commitment on
participation in online communities. Information Systems Research, 22(4), 841–854.
https://doi.org/10.1287/isre.1090.0265
104
Bechtoldt, M. N., Choi, H.-S., & Nijstad, B. A. (2012). Individuals in mind, mates by heart:
Individualistic self-construal and collective value orientation as predictors of group
creativity. Journal of Experimental Social Psychology, 48(4), 838–844.
https://doi.org/10.1016/j.jesp.2012.02.014
Bechtoldt, M. N., de Dreu, C. K. W., Nijstad, B. A., & Choi, H.-S. (2010). Motivated
information processing, social tuning, and group creativity. Journal of Personality and
Social Psychology, 99(4), 622.
Beersma, B., & de Dreu, C. K. W. (2005). Conflict’s consequences: Effects of social motives on
postnegotiation creative and convergent group functioning and performance. Journal of
Personality and Social Psychology, 89(3), 358–374. https://doi.org/10.1037/0022-
3514.89.3.358
Bekkers, R. (2004). Stability, reliability and validity of social value orientation (SSRN Scholarly
Paper No. ID 2274560). Rochester, NY: Social Science Research Network. Retrieved
from http://papers.ssrn.com/abstract=2274560
Bem, D. J., & Lord, C. G. (1979). Template matching: A proposal for probing the ecological
validity of experimental settings in social psychology. Journal of Personality and Social
Psychology, 37(6), 833-846.
Benkler, Y. (2002). Coase’s Penguin, or, Linux and the nature of the firm. The Yale Law
Journal, 112(3), 369–446. https://doi.org/10.2307/1562247
Benkler, Y. (2006). The wealth of networks: How social production transforms markets and
freedom. New Haven, CT: Yale University Press.
105
Berinsky, A. J., Huber, G. A., & Lenz, G. S. (2012). Evaluating online labor markets for
experimental research: Amazon.com’s Mechanical Turk. Political Analysis, 20(3), 351–
368. https://doi.org/10.1093/pan/mpr057
Bogaert, S., Boone, C., & Declerck, C. (2008). Social value orientation and cooperation in social
dilemmas: A review and conceptual model. British Journal of Social Psychology, 47(3),
453–480. https://doi.org/10.1348/014466607X244970
Bornstein, G., & Ben-Yossef, M. (1994). Cooperation in intergroup and single-group social
dilemmas. Journal of Experimental Social Psychology, 30(1), 52–67.
https://doi.org/10.1006/jesp.1994.1003
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. J., Lacetera, N., & Lakhani, K. R. (2011). Incentives and problem uncertainty in
innovation contests: An empirical analysis. Management Science, 57(5), 843–863.
Boudreau, K. J., & Lakhani, K. R. (2009). How to manage outside innovation. MIT Sloan
Management Review, 50(4), 69–76.
Boudreau, K. J., & Lakhani, K. R. (2013). Using the crowd as an innovation partner. Harvard
Business Review, (April), 61–69.
Boudreau, K. J., & Lakhani, K. R. (2015). “Open” disclosure of innovations, incentives and
follow-on reuse: Theory on processes of cumulative innovation and a field experiment in
computational biology. Research Policy, 44(1), 4–19.
https://doi.org/10.1016/j.respol.2014.08.001
106
Boudreau, K. J., Lakhani, K. R., & Menietti, M. (2014). Performance responses to competition
across skill-levels in rank order tournaments: Field evidence and implications for
tournament design. Working Paper.
Brabham, D. C. (2008). 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. https://doi.org/10.1177/1354856507084420
Brown, V., Tumeo, M., Larey, T. S., & Paulus, P. B. (1998). Modeling cognitive interactions
during group brainstorming. Small Group Research, 29(4), 495–526.
https://doi.org/10.1177/1046496498294005
Brucks, W. M., & van Lange, P. A. M. (2007). When prosocials act like proselfs in a commons
dilemma. Personality and Social Psychology Bulletin, 33(5), 750–758.
https://doi.org/10.1177/0146167206298569
Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechanical Turk: A new source
of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6(1), 3–5.
https://doi.org/10.1177/1745691610393980
Butler, B. S., Sproull, L., Kiesler, S., & Kraut, R. E. (2007). Community effort in online groups:
Who does the work and why? In S. Weisband (Ed.), Leadership at a distance: Research
in technologically-supported work (pp. 171–194). Mahwah, NJ: Lawrence Erlbaum.
Camacho, L. M., & Paulus, P. B. (1995). The role of social anxiousness in group brainstorming.
Journal of Personality and Social Psychology, 68(6), 1071.
Carnevale, P. J., & Probst, T. M. (1998). Social values and social conflict in creative problem
solving and categorization. Journal of Personality and Social Psychology, 74(5), 1300–
1309. http://dx.doi.org.libproxy2.usc.edu/10.1037/0022-3514.74.5.1300
107
Connolly, T., Routhieaux, R. L., & Schneider, S. K. (1993). On the effectiveness of group
brainstorming: test of one underlying cognitive mechanism. Small Group Research,
24(4), 490–503.
Cornelissen, G., Dewitte, S., & Warlop, L. (2011). Are social value orientations expressed
automatically? Decision making in the dictator game. Personality and Social Psychology
Bulletin, 37(8), 1080–1090. https://doi.org/10.1177/0146167211405996
Coskun, H., & Yilmaz, O. (2009). A new dynamical model of brainstorming: Linear, nonlinear,
continuous (simultaneous) and impulsive (sequential) cases. Journal of Mathematical
Psychology, 53(4), 253–264. https://doi.org/10.1016/j.jmp.2009.03.003
de Cremer, D., & van Lange, P. A. M. (2001). Why prosocials exhibit greater cooperation than
proselfs: The roles of social responsibility and reciprocity. European Journal of
Personality, 15(S1), S5–S18. https://doi.org/10.1002/per.418
de Dreu, C. K. W. (2006). Rational self-interest and other orientation in organizational behavior:
A critical appraisal and extension of Meglino and Korsgaard (2004). Journal of Applied
Psychology, 91(6), 1245.
de Dreu, C. K. W. (2010). Social value orientation moderates ingroup love but not outgroup hate
in competitive intergroup conflict. Group Processes & Intergroup Relations, 13(6), 701–
713. https://doi.org/10.1177/1368430210377332
de Dreu, C. K. W., & Boles, T. L. (1998). Share and share alike or winner take all? The influence
of social value orientation upon choice and recall of negotiation heuristics.
Organizational Behavior and Human Decision Processes, 76(3), 253–276.
https://doi.org/10.1006/obhd.1998.2806
108
de Dreu, C. K. W., & Nauta, A. (2009). Self-interest and other-orientation in organizational
behavior: Implications for job performance, prosocial behavior, and personal initiative.
Journal of Applied Psychology, 94(4), 913.
de Dreu, C. K. W., & Nijstad, B. A. (2008). Mental set and creative thought in social conflict:
threat rigidity versus motivated focus. Journal of Personality and Social Psychology,
95(3), 648–661. http://dx.doi.org.libproxy2.usc.edu/10.1037/0022-3514.95.3.648
de Dreu, C. K. W., Nijstad, B. A., Bechtoldt, M. N., & Baas, M. (2011). Group creativity and
innovation: A motivated information processing perspective. Psychology of Aesthetics,
Creativity, and the Arts, 5(1), 81–89. https://doi.org/10.1037/a0017986
de Dreu, C. K. W., Nijstad, B. A., & Knippenberg, D. van. (2008). Motivated information
processing in group judgment and decision making. Personality and Social Psychology
Review, 12(1), 22–49. https://doi.org/10.1177/1088868307304092
de Dreu, C. K. W., Weingart, L. R., & Kwon, S. (2000). Influence of social motives on
integrative negotiation: A meta-analytic review and test of two theories. Journal of
Personality and Social Psychology, 78(5), 889–905. https://doi.org/10.1037/0022-
3514.78.5.889
de Dreu, C. K. W., & West, M. A. (2001). Minority dissent and team innovation: The importance
of participation in decision making. Journal of Applied Psychology, 86(6), 1191–1201.
de Kwaadsteniet, E. W., van Dijk, E., Wit, A., & de Cremer, D. (2006). Social dilemmas as
strong versus weak situations: Social value orientations and tacit coordination under
resource size uncertainty. Journal of Experimental Social Psychology, 42(4), 509–516.
https://doi.org/10.1016/j.jesp.2005.06.004
109
Dennis, A. R., & Valacich, J. S. (1993). Computer brainstorms: More heads are better than one.
Journal of Applied Psychology, 78(4), 531–537.
Dholakia, U. M., Bagozzi, R. P., & Pearo, L. K. (2004). A social influence model of consumer
participation in network- and small-group-based virtual communities. International
Journal of Research in Marketing, 21(3), 241–263.
https://doi.org/10.1016/j.ijresmar.2003.12.004
Diehl, M., & Stroebe, W. (1987). Productivity loss in brainstorming groups: Toward the solution
of a riddle. Journal of Personality and Social Psychology, 53(3), 497–509.
http://dx.doi.org /10.1037/0022-3514.53.3.497
Dugosh, K. L., & Paulus, P. B. (2005). Cognitive and social comparison processes in
brainstorming. Journal of Experimental Social Psychology, 41(3), 313–320.
