University of Southern California Dissertations and Theses

Crowdsourced art: activating creative participation in online spaces

(USC Thesis Other)

Crowdsourced art: activating creative participation in online spaces

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content !

CROWDSOURCED ART:

ACTIVATING CREATIVE PARTICIPATION

IN ONLINE SPACES

by

Ioana Literat

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 2015

Copyright 2015 Ioana Literat
! ii!
DEDICATION

This dissertation is dedicated to every single Mechanical Turk worker who contributed to
my research. Because you saw an odd request to draw a snail called Hashtag for 5 cents
and you said “why not?”. Because, for those 5 cents, instead of doing the bare minimum,
you took your time and painted the sky and the grass and the swirls on Hashtag’s shell.
Because you know and I know that it is not about those 5 cents at all.

I continue to be in awe of your creativity, inspiration and humor.

! iii!
ACKNOWLEDGMENTS

First of all, I would like to thank my advisor and dissertation chair, Henry
Jenkins, for his unwavering support and dedication throughout the years. I never felt
alone in this process, I always knew I could count on him and I am so thankful for his
mentorship and feedback. He has enriched my ideas, challenged my thinking, and shaped
me into the young scholar I am today. An ideal mentor, Henry gave me the confidence to
carve my own path, while being there for me whenever I needed help or advice.
I want to thank my brilliant committee members, Larry Gross, Andrea
Hollingshead and Daren Brabham. It was in Larry’s class that I wrote my first paper on
crowdsourced art and he has been one of my biggest supporters ever since. Thank you for
believing in me. I am grateful to Andrea for encouraging me to jump head-first into the
word of quantitative research and for all the help she has given me along the way. This
dissertation would not have been possible without her, and I am forever grateful to her for
making me a more fearless and well-rounded scholar. I would like to thank Daren for his
feedback and advice; as a young researcher, I look up to him as a role model and I am in
awe of his achievements.
I am grateful to the Annenberg School for Communication and to the USC
Graduate School for their financial support and for providing me with a stimulating
intellectual environment for my research. At Annenberg, I always felt free to be myself
and to explore a wide variety of research interests. I appreciate this freedom
tremendously, I realize that it is very rare and I feel incredibly lucky. I could not have
chosen a better destination for my graduate studies.
! iv!
Special thanks to Neta Kligler-Vilenchik, who shared her statistics wisdom with
me when I needed it most, and to Amanda Ford for all her logistical support. I am also
grateful to the 2013 Oxford Internet Institute cohort: it was in conversations with them
that the idea of a children’s book about the Internet took shape.
On the personal front, I want to thank my parents, Carmen and Marius Dan
Literat, who made all this possible. They sacrificed everything so that I can follow my
dreams. For this, I thank them every day.
I am incredibly grateful to my boyfriend Michael, my favorite “solutions
consultant.” Thank you for believing in me and being there for me throughout this
process. Thank you, also, for always making sure to move snails away from the pavement
and onto the grass after it rains, so they wouldn't be stepped on. I know you do it out of
love for me and I appreciate it – as do all the snails whose lives you've saved.
Thank you to Pepito and Purrrcy, our beloved pets, who sat by me (sometimes
on me) every day as I worked on this dissertation. Feeling their warm bodies next to mine
and listening to them snoring – Pepi loudly and coarsely, Purrcy softly and daintily –
made for the best kind of work environment and reminded me, even in the pits of
statistics hell, that I am loved and needed.
Last but not least, I want to thank all the online participants who contributed to
this research. I am especially grateful to the Mechanical Turk workers, whom this
dissertation is dedicated to.

! v!
TABLE OF CONTENTS
Dedication ii
Acknowledgments iii
List of Tables vii
List of Figures viii
Abstract ix
Chapter 1: Introduction 1
Goals and Rationale 3
Theoretical Framework 5
Empirical Research Design 7
Significance of the Research 11

Chapter 2: Interrogating Participation Across Disciplinary Boundaries 13
Political Philosophy 14
Definitions and Key Debates 14
Degrees of Participation 16
Strengths and Opportunities 17
Cultural Studies 19
Definitions and Key Debates 19
Degrees of Participation 21
Strengths and Opportunities 24
Art 27
Definitions and Key Debates 27
Degrees of Participation 28
Strengths and Opportunities 30
Conclusion 32

Chapter 3: Crowdsourced Art and Collective Creativity 33
A Primer on Crowdsourcing 33
Online Crowdsourced Art: Cases and Typology 38
Crowdsourced Art and Relational Aesthetics 45
Redefining the Role of the Artist: The Alpha Artist in Crowdsourced Art 49
Redefining the Role of the Public: Qualifying Participation 53
Conclusion: Agency and Structure in Crowdsourced Art 62

Chapter 4: Methodology 67
Participants 67
Materials and Procedure 69
Measures 76
Data Analysis 80
! vi!