https://doi.org/10.1016/j.jesp.2004.05.009
Dugosh, K. L., Paulus, P. B., Roland, E. J., & Yang, H.-C. (2000). Cognitive stimulation in
brainstorming. Journal of Personality and Social Psychology, 79(5), 722–735.
http://dx.doi.org/10.1037/0022-3514.79.5.722
Eek, D., & Gärling, T. (2006). Prosocials prefer equal outcomes to maximizing joint outcomes.
British Journal of Social Psychology, 45(2), 321–337.
https://doi.org/10.1348/014466605X52290
Eek, D., & Gärling, T. (2008). A new look at the theory of social value orientations: Prosocials
neither maximize joint outcome nor minimize outcome differences but prefer equal
outcomes. In New Issues and Paradigms in Research on Social Dilemmas (pp. 10–26).
Springer, Boston, MA. https://doi.org/10.1007/978-0-387-72596-3_2
110
Eisenberg, N., Guthrie, I. K., Cumberland, A., Murphy, B. C., Shepard, S. A., Zhou, Q., & Carlo,
G. (2002). Prosocial development in early adulthood: a longitudinal study. Journal of
Personality and Social Psychology, 82(6), 993–1006.
Eisenberger, R., Kuhlman, D. M., & Cotterell, N. (1992). Effects of social values, effort training,
and goal structure on task persistence. Journal of Research in Personality, 26(3), 258–
272. https://doi.org/10.1016/0092-6566(92)90043-4
Elashoff, J. D. (1969). Analysis of covariance: A delicate instrument. American Educational
Research Journal, 6(3), 383–401. https://doi.org/10.3102/00028312006003383
Faraj, S., & Johnson, S. L. (2011). Network exchange patterns in online communities.
Organization Science, 22(6), 1464–1480. https://doi.org/10.1287/orsc.1100.0600
Fuge, M., Tee, K., Agogino, A., & Maton, N. (2014). Analysis of collaborative design networks:
A case study of OpenIDEO. Journal of Computing and Information Science in
Engineering, 14(2), 021009–021009. https://doi.org/10.1115/1.4026510
Galletta, D. F., Marks, P. V., Polak, P., & McCoy, S. (2003). What leads us to share valuable
knowledge? An experimental study of the effects of managerial control, group
identification, and social value orientation on knowledge-sharing behavior. In
Proceedings of the 36th Annual Hawaii International Conference on System Sciences.
Washington, DC: IEEE. https://doi.org/10.1109/HICSS.2003.1174271
Gallupe, R. B., Bastianutti, L. M., & Cooper, W. H. (1991). Unblocking brainstorms. Journal of
Applied Psychology, 76(1), 137–142.
Gallupe, R. B., Dennis, A. R., Cooper, W. H., Valacich, J. S., Bastianutti, L. M., & Nunamaker,
J. F., Jr. (1992). Electronic brainstorming and group size. The Academy of Management
Journal, 35(2), 350–369. https://doi.org/10.2307/256377
111
George, J. M. (2007). Creativity in organizations. The Academy of Management Annals, 1(1),
439–477. https://doi.org/10.1080/078559814
Goncalo, J. A., & Duguid, M. M. (2012). Follow the crowd in a new direction: When conformity
pressure facilitates group creativity (and when it does not). Organizational Behavior and
Human Decision Processes, 118(1), 14–23. https://doi.org/10.1016/j.obhdp.2011.12.004
Goncalo, J. A., Flynn, F. J., & Kim, S. H. (2010). Are two narcissists better than one? The link
between narcissism, perceived creativity, and creative performance. Personality and
Social Psychology Bulletin, 36(11), 1484–1495.
https://doi.org/10.1177/0146167210385109
Goncalo, J. A., & Kim, S. H. (2010). Distributive justice beliefs and group idea generation: Does
a belief in equity facilitate productivity? Journal of Experimental Social Psychology,
46(5), 836–840. https://doi.org/10.1016/j.jesp.2010.03.007
Goncalo, J. A., & Staw, B. M. (2006). Individualism–collectivism and group creativity.
Organizational Behavior and Human Decision Processes, 100(1), 96–109.
https://doi.org/10.1016/j.obhdp.2005.11.003
Grant, A. M., & Berry, J. W. (2011). The necessity of others is the mother of invention: Intrinsic
and prosocial motivations, perspective taking, and creativity. Academy of Management
Journal, 54(1), 73–96. https://doi.org/10.5465/AMJ.2011.59215085
Green, M., Seepersad, C. C., & Hölttä-Otto, K. (2014). Crowd-sourcing the evaluation of
creativity in conceptual design: A pilot study. In Proceedings of ASME 2014
International Design Engineering Technical Conferences and Computers and
Information in Engineering Conference, (pp.V007T07A016). Buffalo, NY: ASME.
https://doi.org/10.1115/DETC2014-34434
112
Hackman, J. R., & Oldham, G. R. (1976). Motivation through the design of work: test of a
theory. Organizational Behavior and Human Performance, 16(2), 250–279.
https://doi.org/10.1016/0030-5073(76)90016-7
Hardin, R. (1971). Collective action as an agreeable n-prisoners’ dilemma. Behavioral Science,
16(5), 472–481. https://doi.org/10.1002/bs.3830160507
Heilman, M., & Smith, N. (2010). Rating computer-generated questions with Mechanical Turk.
Presented at the 11th Annual Conference of the North American Chapter of the
Association for Computational Linguistics, Los Angeles, CA.
Hilbig, B. E., & Zettler, I. (2009). Pillars of cooperation: Honesty–humility, social value
orientations, and economic behavior. Journal of Research in Personality, 43(3), 516–519.
https://doi.org/10.1016/j.jrp.2009.01.003
Howe, J. (2008). Crowdsourcing: Why the power of the crowd is driving the future of business.
New York, NY: Crown Business.
Hülsheger, U. R., Anderson, N., & Salgado, J. F. (2009). Team-level predictors of innovation at
work: a comprehensive meta-analysis spanning three decades of research. The Journal of
Applied Psychology, 94(5), 1128–1145. https://doi.org/10.1037/a0015978
Iedema, J., & Poppe, M. (1994). Effects of social value orientation on expecting and learning
others’ orientations. European Journal of Social Psychology, 24(5), 565–579.
https://doi.org/10.1002/ejsp.2420240504
Janis, I. L. (1982). Groupthink: Psychological studies of policy decisions and fiascoes (2nd ed.).
Boston, MA: Cengage Learning.
113
Janssen, O., & Huang, X. (2008). Us and me: Team identification and individual differentiation
as complementary drivers of team members’ citizenship and creative behaviors. Journal
of Management, 34(1), 69–88. https://doi.org/10.1177/0149206307309263
Jeppesen, L. B., & Lakhani, K. R. (2010). Marginality and problem-solving effectiveness in
broadcast search. Organization Science, 21(5), 1016–1033.
https://doi.org/10.1287/orsc.1090.0491
Kanagaretnam, K., Mestelman, S., Nainar, K., & Shehata, M. (2009). The impact of social value
orientation and risk attitudes on trust and reciprocity. Journal of Economic Psychology,
30(3), 368–380. https://doi.org/10.1016/j.joep.2008.12.003
Kane, A. A., Argote, L., & Levine, J. M. (2005). Knowledge transfer between groups via
personnel rotation: Effects of social identity and knowledge quality. Organizational
Behavior and Human Decision Processes, 96(1), 56–71.
https://doi.org/10.1016/j.obhdp.2004.09.002
Karau, S. J., & Williams, K. D. (1993). Social loafing: A meta-analytic review and theoretical
integration. Journal of Personality and Social Psychology, 65(4), 681–706.
https://doi.org/10.1037/0022-3514.65.4.681
Kelley, H. H., & Stahelski, A. J. (1970). Social interaction basis of cooperators’ and competitors’
beliefs about others. Journal of Personality and Social Psychology, 16(1), 66–91.
http://dx.doi.org.libproxy2.usc.edu/10.1037/h0029849
Kittur, A., Chi, E. H., & Suh, B. (2008). Crowdsourcing user studies with Mechanical Turk. In
Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp.
453–456). New York, NY, USA: ACM. https://doi.org/10.1145/1357054.1357127
114
Kittur, A., Smus, B., Khamkar, S., & Kraut, R. E. (2011). CrowdForge: Crowdsourcing complex
work. In Proceedings of the 24th Annual ACM Symposium on User Interface Software
and Technology (pp. 43–52). New York, NY: ACM.
https://doi.org/10.1145/2047196.2047202
Knight, G. P., & Dubro, A. (1984). Cooperative, competitive, and individualistic social values:
An individualized regression and clustering approach. Journal of Personality and Social
Psychology, 46(1), 98–105.
Kuhlman, D. M., & Marshello, A. F. (1975a). Individual differences in game motivation as
moderators of preprogrammed strategy effects in prisoner’s dilemma. Journal of
Personality and Social Psychology, 32(5), 922–931.
Kuhlman, D. M., & Marshello, A. F. (1975b). Individual differences in the game motives of
own, relative, and joint gain. Journal of Research in Personality, 9(3), 240–251.
https://doi.org/10.1016/0092-6566(75)90020-3
Kuhlman, D. M., & Wimberley, D. L. (1976). Expectations of choice behavior held by
cooperators, competitors, and individualists across four classes of experimental games.
Journal of Personality and Social Psychology, 34(1), 69–81.
http://dx.doi.org.libproxy1.usc.edu/10.1037/0022-3514.34.1.69
Lakhani, K. R., & von Hippel, E. (2003). How open source software works: “free” user-to-user
assistance. Research Policy, 32(6), 923–943. https://doi.org/10.1016/S0048-
7333(02)00095-1
Lakhani, K. R., & Wolf, R. (2005). Why hackers do what they do: Understanding motivation and
effort in free/open source software projects. In Perspectives on free and open source
software (pp. 3–22). Cambridge, MA: MIT Press.
115
Lampel, J., & Bhalla, A. (2007). The role of status seeking in online communities: Giving the
gift of experience. Journal of Computer-Mediated Communication, 12(2), 434–455.
https://doi.org/10.1111/j.1083-6101.2007.00332.x
Larey, T. S., & Paulus, P. B. (1999). Group preference and convergent tendencies in small
groups: A content analysis of group brainstorming performance. Creativity Research
Journal, 12(3), 175–184. https://doi.org/10.1207/s15326934crj1203_2
Lerner, J., & Triole, J. (2000). The simple economics of open source (Working Paper No. 7600).
National Bureau of Economic Research. Retrieved from
http://www.nber.org/papers/w7600
Levine, S. S., & Prietula, M. J. (2014). Open collaboration for innovation: Principles and
performance. Organization Science, 25(5), 1414–1433.
https://doi.org/10.1287/orsc.2013.0872
Liebrand, W. B. G. (1984). The effect of social motives, communication and group size on
behaviour in an N-person multi-stage mixed-motive game. European Journal of Social
Psychology, 14(3), 239–264. https://doi.org/10.1002/ejsp.2420140302
Liebrand, W. B. G., Jansen, R. W. T. L., Rijken, V. M., & Suhre, C. J. M. (1986). Might over
morality: Social values and the perception of other players in experimental games.
Journal of Experimental Social Psychology, 22(3), 203–215.
https://doi.org/10.1016/0022-1031(86)90024-7
Liebrand, W. B. G., & McClintock, C. G. (1988). The ring measure of social values: A
computerized procedure for assessing individual differences in information processing
and social value orientation. European Journal of Personality, 2(3), 217–230.
https://doi.org/10.1002/per.2410020304
116
Liebrand, W. B. G., & van Run, G. J. (1985). The effects of social motives on behavior in social
dilemmas in two cultures. Journal of Experimental Social Psychology, 21(1), 86–102.
https://doi.org/10.1016/0022-1031(85)90008-3
Liebrand, W. B. G., Wilke, H. A. M., Vogel, R., & Wolters, F. J. M. (1986). Value orientation
and conformity: A study using three types of social dilemma games. Journal of Conflict
Resolution, 30(1), 77–97. https://doi.org/10.1177/0022002786030001006
Litman, L., Robinson, J., & Rosenzweig, C. (2015). The relationship between motivation,
monetary compensation, and data quality among US- and India-based workers on
Mechanical Turk. Behavior Research Methods, 47(2), 519–528.
https://doi.org/10.3758/s13428-014-0483-x
Lozano, J. H. (2018). The situational strength hypothesis and the measurement of personality.
Social Psychological and Personality Science, 9(1), 70–79.
https://doi.org/10.1177/1948550617699256
MacCrimmon, K. R., & Messick, D. M. (1976). A framework for social motives. Behavioral
Science, 21(2), 86–100. https://doi.org/10.1002/bs.3830210203
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. https://doi.org/10.1016/j.jsis.2013.07.004
Malhotra, A., & Majchrzak, A. (2014). Managing crowds in innovation challenges (SSRN
Scholarly Paper No. ID 2535980). Rochester, NY: Social Science Research Network.
Retrieved from http://papers.ssrn.com/abstract=2535980
Malone, T. W., Laubacher, R., & Dellarocas, C. (2010). The collective intelligence genome. MIT
Sloan Management Review, 51(3), 21–31.
117
Marwell, G., & Oliver, P. (1993). The critical mass in collective action (Reprint edition).
Cambridge, UK; New York, USA: Cambridge University Press.
Mason, W., & Suri, S. (2012). Conducting behavioral research on Amazon’s Mechanical Turk.
Behavior Research Methods, 44(1), 1–23. https://doi.org/10.3758/s13428-011-0124-6
McClintock, C. G. (1972). Social motivation - A set of propositions. Behavioral Science, 17(5),
438–454.
McClintock, C. G., & Allison, S. T. (1989). Social value orientation and helping behavior.
Journal of Applied Social Psychology, 19(4), 353–362. https://doi.org/10.1111/j.1559-
1816.1989.tb00060.x
McClintock, C. G., & Liebrand, W. B. (1988). Role of interdependence structure, individual
value orientation, and another’s strategy in social decision making: A transformational
analysis. Journal of Personality and Social Psychology, 55(3), 396–409.
http://dx.doi.org.libproxy1.usc.edu/10.1037/0022-3514.55.3.396
McClintock, C. G., Messick, D. M., Kuhlman, D. M., & Campos, F. T. (1973). Motivational
bases of choice in three-choice decomposed games. Journal of Experimental Social
Psychology, 9(6), 572–590. https://doi.org/10.1016/0022-1031(73)90039-5
Messick, D. M., & McClintock, C. G. (1968). Motivational bases of choice in experimental
games. Journal of Experimental Social Psychology, 4(1), 1–25.
https://doi.org/10.1016/0022-1031(68)90046-2
Miller, G. A., & Chapman, J. P. (2001). Misunderstanding analysis of covariance. Journal of
Abnormal Psychology, 110(1), 40–48.
Milliken, F. J., Bartel, C. A., & Kurtzberg, T. R. (2003). Diversity and creativity in work groups:
A dynamic perspective on the affective and cognitive processes that link diversity and
118
performance. In P. B. Paulus & B. A. Nijstad, Group creativity : Innovation through
collaboration (pp. 32–62). Cary, NC: Oxford University Press.
Murphy, R. O., & Ackermann, K. A. (2014). Social value orientation: Theoretical and
measurement issues in the study of social preferences. Personality and Social Psychology
Review, 18(1), 13–41. https://doi.org/10.1177/1088868313501745
Murphy, R. O., Ackermann, K. A., & Handgraaf, M. (2011). Measuring social value orientation.
Judgment and Decision Making, 6(8), 771–781.
Nauta, A., Dreu, C. K. W. D., & Vaart, T. van der. (2002). Social value orientation,
organizational goal concerns and interdepartmental problem-solving behavior. Journal of
Organizational Behavior, 23(2), 199–213. https://doi.org/10.1002/job.136
Nemeth, C. J., Personnaz, B., Personnaz, M., & Goncalo, J. A. (2004). The liberating role of
conflict in group creativity: A study in two countries. European Journal of Social
Psychology, 34(4), 365–374. https://doi.org/10.1002/ejsp.210
Nijstad, B. A., Berger-Selman, F., & Dreu, C. K. W. D. (2014). Innovation in top management
teams: Minority dissent, transformational leadership, and radical innovations. European
Journal of Work and Organizational Psychology, 23(2), 310–322.
https://doi.org/10.1080/1359432X.2012.734038
Nijstad, B. A., & de Dreu, C. K. W. (2012). Motivated information processing in organizational
teams: Progress, puzzles, and prospects. Research in Organizational Behavior, 32, 87–
111. https://doi.org/10.1016/j.riob.2012.11.004
Nijstad, B. A., Stroebe, W., & Lodewijkx, H. F. M. (1999). Persistence of brainstorming groups:
How do people know when to stop? Journal of Experimental Social Psychology, 35(2),
165–185. https://doi.org/10.1006/jesp.1998.1374
119
Nijstad, B. A., Stroebe, W., & Lodewijkx, H. F. M. (2002). Cognitive stimulation and
interference in groups: Exposure effects in an idea generation task. Journal of
Experimental Social Psychology, 38(6), 535–544. https://doi.org/10.1016/S0022-
1031(02)00500-0
Oldham, G. R., & Cummings, A. (1996). Employee creativity: Personal and contextual factors at
work. The Academy of Management Journal, 39(3), 607–634.
https://doi.org/10.2307/256657
Olekalns, M., & Smith, P. L. (1999). Social value orientations and strategy choices in
competitive negotiations. Personality and Social Psychology Bulletin, 25(6), 657–668.
https://doi.org/10.1177/0146167299025006002
Olson, M. (1965). The logic of collective action: Public goods and the theory of groups (Revised
edition). Cambridge, MA: Harvard University Press.
Osborn, A. F. (1953). Applied imagination: Principles and procedures of creative thinking. New
York: Charles Scribner’s Sons.
Ostrom, E. (1990). Governing the commons: The evolution of institutions for collective action.
Cambridge, UK; New York, NY: Cambridge University Press.
Parks, C. D., & Rumble, A. C. (2001). Elements of reciprocity and social value orientation.
Personality and Social Psychology Bulletin, 27(10), 1301–1309.
https://doi.org/10.1177/01461672012710006
Paulus, Paul B. (2000). Groups, teams, and creativity: The creative potential of idea-generating
groups. Applied Psychology, 49(2), 237–262. https://doi.org/10.1111/1464-0597.00013
Paulus, Paul B., Brown, V., & Ortega, A. H. (1999). Group creativity. In Social creativity (Vol.
2, pp. 151–176). Cresskill, N.J.: Hampton.
120
Paulus, Paul B., & Dzindolet, M. T. (1993). Social influence processes in group brainstorming.
Journal of Personality and Social Psychology, 64(4), 575–586.
http://dx.doi.org.libproxy.usc.edu/10.1037/0022-3514.64.4.575