Chapter 5: Results 82
Quantity of Submissions 82
Quality of Submissions 84
Usefulness 84
Novelty 87
Time Spent on Task 90
Rebellious Contributions 94
Participants’ Motivations 95
Discussion 98
Limitations 105

Chapter 6: Crowdsourcing Creative Tasks on Alternative Platforms 109
hitRECord 110
Platform Description 110
Project Implementation 112
Results 113
Key Takeaways 114
Peer 2 Peer University (P2PU) 120
Platform Description 120
Project Implementation 122
Results 125
Key Takeaways 127

Chapter 7: Conclusion 134
Crowdsourcing Creative Tasks: What Works and What Doesn’t? 135
Reassessing Participation and Its Outcomes 140
An Interdisciplinary Approach to Participation 147

Epilogue: A Children’s Book about the Internet, by the Internet 150

Bibliography 153

Appendices
Appendix A: Mechanical Turk Task Instructions 166
Appendix B: Coding Scheme for Measuring Quality 169
Appendix C: Coding Scheme for Categorizing Participant Motivations 172
Appendix D: Distribution of Values According to Reward Amounts 175
Appendix E: Participant Motivations for Each Reward Amount
and Incentive 180
Appendix F: Examples of Rebellious Contributions Submitted for
Illustration Tasks 182
Appendix G: Examples of Illustration Submissions 185
Appendix H: hitRECord Task Instructions 187
Appendix I: P2PU Course Description (“About” Page) 188

! vii!
LIST OF TABLES

Table 1: A Typology of Online Crowdsourced Art 44

Table 2: The Levels of Creative Participation in Crowdsourced Art 55

Table 3: List of Mechanical Turk Tasks (Experimental Conditions) 74

Table 4: Intercoder Reliability Results for Writing Tasks 78

Table 5: Intercoder Reliability Results for Illustration Tasks 79

Table 6: Quantity of Submissions: Summary of Results 83

Table 7: Usefulness of Submissions: Summary of Results 86

Table 8: Novelty of Submissions: Summary of Results 89

Table 9: Time Spent on Task: Summary of Results 92
! viii!
LIST OF FIGURES

Figure 1: The pyramid of participation: a schematic depiction of the
levels of creative participation 56

Figure 2: A screenshot of a writing task (control condition) 70

Figure 3: A screenshot of an illustration task (control condition) 71

Figure 4: Participants’ motivations for accepting the HITs about
Hashtag the Snail 96

Figure 5: A word cloud of the words most commonly included in
written submissions 102

Figure 6: A screenshot of the hitRECord collaboration, featuring some
of the submitted records 113

Figure 7: The hitRECord conversation page 117

Figure 8: Introduction to the letter-writing campaign to Jeff Bezos, as
posted on Dynamo 140

! ix!
ABSTRACT

Recent years have seen a proliferation of artistic projects that are based on open
public participation and take place entirely online. Building on the concept of
crowdsourcing as a strategy for content generation, I describe such projects as online
crowdsourced art. From a theoretical perspective, this research introduces a useful
typology for crowdsourced art, explores its aesthetic, cultural and technological
implications, and discusses the different levels of participation that crowdsourced art
projects enable. As a practical application of my research, to better understand the
conditions that most effectively foster creative participation online, I crowdsourced the
development of a children's book about digital culture – collecting 4200 written and
visual submissions on Amazon’s Mechanical Turk – and determined, through large-scale
statistical analysis, the impact that key incentives have on the process of creative
crowdsourcing.
This work – at the intersection of communication, Internet studies, and art history
– represents the first attempt to map out this emerging field, and makes a vital
contribution to our understanding of online creative participation, crowdsourcing and
collective creativity. The proposed reconceptualization of participatory processes (as
receptive, executory and structural) enhances the current literature on online
participation, facilitating a more sophisticated understanding of the ways people engage
and collaborate in digital spaces. From a practical perspective, my findings help
illuminate the factors that have the greatest impact on the quality and quantity of online
participation, determining the role of financial rewards, bonuses, project purpose and
authorship credit in creative crowdsourcing initiatives.
!1
CHAPTER 1:
INTRODUCTION
The increasingly complex and expressive affordances of computer-mediated
communication platforms have changed the way we work, play, and relate to one another.
The advent of the Web 2.0 era with its alluring promise of feasible large-scale
participation and global interconnectedness has facilitated a cultural shift in traditional
notions of authorship, creativity and individual expression. This shift has been apparent
in fields such as business, entrepreneurship, entertainment, journalism, but also in the
arts. As companies realized the generative potential of networked online communities,
the practice of crowdsourcing became a profitable and efficient strategy of harnessing
these communities’ knowledge and creativity in order to create content, solve problems
and effectively perform corporate research and development tasks (Howe, 2006a).
Beyond these corporate contexts, the unique affordances of the Internet as a platform for
reaching a global community of potential content generators has also stimulated the rise
of an impressive number of successful online ventures built exclusively on the
phenomenon of crowdsourcing: Wikipedia, YouTube, eBay, iStockPhoto and Threadless
are only a few of the most prominent examples. Reflecting these social and cultural
trends in content generation, the arts soon followed suit, and the practice of online
crowdsourced art emerged as a significant development across the entire artistic
spectrum, from visual arts to music to creative writing.
!2
In today’s art world, “the art object is no longer necessarily the primary focus of
the encounter with art” (Beech, 2010, p. 20). In response to a perceived erosion of social
bonds, many artists have shifted their attention from creating material objects to
orchestrating and facilitating social encounters. Such projects – generally described as
participatory art - have received substantial attention in recent scholarship (Beech, 2008,
2010; Bishop, 2004, 2006a, 2006b; Downey, 2009; Foster, 2006; Kester, 2004, 2011), but
these accounts have focused on theorizing the embodied, face-to-face interaction between
participants. Given the growing number of contemporary artists who are using
technology, particularly the interactive Web, as a means to engage the public in the
collective production of art, there is a critical need to expand this research beyond
embodied contexts and into the realm of technologically-mediated participatory art as
well.
Recent years have seen a proliferation of art projects that are based on open public
participation and take place entirely online. Aaron Koblin and Chris Milk rely on Johhny
Cash fans to collectively draw a tribute to the artist, frame by frame; Ridley Scott enlists
YouTube users to film a feature-length documentary created entirely by amateur
filmmakers all over the world; Eric Whiteacre conducts virtual choirs made up of
thousands of global voices recorded by webcams and connected by the Internet; and
acclaimed director Paul Verhoeven invites his audience to participate in the collaborative
screenwriting of award-winning films. Building on the concept of crowdsourcing as a
strategy for content generation (Howe, 2006; Brabham, 2013), I describe such projects as
online crowdsourced art, which I define as the practice of using the Internet as a
!3
participatory platform to directly engage the public in the creation of artwork (Literat,
2012).
As illustrated by the historical background of participatory art, the idea of
integrating the creative input of the “crowd” into the artistic process is by no means new.
Whereas the happenings of the 60s and 70s or the conceptual art in modern gallery
exhibitions made it possible for a small audience to participate – always dependent on
their physical presence at the site of production or exhibition – the rise of Web 2.0
enabled the extension of this invitation for creative input to a global community of
potential participants. Thus, the integration of crowdsourcing into the process of artistic
production, while an innovative practice, is also a foreseeable outgrowth of recent trends
in culture and commerce, coupled with the emergence of a feasible platform that could
make this collaboration possible. However, in view of its reliance on the artistic
contribution of a large pool of usually anonymous participants, this type of art raises
consequential questions about notions of collective creativity, authorship, collaboration,
and the shifting structure of artistic production in the new digital environment.
Goals and Rationale
The purpose of my research on crowdsourced art is twofold. From a theoretical
standpoint, I want to lay a general foundation for crowdsourced art, creating a useful
typology for crowdsourced art and developing a nuanced framework of participation that
illuminates how the formal structure of a participatory project allows for different levels
or degrees of creative engagement.
!4
Subsequently, on an empirical level, I aim to investigate, through an experimental
design, the incentive structures that shape creative collaboration in online environments.
My principal interest in this endeavor is to understand the conditions that most effectively
foster online creative participation. Specifically, I am interested in the impact of four key
factors (amount of reward; possibility of bonus payment; stated purpose of the project;
attribution of authorship credit) on the quantity and quality of the resulting submissions,