Paulus, Paul B., Kohn, N. W., Arditti, L. E., & Korde, R. M. (2013). Understanding the group
size effect in electronic brainstorming. Small Group Research, 44(3), 332–352.
https://doi.org/10.1177/1046496413479674
Paulus, Paul B., Larey, T. S., & Ortega, A. H. (1995). Performance and perceptions of
brainstormers in an organizational setting. Basic and Applied Social Psychology, 17(1
and 2), 249–265.
Paulus, Paul B., Larey, T. S., Putman, V. L., Leggett, K. L., & Roland, E. J. (1996). Social
influence processing in computer brainstorming. Basic and Applied Social Psychology,
18(1), 3–14. https://doi.org/10.1207/s15324834basp1801_2
Paulus, Paul B., & Yang, H.-C. (2000). Idea generation in groups: A basis for creativity in
organizations. Organizational Behavior and Human Decision Processes, 82(1), 76–87.
https://doi.org/10.1006/obhd.2000.2888
Platow, M. J., & Shave, R. (1995). Social value orientations and the expression of achievement
motivation. Journal of Social Psychology, 135(1), 71–81.
Powell, W. W. (1990). Neither market nor hierarchy: Network forms of organization. In B. M.
Staw & L. L. Cummings (Eds.), Research in organizational behaviour: An annual series
of analytical essays and critical reviews (Vol. 12, pp. 295–336). Greenwich, CT: JAI
Press.
121
Pulford, B. D., Krockow, E. M., Colman, A. M., & Lawrence, C. L. (2016). Social value
induction and cooperation in the Centipede game. PLOS ONE, 11(3), e0152352.
https://doi.org/10.1371/journal.pone.0152352
Quinn, A. J., & Bederson, B. B. (2011). Human computation: A survey and taxonomy of a
growing field. In Proceedings of the SIGCHI Conference on Human Factors in
Computing Systems (pp. 1403–1412). New York, NY: ACM.
https://doi.org/10.1145/1978942.1979148
Roch, S. G., & Samuelson, C. D. (1997). Effects of Environmental Uncertainty and Social Value
Orientation in Resource Dilemmas. Organizational Behavior and Human Decision
Processes, 70(3), 221–235. https://doi.org/10.1006/obhd.1997.2707
Sawyer, K. (2007). Group genius: The creative power of collaboration. New York, NY: Basic
Books.
Shalley, C. E., Zhou, J., & Oldham, G. R. (2004). The effects of personal and contextual
characteristics on creativity: Where should we go from here? Journal of Management,
30(6), 933–958. https://doi.org/10.1016/j.jm.2004.06.007
Shepherd, M. M., Briggs, R. O., Reinig, B. A., Yen, J., & Nunamaker, J. F., Jr. (1995). Invoking
social comparison to improve electronic brainstorming: Beyond anonymity. Journal of
Management Information Systems, 12(3), 155–170.
Shin, S. J., & Zhou, J. (2007). When is educational specialization heterogeneity related to
creativity in research and development teams? Transformational leadership as a
moderator. Journal of Applied Psychology, 92(6), 1709–1721.
http://dx.doi.org.libproxy.usc.edu/10.1037/0021-9010.92.6.1709
122
Simpson, B. (2004). Social values, subjective transformations, and cooperation in social
dilemmas. Social Psychology Quarterly, 67(4), 385–395.
Smith, S. M. (2003). The constraining effects of initial ideas. In Paul B. Paulus & B. A. Nijstad,
Group creativity : Innovation through collaboration (pp. 15–31). Cary, NC: Oxford
University Press.
Surowiecki, J. (2005). The wisdom of crowds. New York, NY: Anchor.
Taggar, S. (2002). Individual creativity and group ability to utilize individual creative resources:
A multilevel model. The Academy of Management Journal, 45(2), 315–330.
https://doi.org/10.2307/3069349
Tajfel, H., & Turner, J. (1979). The social identity theory of intergroup conflict. In W. G. Austin
& S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–48).
Monterey, CA: Brooks/Cole.
Terwiesch, C., & Xu, Y. (2008). Innovation contests, open innovation, and multiagent problem
solving. Management Science, 54(9), 1529–1543.
https://doi.org/10.1287/mnsc.1080.0884
Thompson, L., & Brajkovich, L. F. (2003). Improving the creativity of organizational work
groups [and executive commentary]. The Academy of Management Executive (1993-
2005), 17(1), 96–111.
Utz, S. (2004). Self-construal and cooperation: Is the interdependent self more cooperative than
the independent self? Self and Identity, 3(3), 177–190.
https://doi.org/10.1080/13576500444000001
123
Valacich, J. S., Dennis, A. R., & Connolly, T. (1994). Idea generation in computer-based groups:
A new ending to an old story. Organizational Behavior and Human Decision Processes,
57(3), 448–467. https://doi.org/10.1006/obhd.1994.1024
Van de Ven, A. H., & Delbecq, A. L. (1974). The effectiveness of nominal, Delphi, and
interacting group decision making processes. The Academy of Management Journal,
17(4), 605–621. https://doi.org/10.2307/255641
van Dijk, E., & de Cremer, D. (2006). Self-benefiting in the allocation of scarce resources:
Leader-follower effects and the moderating effect of social value orientations.
Personality and Social Psychology Bulletin, 32(10), 1352–1361.
https://doi.org/10.1177/0146167206290338
van Kleef, G. A., & van Lange, P. A. M. (2008). What other’s disappointment may do to selfish
people: Emotion and social value orientation in a negotiation context. Personality and
Social Psychology Bulletin, 34(8), 1084–1095.
https://doi.org/10.1177/0146167208318402
van Lange, P. A. M. (1999). The pursuit of joint outcomes and equality in outcomes: An
integrative model of social value orientation. Journal of Personality and Social
Psychology, 77(2), 337–349. http://dx.doi.org/10.1037/0022-3514.77.2.337
van Lange, P. A. M. (2000). Beyond self-interest: A set of propositions relevant to interpersonal
orientations. European Review of Social Psychology, 11(1), 297–331.
https://doi.org/10.1080/14792772043000068
van Lange, P. A. M., Klapwijk, A., & Munster, L. M. V. (2011). How the shadow of the future
might promote cooperation. Group Processes & Intergroup Relations, 14(6), 857–870.
https://doi.org/10.1177/1368430211402102
124
van Lange, P. A. M., & Liebrand, W. B. G. (1991). Social value orientation and intelligence: A
test of the goal prescribes rationality principle. European Journal of Social Psychology,
21(4), 273–292. https://doi.org/10.1002/ejsp.2420210402
van Lange, P. A. M., Otten, W., de Bruin, E. M., & Joireman, J. A. (1997). Development of
prosocial, individualistic, and competitive orientations: theory and preliminary evidence.
Journal of Personality and Social Psychology, 73(4), 733–746.
van Lange, P. A. M., & Semin-Goossens, A. (1998). The boundaries of reciprocal cooperation.
European Journal of Social Psychology, 28(5), 847–854.
https://doi.org/10.1002/(SICI)1099-0992(199809/10)28:5<847::AID-EJSP886>3.0.CO;2-
L
van Lange, P. A. M., van Vugt, M., Meertens, R. M., & Ruiter, R. A. C. (1998). A social
dilemma analysis of commuting preferences: The roles of social value orientation and
trust. Journal of Applied Social Psychology, 28(9), 796–820.
https://doi.org/10.1111/j.1559-1816.1998.tb01732.x
von Hippel, E. (1994). “Sticky information” and the locus of problem solving: Implications for
innovation. Management Science, 40(4), 429–439.
von Hippel, E. (2006). Democratizing innovation. Cambridge, Mass.: The MIT Press.
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.
Wasko, M. M., Faraj, S., & Teigland, R. (2004). Collective action and knowledge contribution in
electronic networks of practice. Journal of the Association for Information Systems,
5(11/12), 493–513.
125
Wu, H., Corney, J., & Grant, M. (2014). Relationship between quality and payment in
crowdsourced design. In Proceedings of the 2014 IEEE 18th International Conference on
Computer Supported Cooperative Work in Design (CSCWD) (pp. 499–504).
https://doi.org/10.1109/CSCWD.2014.6846895
Yamagishi, T. (1995). Social dilemmas. In K. S. Cook, G. A. Fine, & J. S. House (Eds.),
Sociological perspectives on social psychology. Boston, MA: Allyn & Bacon.
Yu, L., & Nickerson, J. V. (2011). Cooks or cobblers? Crowd creativity through combination. In
Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp.
1393–1402). New York, NY, USA: ACM. https://doi.org/10.1145/1978942.1979147
Yu, L., & Nickerson, J. V. (2013). An internet-scale idea generation system. ACM Transactions
on Interactive Intelligent Systems, 3(1), 2:1–2:24.
https://doi.org/10.1145/2448116.2448118
Yu, L., Nickerson, J. V., & Sakamoto, Y. (2012). Collective creativity: Where we are and where
we might go. In Proceedings of the Collective Intelligence Conference, 2012. Retrieved
from http://arxiv.org/abs/1204.3890
Zajonc, R. B. (1965). Social facilitation. Science, 149(3681), 269–274.
Ziegler, R., Diehl, M., & Zijlstra, G. (2000). Idea production in nominal and virtual groups: Does
computer-mediated communication improve group brainstorming? Group Processes &
Intergroup Relations, 3(2), 141–158.
126
Appendices
Appendix A: Summary of Crowdsourced Idea Generation Projects in Pilot Study 1

Project
Name Sponsor Task Topic Task Type Task Structure Incentives Communication
Banchile Inversiones Product development Self/Community Cooperative Social Yes
BrightIdeas Bayada
Product development/ Internal service
improvement (to employees) Self/Community Cooperative Social Yes
The
Innovation
Hub Cerebral Palsy Alliance Product development Self/Other Cooperative Social Unclear
World
Cerebral
Palsy Day Cerebral Palsy Alliance Product development Community Cooperative/Competitive Social Yes
Citrix Product development Self/Community Cooperative Social Yes
City of
Atlanta City of Atlanta Public service development Community Competitive Social No
Imagine
Huntsville City of Huntsville Public service development Community Cooperative Social Yes
City of Minneapolis Public service development Community Cooperative Social Yes
What to Fix
Columbia Columbia University Community service development Community Cooperative Social Yes
Sunshot
Catalyst Department of Energy Technology solution Self Competitive Social/Monetary Yes
ePolicyWorks Department of Labor Public service development Other Cooperative Social Yes
EA Sports Electronic Arts Inc. Product development Self Cooperative/Competitive Social Yes
Georgetown University Community service development Community Cooperative Social Yes
The Fix
California Innovate Your State Public service development Self/Community Competitive Social/Monetary
Unclear, seemed
not
Kane Product development Self/Community Cooperative Social/Monetary Yes
127
Laptime Club Magneti Marelli Product development Self Cooperative/Competitive Social Yes
Global
Innovation
Competition Making All Voices Count Public service development Self/Community Competitive Social/Monetary No
WTF@MSA MSA Internal service improvement Community Cooperative Social Yes
Investigate
This! NBC Product development Self/Community Cooperative Social Yes
GeoVation Ordnance Survey Tech solution Self Competitive Social/Monetary Yes
Ottawa Public Library Public service development Community Cooperative Social Yes
We Innovate Princess Cruises Product development Self/Community Cooperative Social/Monetary Yes
State of Minnesota Public service development Self/Community Competitive Social
Unclear, seemed
not
Scentsy
Family
IdeaShare Scentsy Product development Self/Community Cooperative/Competitive Social Yes
Techsmith Product development Self Cooperative Social Yes