the time spent by participants on the task, and the number of rebellious (i.e. off-task)
contributions. The explication of these relationships constitutes the vital empirical thrust
of the current research.
Another empirical question – albeit of secondary importance – concerned the self-
stated motivations that drive contributors to engage in creative activities online. The
idiosyncrasies of the online context of crowdsourced art are reflected in the wide array of
potential interaction modes (in terms of both group protocols and technological tools),
but also in the specific motivations that draw participants to such projects. Whereas in
offline relational art, the incentive to participate is dependent on one’s physical presence
at the site of the art project (often serendipitously, without prior knowledge of the artists’
plans or even the existence of a project), in crowdsourced art, the participants’
motivations are more deliberate. So far, the nature of these incentives has not received
any scholarly attention. Thus, a secondary goal of this study was to better understand the
motivations that stimulate participation in crowdsourced art.
Last but not least, beyond these research questions, a key goal driving this
endeavor – and driving my research in general – is the desire to enable creative
!5
expression, and to extend open opportunities for creative participation and collaboration.
This initiative was born out of my belief that creativity should not be - and was never
meant to be - the prerogative of a few chosen individuals. Nor should creativity be
limited to the institutional space of museums, galleries, and art fairs. The Internet
provides a unique platform to engage a potentially global community in the creation of
collaborative art, and I think there is something undeniably special about
that. Furthermore, I deliberately operate with a broad definition of art - in my mind, there
should be no elitist separation between art and creativity. Art is creativity, and creativity
can turn into art. I cherish collaboration and openness, and I strive to facilitate collective
experiences that are fundamentally participatory, creative and meaningful.
Theoretical Framework
Although the online medium functions as a platform for the creative production of
art, crowdsourced art is significantly different from net art or individually-created new
media art, since the aesthetic impulse here is first and foremost social. As such,
crowdsourced art is rooted in a rich heritage of artistic participation and collaboration,
and its emergence can be understood within this aesthetic context as a digital embodiment
of relational aesthetics. However, in Bourriaud’s original formulation (2002), relational
art refers to interpersonal, embodied, face-to-face interactions, and does not account for
technologically-mediated participation. If Bourriaud views relational art as rooted in the
“sphere of human interactions” (p. 14), crowdsourced art refers to the sphere of human
relations as mediated by technology (Literat, 2012); indeed, the novelty and importance
!6
of the current research stems partly from the expansion of the relational aesthetics
literature to online contexts. The role of the Internet as an enabling platform for the
production of collaborative art carries vital implications in terms of the interactions
between the participants: “significantly, because there is no physical encounter in online
crowdsourced art, the social aspect of artistic co-production is to be found in the formal
(digital) fiber of the art project; form and socialization are necessarily interlinked, jointly
exposing the technological structure that facilitates and mediates this digital
encounter” (Literat, 2012, p. 2972).
Bishop (2006b) identifies three main motivations invoked by artists working
under the relational aesthetic model: activation (the stimulation of empowerment and
agency), authorship (the cessation of authorship in an effort to make the artistic process
more democratic and egalitarian), and community (the restoration of weakened social
bonds through the collective elaboration of meaning). This last motivation is of particular
interest and consequence. Bishop (2004) notes that the conceptualization of the activated
audience as an ad-hoc community is a core feature of relational art: “rather than a one-to-
one relationship between work of art and viewer, relational art sets up situations in which
viewers are not just addressed as a collective, social entity, but are actually given the
wherewithal to create a community, however temporary or utopian this may be” (p. 54).
Bishop (2006b) also makes a useful distinction between participatory art in the relational
aesthetics tradition and, respectively, interactive art. She posits that the former is
concerned with the social dimension of participation, rather than the physical activation
of the individual viewer that lies at the core of interactive art. The aim of participatory
!7
art, and of relational aesthetics in particular, is thus to “collapse the distinction between
performer and audience, professional and amateur, production and reception” (p. 10).
Nonetheless, in spite of the progressive and prosocial ambitions of participatory
art, several critics have addressed the lack of nuance and objectivity within the utopian
impetus of relational aesthetics. As Beech (2008) argues, while the artists working in this
vein may have sincere and socially significant motivations, participation should not be
glorified as a means to solve a whole range of cultural problems (from art world elitism
to audience passivity to broken social bonds). In light of the goals of participatory art,
and given the micro-utopian ethos that Bourriaud identified in relational aesthetics, such
art is seen, almost by default, as political and emancipatory. The actual quality of the
relationships produced within relational art projects is rarely brought into question
(Bishop, 2004; Foster, 2006).
Empirical Research Design
On an empirical level, to answer my principal research question regarding the
factors that most effectively foster creative participation online, I used Amazon’s
Mechanical Turk platform to crowdsource the development of a children's book about
digital culture (a children's book about the Internet, by the Internet!), collecting 4200
!8
creative contributions (2100 written submissions and 2100 illustrations). In each
1
experimental condition, I modified a key variable in order to determine, through
statistical analysis, the impact that the following incentives have on the process of
creative participation online:
•
Amount of financial reward (5 cents, 10 cents, or 20 cents)
•
Provision of supplemental bonus for highest-quality submissions ($1 bonus,
$5 bonus, or no bonus)
•
Stated purpose of the project (charity, commercial, or no mention)
•
Attribution of authorship (authorship will be credited, authorship will not be
credited, or no mention).
In terms of dependent variables, my goal was to assess the impact of the above
variables on the following measures:
•
Quantity of submissions (operationalized as the time it takes for all
assignments to be fulfilled)
•
Quality of submissions (conceptualized as usefulness and, respectively,
novelty)
•