Texas Health and Human
Services Public service development Self/Community Competitive Social No
Proyecto
FAST
The City of Seseña and the
Center for Excellence in Spain Improve personal business plans Self Cooperative Social Yes
The University of Calgary Community service development Community Cooperative Social Yes
SAVE The White House Public service development Community Competitive Social Yes
Idea Farm TTI Group Product development Self/Community Cooperative/Competitive Social Unclear
Ideaction UC Berkeley Community service development Community Cooperative Social Yes

University of North Carolina
Wilmington Community service development Community Cooperative
Unclear, seemed
social Yes
Veteran' Administration Public service development Self/Community/Other Cooperative Social/Monetary Yes
Western Australia Police Public service development Self/Community Cooperative Social/Monetary Yes
128
Appendix B: Measures of SVO

The Slider Measure

1
2
3
4
5
6
You
Other
You
Other
You
Other
You
Other
You
Other
You
Other
50
100
54 59 63 68 72 76 81 85
96 94 93 91 89 87 85 98
You receive
Other receives
50
100
56 63 69 75 81 88 94 100
88 81 75 69 63 56 50 94
You receive
Other receives
50
100
54 59 63 68 72 76 81 85
79 68 58 47 36 26 15 89
You receive
Other receives
85
85
87 89 91 93 94 96 98 100
76 72 68 63 59 54 50 81
You receive
Other receives
85
85
85 85 85 85 85 85 85 85
68 59 50 41 33 24 15 76
You receive
Other receives
100
50
98 96 94 93 91 89 87 85
41 37 33 28 24 19 15 46
You receive
Other receives
In this task you have been randomly paired with another person, whom we will refer to as the other. This other person is someone you
do not know and will remain mutually anonymous. All of your choices are completely confidential. You will be making a series of
decisions about allocating resources between you and this other person. For each of the following questions, please indicate the
distribution you prefer most by marking the respective position along the midline. You can only make one mark for each question.
Your decisions will yield money for both yourself and the other person. In the example below, a person has chosen to distribute money
so that he/she receives 50 dollars, while the anonymous other person receives 40 dollars.
There are no right or wrong answers, this is all about personal preferences. After you have made your decision, write the resulting
distribution of money on the spaces on the right. As you can see, your choices will influence both the amount of money you receive
as well as the amount of money the other receives.
Example:
Instructions
30
80
35 40 45 50 55 60 65 70
60 50 40 30 20 10 0 70
You receive
Other receives
40
50
You
Other
a
129
Triple-Dominance Measure
Imagine that you have been randomly paired with another person, whom you do not
know and will not knowingly meet or communicate in the future. Both of you are making
choices among three options (A, B, or C) by circling one of the options. Each option represents a
different point allocation to yourself and another person. Therefore, your choices determine the
number of points you received and the number of points the other person receives. Also, the
other person’s choices determine the number of points you receive and the number of points s/he
receives. The points are important to you and also to the other person. The more points you get,
the better off you are. Also, the more points the other person gets, the better off s/he will be.