Time spent on task
•
Number of rebellious contributions
In addition to Mechanical Turk, my initial plan was to crowdsource the children’s book on two
1
other online platforms: hitRECord and Peer 2 Peer University (P2PU). By launching and studying
the same creative participation project on three different online platforms, I hoped to achieve an
insightful comparative analysis of the social and technical features that determine the nature and
quality of online creative participation. Unfortunately, in spite of a concerted effort, the attempts
to crowdsource the book on hitRECord and on P2PU failed to elicit sustainable participation.
Chapter 6 describes both implementations and discusses the key conclusions that emerged from
those experiences.
!9
I hypothesized that the quantity and quality (i.e. usefulness and novelty) of the
submissions, as well as the time spent on task would increase when:
a. there is a potential bonus payment.
b. the stated purpose of the task is charity, and will decrease when the
stated purpose is commercial.
c. there is the possibility of gaining credit for one’s work, and will
decrease when there is no attribution of authorship credit.
d. the amount of the financial reward is greater, and will decrease when
the reward is lower.
In regards to the number of rebellious submissions, I expected it would increase
when:
a. there is a potential bonus payment.
b. the stated purpose of the task is commercial, and will decrease when the
stated purpose is charity.
c. there is no attribution of authorship credit, and will decrease when there
is the possibility of gaining credit for one’s work.
d. the amount of the financial reward is lower.
Each task on Mechanical Turk also contained an optional open-ended question
about the participants’ motivations for selecting that particular task. These responses
were also coded and categorized. In regards to participant motivations, I expected that
!10
participation would be motivated primarily by the fun and unusual nature of the task,
rather than the monetary rewards.
The choice of using a children’s book as the cross-platform creative project for
this study is both practical and personal. On a practical level, a children’s book is a non-
intimidating creative text of moderate length, which facilitates both written and visual
contributions. But beyond that, on a personal level, I am passionate about children’s
books and I am passionate about the positive potential of the Internet; looking at
published children’s books on digital culture, it is overwhelming how negatively-framed
they all are. Don’t talk to strangers on Facebook, don't give our personal information,
don’t download pirated material, don't… don't… don't. Conversely, I wanted to develop a
children’s book that tells kids about the positive potential of the Internet, especially the
ability to find community and connect with people from all over the world. And since the
book would be written and illustrated entirely by online participants, it seemed fitting that
the subject matter would be thematically connected to digital culture and online
communities.
The finalized book will be available in two forms: as a free online e-book, and as
a printed book for those that want to purchase it, with all profits going to The Modern
Story, a grassroots NGO that implements digital storytelling programs for
underprivileged children in Indian public schools. I had the pleasure of working as the
field coordinator of this program in India, before starting my PhD studies, and I can attest
to the value of this program and to the fact that any money raised with this book would
directly benefit the children and their schools. In addition, given this organization’s
!11
domain of activity, they were an ideal candidate for receiving the proceeds. In a sense, the
prosocial circle seemed complete, with a children’s book about digital culture, written
and illustrated through digital participation, supporting digital literacy education for
underprivileged children.
Significance of the Research
Situated at the intersection of art, education, and communication research, this
research makes a vital contribution to our understanding of participatory art,
crowdsourcing and collective creativity. From a theoretical perspective, the present work
lays the foundation for a more nuanced conceptualization of online creative participation.
Significantly, it will also help gain a better understanding of participatory art in online
contexts, thus making a tremendously valuable contribution to the literature on
participatory art and relational aesthetics, which has focused so far on embodied, face-to-
face interactions. By extending this body of literature to online art, the current research is
also building a much-needed bridge between art history and Internet studies – two
disciplines that have insufficiently been put into conversation with one other, and which
would benefit from a more harmonious integration.
From a practical perspective, my findings will help determine the conditions that
most effectively foster creative participation online. The outcome of this investigation
will thus be useful not only to theorists and scholars, but also to companies, artists and
practitioners who are considering relying on public participation for their creative
projects. Furthermore, the secondary research question will help illuminate the
!12
motivations that drive participants to engage in the collective production of online art,
given the lack of financial incentives. Determining participant motivations has been an
important goal of recent crowdsourcing scholarship (Brabham, 2010; Chandler &
Kapelner, 2010, Zheng, Li & Hou, 2011), so the hope is that the current project may shed
light on these incentives from a novel perspective.
In light of the various implications described above, this study will be a valuable
cross-disciplinary investigation with important artistic, socio-cultural, and educational
outcomes. While modernism implied a shift from the public to the private, relational art
can be seen as a reversal of this dynamic (Bishop, 2004). In this sense, crowdsourced art
is an important harbinger of the (re)valuation of public participation in the cultural
sphere, with ramifications that stretch beyond the confines of the art world. Furthermore,
its reliance on the Internet as a connecting platform makes an enticing claim about the
current and future role of communication technologies in facilitating community, global
dialogue, and collective creativity. In the words of Nicolas Bourriaud (2002), “art was
intended to prepare and announce a future world: today it is modeling possible universes”
(p.13). If so, the universe that crowdsourced art is modeling is a compelling one:
fundamentally open, participatory, and creative, yet yearning for meaningful social
connections.
!13
CHAPTER 2:
INTERROGATING PARTICIPATION ACROSS DISCIPLINARY BOUNDARIES
Although the concept of participation is an increasingly ubiquitous buzzword
across a wide variety of disciplines, the specifics of its scope and applications are
insufficiently interrogated. As Delwiche and Henderson (2012) remind us, participatory
culture studies are, at heart, an interdisciplinary project. A comparative consideration of
specific understandings of participation across social domains (in this case: cultural,
political, artistic) emerges, therefore, as a highly valuable theoretical endeavor. Moreover,
Carpentier (2009) is right in pointing to the ideological nature of the struggle to define
and evaluate participation. Indeed, this attempt is more than a semantic project; what is at
stake is the dynamic redefinition of social relations, which carries both practical and
ideological implications. As Henry Jenkins (2013) recently observed, “it becomes more
and more urgent to develop a more refined vocabulary that allows us to better distinguish
between different models of participation and to evaluate where and how power shifts
may be taking place” (p. 5). A key conclusion that emerged from my own research is that
participation cannot be conceived as a blanket term, and that it is vital to interrogate the
different degrees or levels of meaningful participation, as they are understood in different
social domains. This chapter will therefore examine the conceptualization of participation
across three domains - political, cultural, and artistic - considering, in each of these cases:
a) the definition and key debates around participation, b) the degrees of meaningful
!14
participation, and c) the strengths of each of these perspectives, particularly in terms of
opportunities for theoretical cross-pollination.
Political Philosophy
Definitions and key debates
(Re)Rooting the concept of participation in the literature of political philosophy,
Carpentier (2011a) defines participation as “a political - in the broad meaning of the
concept of the political - process where the actors involved in decision-making processes
are positioned towards each other through power relationships that are (to an extent)
egalitarian” (p. 31). Significantly, he distinguishes participation from the related concepts
of access and interaction, which he considers as its conditions of possibility. In other
words, access and interaction are necessary for participation to occur, but are distinct
from it because they place less emphasis on power dynamics and decision-making
(Carpentier, 2011b). Although one must feel “invited, committed and/or empowered” to
take part in a participatory process, “the presence of a participatory culture cannot be
conflated with participation itself and its logics of equal(ised) power
relations” (Carpentier & Dahlgren, 2011, p. 8). In the political sense, then, participation is
seen to have a strong material or actionist dimension (Carpentier & Dahlgren, 2011),
which cannot be separated from its ideological conceptualization.
A broad understanding of the political is at the core of Carpentier and Dalhgren’s
arguments. Following the path laid by Chantal Mouffe and Carole Pateman, Carpentier
(2011a) posits that we cannot and should not confine the political (especially its
!15
implications for power and decision-making) to the realm of institutionalized politics.
Rather, “all social spheres are the potential objects of claims towards democratization and
increased participation, although these claims (and the struggles provoked) do not lead
necessarily to their realization, and the resistance in some societal realms turns out to be
more substantial than in others” (Carpentier, 2011a, p. 24).
Carpentier’s definition of participation is heavily influenced by the work of
Carole Pateman (1970), who defines participation as influence in or (even) equal say in
decision-making processes. Relying on a theoretical foundation laid by Rousseau, Mill,
and Cole, Pateman’s theory of participatory democracy posits that a participatory society
is a necessary prerequisite for the existence of a truly democratic political system. In
order for maximum participation to be feasible, “social training” for democracy must take
place in spheres other than the political realm. Importantly, these alternative spheres -
which she views as “political systems in their own right” (p. 43) - include the domain of
arts and culture, although Pateman’s focus is primarily on “industry”, i.e. one’s work
environment. Since individuals participating in these alternative spheres are being
socialized for democracy, Pateman views the process of participation as a quintessentially
educative one. She concludes, “the major function of participation in the theory of
participatory democracy is therefore an educative one, educative in the very widest sense,
including both the psychological aspect and the gaining of practice in democratic skills
and procedures” (Pateman, 1970, p. 42).
!16
Degrees of participation
Compared to discussions of participation in cultural studies and art, political
philosophers have been relatively quick to indicate the vagueness of the term, and to
identify the need for a more nuanced understanding of the various degrees or levels of
participation. This may be explained, at least in part, by the fact that participation in
politics is comparatively more normative, and, given the goals involved in this process,
the stakes are often higher than they are in other domains.
In an influential 1969 article, Arnstein introduced her so-called “ladder of
participation,” which identifies eight types (or “rungs”) of participation: manipulation,
therapy, informing, consultation, placation, partnership, delegated power, and, finally,
citizen control. Out of these, the first two are non-participatory, and only the latter three
rungs constitute - or have the potential to constitute - citizen power. Pateman (1970)
distinguishes between partial participation and full participation. In her view, partial
participation is “a process in which two or more parties influence each other in the
making of decisions but the final power to decide rests with one party only” (p. 70), while
full participation occurs when “each individual member of a decision-making body has
equal power to determine the outcome of decisions” (p. 71).
Building on Pateman’s work, Carpentier (2009) makes a useful distinction
between minimalist and maximalist forms of democratic participation. Under the
minimalist model, participation is principally understood as political representation by
means of elections, and the concept of politics is limited to the realm of institutional
decision-making. On the other hand, under the maximalist model, participation is seen as
!17
a combination of representation and participation, and the concept of politics is
deliberately broad, allowing for an increased range of participatory opportunities across a
variety of social sectors.
Strengths and opportunities
Because Carpentier, Dahlgren, and Pateman operate with a very broad concept of
the political, these perspectives are easily - and usefully - applied to cultural or artistic
contexts as well. In particular, the way political philosophers frame participation is
valuable because it foregrounds the concept of power, which is often critically missing
from discussions of participation in cultural or artistic contexts. As participation is
becoming a ubiquitous buzzword across an increasingly diverse range of contexts, it is
crucial to keep in mind that the concept is rooted in a rich tradition of political philosophy
and “is strongly related to the power logics of decision-making, whether it is explicit or