Here is an example:
A B C
You get 50 60 40
Other gets 50 30 0

The first row (You get) represents the number of points that you will receive while the second
row (Other gets) represents the number of points that the other person will receive. If you chose
A, you would receive 50 points and the other person would receive 50 points. If you chose B,
you would receive 60 points and the other person would receive 30 points. And if you chose C,
you would receive 40 points and the other person would receive 0 point.

Please keep in mind that there are no right or wrong answers – choose the option that you
find most attractive. Also, remember that the points are valuable – the more you get, the better
for you. Likewise, the more points the other person gets, the better for him/her.

130
For each of the twelve choice situations, circle A, B, or C, depending on which column you
prefer most.

1) A B C
You get 480 540 480
Other gets 80 280 480

2) A B C
You get 560 500 500
Other gets 300 500 100

3) A B C
You get 520 520 580
Other gets 520 120 320

4) A B C
You get 510 560 510
Other gets 510 300 110

5) A B C
You get 550 500 500
Other gets 300 100 500

6) A B C
You get 480 490 540
Other gets 100 490 300

7) A B C
You get 500 560 490
Other gets 100 300 490

8) A B C
You get 560 500 490
Other gets 300 500 90

9) A B C
You get 500 500 570
Other gets 500 100 300

10) A B C
You get 480 520 480
Other gets 480 300 180

11) A B C
You get 470 330 440
Other gets 300 110 440

12) A B C
You get 460 510 530
Other gets 100 510 320
131
Appendix C: Materials for Pilot Study 2 (Pilot Experiments 1-3)

Recruitment Ad
Title: Short 10-15min web-survey on online idea contribution

Hi there!
I am a PhD student at the Annenberg School for Communication in the University of Southern
California. I’m conducting a study on online idea generation. You must live in the US and be 18
years old or older to participate in this study.
This study includes two main steps:
1. Informed consent and instructions: Click on the link below, which will lead you to a
Qualtrics survey. You will read detailed information about the study there. Once you agree to
participate in this study, instructions about the task will be given.
2. Main Survey: You will complete a questionnaire in the Qualtrics survey. A confirmation
code for MTurk payment will be given at the end of this survey.
The entire process will take 15 minutes. Payment will be processed within 48 hours of
completion (of all three steps).
Make sure to leave this window open as you complete the survey. When you are finished, you
will return to this page to paste the code into the box.
Please click on the link below to start the study.

Tasks
Self-oriented tasks
1. How might the current payment process in Amazon Mechanical Turk be improved?
2. Please suggest changes to Amazon Mechanical Turk to improve the efficiency of task
completion in the system.
3. Help come up with a name for a new Amazon Mechanical Turk service.

Community-oriented tasks
1. How can we improve the local public transportation system to promote its use?
2. What should the public park in our local neighborhood be like?
3. How can we reduce cost and improve service?

Other-oriented tasks:
1. How to help urban slum communities become more resilient to the effects of climate
change?
2. How to use mobile technology to facilitate health care access in the developing
countries?
3. How to improve education and expand learning opportunities for refugees around the
world?

132
Measure of Task Interest
After viewing each task listed in the previous section:

Are you interested in contributing your ideas to it?
Please select the answer that best represents your interest

Extremely
uninterested
Moderately
uninterested
Slightly
uninterested

Neither
interested
nor
uninterested
Slightly
interested

Moderately
interested
Extremely
interested

Demographic Measures
What is your age?
o 18-24 years
o 25-34 years
o 35-44 years
o 45-54 years
o 55-64 years
o 65 years or more
What is the highest level of education you have completed?
o Less than High School
o High School / GED
o Some College
o 2-year College Degree
o 4-year College Degree
o Master’s Degree
o Doctoral Degree
o Professional Degree (JD, MD)
What is your gender?
o Male
o Female
o Other
133

Appendix D: Materials for Pilot Study 2 (Pilot Experiments 4)

Recruitment Ad
Title: Short 15min online idea generation activity + web survey

Hi there!

I am a PhD student at the Annenberg School for Communication in the University of Southern
California. I’m conducting a study on online idea generation. You must live in the US and be 18
years old or older to participate in this study.

This study includes two main steps:

1. Informed consent and instructions: Click on the link below, which will lead you to a
Qualtrics survey. You will read detailed information about the study there. Once you agree to
participate in this study, instructions about the task will be given.

2. Main Survey: You will complete a questionnaire in the Qualtrics survey and develop ideas
to help solve questions assigned to you. A confirmation code for MTurk payment will be given
at the end of this survey.

The entire process will take approximately 15-20 minutes. Payment will be processed within 48
hours of completion (of all three steps).

Make sure to leave this window open as you complete the survey. When you are finished, you
will return to this page to paste the code into the box.

Please click on the link below to start the study/

Tasks
Self-oriented task:
How might Amazon Mechanical Turk improve task completion efficiency of workers?

Community-oriented task:
What features should the public park in our local neighborhood have?

Other-oriented task:
How can we use mobile technology to facilitate health care access in the developing countries?

Demographic Measures (the same as in Appendix C)

134
Appendix E: Materials for the Crowdsourced Idea Evaluation Task

Recruitment Ad
Title: Short 10min idea evaluation activity

Hi there!

I am a PhD student at the Annenberg School for Communication in the University of Southern
California. I’m conducting a study on online idea generation and need your help to evaluate ideas
generated by previous participants. You must live in the US and be 18 years old or older to
participate in this study.

This study includes two main steps:

1. Informed consent and instructions: Click on the link below, which will lead you to a
Qualtrics survey. You will read detailed information about the study there. Once you agree to
participate in this study, instructions about the task will be given.

2. Main Survey: You will rate the ideas in a questionnaire in the Qualtrics survey. A
confirmation code for MTurk payment will be given at the end of this survey.

The entire process will take about 10 minutes. Payment will be processed within 48 hours of
completion (of all three steps).

Make sure to leave this window open as you complete the survey. When you are finished, you
will return to this page to paste the code into the box.

Please click on the link below to start the study/

Scale of Idea Usefulness
After viewing each idea randomly assigned to them:

How useful is the idea? -- To what extent can the idea help solve the question? Please note an
idea that's hard to understand should be rated as not useful.
Not useful
at all
Little
usefulness
Slightly
useful

Somewhat
useful
Moderately
useful

Highly
useful
Extremely
useful

135
Appendix F: Materials for the Main Experiments (Study 1-3)

Recruitment Ad
Title: Short 15-20min online idea generation activity + web survey

Hi there!

I am a PhD student at the Annenberg School for Communication in the University of Southern
California. I’m conducting a study on online idea generation. You must live in the US and be 18
years old or older to participate in this study.

This study includes two main steps:

1. Informed consent and instructions: Click on the link below, which will lead you to a
Qualtrics survey. You will read detailed information about the study there. Once you agree to
participate in this study, instructions about the task will be given.