implicit, formal or informal, minimalist or maximalist and egalitarian or not” (Carpentier
& Dahlgren, 2011, p. 8). In fact, one of the main criticisms of the cultural studies
perspective on participation - and, in particular, of Jenkins’ Convergence Culture (2006a)
- has been that it pays too little attention to the relationship between cultural and political
participation (Couldry, 2011). Responding to Couldry’s criticism, Jenkins (2013) admits
that he has long been “frustrated with the inability of cultural scholars to demonstrate
strong links between ‘cultural resistance’ and more ‘institutionalized’ forms of
politics” (p. 20). A stronger integration of the concepts of power and decision-making
within discussions of participation in cultural studies represents - together with the
!18
educative function of participation, discussed below - a potential bridge between the
cultural and political realms.
In addition, a more pronounced emphasis on power dynamics allows scholars to
better evaluate the outcomes of participation, and to determine whether these
participatory processes actually change the status quo. In the words of Sherry Arnstein,
“participation without redistribution of power is an empty and frustrating process for the
powerless. It allows the powerholders to claim that all sides were considered, but makes
it possible for only some of those sides to benefit” (Arnstein, 1969, p. 216). This
observation is very valuable when applied to cultural or artistic contexts, and it mirrors
my own observations in regards to the sociocultural impact of crowdsourced art: while
participation is a progressive development and has been received with much enthusiasm,
most often it doesn’t fundamentally challenge the underlying power structures that define
the world of fine art (Literat, 2012). Similar observations have been made in regards to
participation in cultural production, especially in view of the strong influence that
commercial media stakeholders continue to wield in this domain (Jenkins, Ford &
Greene, 2013; Schaefer, 2011).
Another lesson to be learned from the political philosophy literature is in regards
to the educative function of participation. As Pateman (1970) argues, we learn to
participate by participating: “participation develops and fosters the very qualities
necessary for it” (p. 42-3). From this perspective, it is easy to see a connection to the way
that Henry Jenkins applies notions of participatory culture within the context of education
(Jenkins et al., 2006) - and indeed, the educational realm is another space in which an
!19
important battle is waged over the meanings and normative implications of participation.
Both Pateman and Jenkins subscribe to the notion that there are certain participatory
skills that need to be cultivated, especially in youth. Even the language they employ is
similar, with Jenkins defining new media literacies as “cultural competencies and social
skills” (Pateman: “participatory skills”) needed for “full participation” (Pateman:
“maximum participation”) in the contemporary cultural landscape.
Cultural Studies
Definitions and key debates
Whereas the first wave of writing on new media cultures relied on the notion of
interactivity to describe the shift in audience-media relations, contemporary theorizations
have re-embraced the concept of participation as a key cultural descriptor (Carpentier,
2012). Jenkins, the foremost investigator of participation in the cultural studies field,
provides a definition anchored in his long-term interest in fan practices: he suggests that
“a participatory culture is one with relatively low barriers to artistic expression, and
strong support for creating and sharing one’s work with others; its members believe that
their contributions matter, and feel some degree of social connection to one
another” (Jenkins et al., 2006, p. 3). He also makes a crucial distinction between
interactivity and participation, with the former emerging from technological design, while
the latter stems from social and cultural practices (Jenkins & Carpentier, 2013).
In Bastard Culture!, Schaefer (2011) argues that participatory culture should not
be reduced to its symptoms - meaning, primarily, the activity of taking part in processes
!20
of cultural production and distribution. Instead, he suggests that the concept is actually “a
complex discourse consisting of the following factors:
a. a rhetoric that advocates social progress through technological
advancement
b. a cultural critique demanding the reconfiguration of power relations
c. the qualities of related technologies
d. how these qualities are used for design and user appropriation
e. the socio-political dynamics related to using the technologies” (p. 14).
As it becomes apparent from this conceptual mapping, Schaefer places a much
greater emphasis on the role of technology in enabling or limiting participation, while
Jenkins is primarily concerned with the social and cultural factors that determine the use
of technology for participatory purposes.
Indeed, the relationship between participation and technology has been a vital -
and most often fruitful - debate in cultural studies. For one, Jenkins (2006a, 2006b, 2013)
has always tried to steer clear of the trap of technological determinism. He acknowledges
that, beyond new communication technologies, there are a multitude of other factors that
have paved the way for the rise of participatory cultures. In Fans, Bloggers, and Gamers
(2006b), he identifies three major trends that explain the participatory ethos of
contemporary media cultures: the emergence of new tools and technologies that facilitate
participatory activities; the increasing prominence of DIY approaches to media
production, particularly in subcultural communities; and an economic environment
encouraging the active production and circulation of media across multiple channels.
!21
Similarly, in Spreadable Media (2013), Jenkins and his co-authors caution against the
assumption that "a more participatory means of circulation can be explained solely (or
even primarily) by this rise of technological infrastructure" (p. 3). While these
technologies surely play a key role in the processes they describe, the authors' interest -
and the strengths of their analysis - lie in their discussion of the social dynamics and
cultural practices that have shaped the development and popularization of these
platforms.
Degrees of participation
Given the proliferation of participatory and pseudo-participatory opportunities
within the new media landscape, theorists in the cultural studies tradition have been
increasingly preoccupied with determining what constitutes meaningful cultural
participation. Often, this question becomes framed as an inquiry into the lower limits of
participation. In a recent dialogue, Jenkins and Carpentier (2013) have debated whether
interpretation, production, curation, and circulation can be seen as potentially meaningful
forms of participation. In Carpentier’s view, the answer is yes, with the exception of
interpretation, which he argues is a form of interaction. Jenkins, however, in light of his
extensive study of fan cultures, is more amenable to the participatory potential of