2. Main Survey: You will help generate ideas to answer a question given to you. You will also
complete a questionnaire in the survey. A confirmation code for MTurk payment will be given at
the end of this survey.

The entire process will take approximately 15-20 minutes. Payment will be processed within 48
hours of completion (of all three steps).

Make sure to leave this window open as you complete the survey. When you are finished, you
will return to this page to paste the code into the box.

Please click on the link below to start the study.

Task
What can MTurk requesters do to ensure workers perform their HITs with good quality?

Manipulation of Task Structure
Cooperative task condition:

“You will collaborate in groups to generate ideas to answer a question posted by the
requester. The HIT you are completing are grouped into batches of 9 assignments. You
will collaborate with 8 other workers who accept the same batch to get the chance to earn
a $15 bonus. You will enter a lottery for the bonus after the HIT is completed. For each
non-redundant idea your batch generates, you get one chance in the lottery for the
bonus. The more non-redundant ideas your batch collectively generates, the more likely
you and other workers in your batch are to receive the bonus.”

136
Competitive task condition:

“You will compete in groups to generate ideas to answer a question posted by the
requester. The HIT you are completing are grouped into batches of 9 assignments. You
will compete with 8 other workers who accept the same batch to get the chance to earn a
$15 bonus. If you generate the most non-redundant ideas in your batch, you will enter a
lottery to win the bonus after the HIT is competed. For each non-redundant idea you
generate, you get one chance in the lottery for the bonus. The more non-redundant ideas
yourself generate, the more likely you are to receive the bonus.”

Manipulation check:

1. [After task instruction] Now please select the answer according to the instruction. This
question has a correct answer. Failure to choose the correct answer may result in rejection
of your work.

In this task, I will:
a) generate non-redundant ideas to win a $5 bonus;
b) collaborate with my group of workers to generate non-redundant ideas and win a $15
bonus;
c) compete with my group of workers to generate non-redundant ideas and win a $15
bonus;
d) evaluate others’ ideas to win a $10 bonus.

2. [After idea generation] Please rate to what extent you agree with the statements below:
a) When I was generating the ideas, I tried to contribute to my batch of workers
b) When I was generating the ideas, I tried to perform better than other workers in my
batch

Strongly
disagree
Disagree Somewhat
disagree
Neither agree
nor disagree
Somewhat
agree
Agree Strongly
Agree

Manipulation of Idea Exposure (Study 2)
Common Idea Condition:

“In the next session, please generate ideas to answer the question:

What can MTurk requesters do to ensure workers perform their HITs with good
quality?

An example of an idea that helps solve the question is:

Offer bonuses for coherent and intelligent answers ”

137
Novel Idea Condition:

“In the next session, please generate ideas to answer the question:

What can MTurk requesters do to ensure workers perform their HITs with good
quality?

An example of an idea that helps solve the question is:

Improve the HIT's website access and provide instructions on what to do if there's a
glitch ”

Manipulation check:
What is the example answer to the proposed question listed above?
Please rephrase the above example in your own words below:

Manipulation of Information about Other’s Performance (Study 3)
Average performance condition:

“In the next session, please generate ideas to answer the question:

What can MTurk requesters do to ensure workers perform their HITs with good
quality?

One average, each worker generates about 7 ideas in our previous activities. ”

Manipulation check:

[After the above information] Please choose: how many ideas on average did workers
generate in our prior activities?
a) 5
b) 6
c) 7
d) 8

High performance condition:

“In the next session, please generate ideas to answer the question:

What can MTurk requesters do to ensure workers perform their HITs with good
quality?

One average, each worker generates about 7 ideas in our previous activities. ”

138
Manipulation check:

[After the above information] Please choose: how many ideas on average did workers
generate in our prior activities?
a) 12
b) 13
c) 14
d) 15

Measure of Task Effort
Please rate to what extent you agree with the statements below:
a) I was motivated to exert efforts when I was generating the ideas
b) I tried hard to generate ideas to answer the question

Strongly
disagree
Disagree Somewhat
disagree
Neither agree
nor disagree
Somewhat
agree
Agree Strongly
Agree

Measure of Task Relevance
Please rate to what extent you agree with the statements below:
In general, I found the question posted by the requester relevant to me

Strongly
disagree
Disagree Somewhat
disagree
Neither agree
nor disagree
Somewhat
agree
Agree Strongly
Agree

Measure of Task Interest
a) In general, how interested are you in contributing your ideas to this type of online idea
generation activities?
b) In general, how interested are you in contributing your ideas to problem solving?

Extremely
uninterested
Moderately
uninterested
Slightly
uninterested

Neither
interested
nor
uninterested
Slightly
interested

Moderately
interested
Extremely
interested

Demographic Measures (the same as in Appendix C)

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Crowdsourced art: activating creative participation in online spaces

PDF

Social motivation and credibility in crowdfunding

PDF

The evolution of multidimensional and multilevel networks in online crowdsourcing

PDF

Generation Z: social media, influencers and brand loyalty in entertainment

PDF

Conspicuous connections as signals of expertise in networks

PDF

Crowdsourcing for integrative and innovative knowledge: knowledge diversity, network position, and semantic patterns of collective reflection

PDF

How social and human capital create financial capital in crowdfunding projects

PDF

Building social Legoland through collaborative crowdsourcing: marginality, functional diversity, and team success

PDF

The social meaning of sharing and geocoding: features and social processes in online communities

PDF

Under the influence: understanding cautious influencer marketing and relations

PDF

Identity, trust, and credibility online: evaluating contradictory user-generated information via the warranting principle

PDF

Memes, meme marketing, and how brands and influencers can leverage them on social media

PDF

Bounded technological rationality: the intersection between artificial intelligence, cognition, and environment and its effects on decision-making

PDF

Dynamic pricing and task assignment in real-time spatial crowdsourcing platforms

PDF

The influencer break-down: a history, categorization, and analysis of success for health & wellness influencers

PDF

Digital creativity and innovation in Chinese social network sites industry

PDF

The formation and influence of online health social networks on social support, self-tracking behavior and weight loss outcomes

PDF

Digital discourse in the fashion industry

PDF

Ecology and network evolution in online innovation contest crowdsourcing

PDF

Why we follow influencers: the role of Asian female fashionistas in Los Angeles

Asset Metadata

Creator Yan, Bei (author)

Core Title Social value orientation, social influence and creativity in crowdsourced idea generation

School Annenberg School for Communication

Degree Doctor of Philosophy

Degree Program Communication

Publication Date 08/03/2018

Defense Date 08/02/2018

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag creativity,crowdsourcing,idea generation,OAI-PMH Harvest,social influence,social value orientation

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Hollingshead, Andrea (committee chair), Carnevale, Peter (committee member), Jian, Lian (committee member)

Creator Email beiyan@usc.edu,beiyanusc@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-49060

Unique identifier UC11668930

Identifier etd-YanBei-6620.pdf (filename),usctheses-c89-49060 (legacy record id)

Legacy Identifier etd-YanBei-6620.pdf

Dmrecord 49060

Document Type Dissertation

Format application/pdf (imt)

Rights Yan, Bei

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