interpretation, although he admits that, in and of itself, interpretation does not necessarily
constitute participation. Given the differences in their disciplinary backgrounds and
theoretical lenses, the dialogue between Jenkins and Carpentier is an extremely
productive one, and their points of disagreement are as interesting as their intersections.
!22
Their conversation also reveals that sometimes it depends how one looks at participation.
As Fish et al. (2011) aptly note, in many cases the same online action or behavior can be
read as participatory or not, depending on one’s particular viewpoint; often, this
determination carries important ideological implications in terms of agency and
empowerment.
Also in regards to the lower limits of participation, a more contentious debate is
whether participation must be a deliberate act. Some cultural commentators (Schaefer,
2011; Kelty, 2012) argue that it does not necessarily have to be deliberate. In Bastard
Culture!, Schaefer (2011) distinguishes between implicit and explicit participation,
although he considers both to be forms of participation. In his view, explicit participation
represents ‘’intrinsically motivated actions exercised in social formations which share a
high degree of interaction, common objectives, and interests” (p. 44). Implicit
participation, on the other hand, occurs on “a rather subliminal level”, when “users are
participating – often without acknowledgement from the companies offering such
services – in shaping and expanding the information infrastructure” (p. 44). In his review
of Schaefer’s work, Carpentier (2012) argues that the consideration of implicit
engagement as a form of participation is problematic, because it makes it difficult to
determine the limits of participation, or the instances in which participation stops being
participation. In light of my own research, I share Carpentier’s more skeptical view and
consider implicit participation to be an oxymoron. Moreover, I fear that such a
conception threatens to deemphasize the sociocultural meaningfulness of participation,
especially in terms of agency, empowerment, and efficacy.
!23
This potential de-emphasis would partly be due to the fact that implicit
participation does not carry the same positive connotations that explicit participation
does. Indeed, implicit participation has often been framed as free labor or exploitation
(Andrejevic, 2009; Fish & Srinivasan, 2012; Terranova, 2000). This is particularly true in
recent times. In fact, a key stimulus behind Spreadable Media was the authors’
disappointment with how corporate forces misconstrue or, worse, combat fan engagement
and participatory culture. Jenkins et al. (2013) caution that “the companies that will thrive
over the long term in a "spreadable media" landscape are those that listen to, care about,
and ultimately aim to speak to the needs and wants of their audiences as crucially as they
do their own business goals” (p. xii). An interesting question is whether these implicit
participants are really that oblivious regarding the moneymaking value of their clicks,
hits and likes. Banks & Deuze (2009) note that scholars writing in the legacy of
Terranova often position participants as unaware that their participation is an
economically-valuable practice; this approach is, according to the authors, patronizing
and often untrue.
Finally, another key distinction in regards to the degrees of participation is
Carpentier’s (2011b) discussion of participation through the media versus participation in
the media. Participation through the media “deals with the opportunities for mediated
participation in public debate and for self-representation in the variety of public spaces
that characterize the social” (p. 520). In contrast, participation in the media concerns the
involvement of non-professionals in media production (content-related participation) and
media decision-making (structural participation). Although, from the perspective of
!24
agency and power, the latter is clearly a more valuable form of engagement than the
former, Carpentier notes that participation through the media can facilitate important
participatory opportunities as well, depending on the particular structure of decision-
making processes and power dynamics involved (Jenkins & Carpentier, 2013).
Strengths and opportunities
As even this brief theoretical overview indicates, cultural studies scholars are well
attuned to the nuances of participation, and have had a particularly fruitful investigation
of the lower limits of engagement. In art, especially, participation is still very much a
blanket term; art scholars would be well served by a more refined investigation of the
various degrees and types of engagement. For instance, the mere attendance at cultural
events or museums is conceptualized as participation (Jenkins & Bertozzi (2007)
represents a notable exception). While this may simply be a semantic idiosyncrasy, the
notion of participation as attendance prevents the development of a more productive and
nuanced dialogue regarding the various types of participatory opportunities available,
particularly in regards to collaboration and content-related participation.
2
Discussions of online art also fall prey to a fallacy in regards to participation, albeit this is a
2
different kind of mislabeling. Specifically, digital art projects that rely on implicit participation
are regularly being described as “participatory” or “crowdsourced” when in fact they are not. For
example, Jonathan Harris’ project We Feel Fine uses innovative data visualization techniques to
display the “feelings” of bloggers across the world: it crawls the web, looking for instances of the
word feel/feeling/felt etc, and then displays these feelings and makes them searchable by location,
gender, type of emotion, and even local weather. This is not a crowdsourced nor participatory
project, as it has often been described in the media, because the bloggers in question are not
deliberately participating. Rather, it should be seen as data mining, made beautiful with the aid of
information visualization techniques.
!25
Another lesson to be learned from the cultural studies approach is the need to pay
closer attention to the role of technology vis-à-vis participation. This necessarily includes
a more objective examination of the positive rhetoric surrounding technology as an
emancipatory and equalizing factor in the facilitation of more widespread participation.
Nonetheless, although my general view on participatory cultures is definitely not one of
technological determinism, I believe that when it comes to art, the role of the Internet as
an enabling platform - from a strictly pragmatic perspective - is comparatively more vital.
This is primarily because the non-digital spaces that may otherwise facilitate participation
in art are, in most cases, rather elitist and do not sufficiently challenge the status quo. As
the history of offline participatory art indicates, prior to the popularization of the Internet,
these participatory artworks usually took place in museums or galleries (with site-specific
art being an important exception). This tradition of embodied engagement continues, as
seen in the recent works of Marina Abramovic (who, in her popular performance The
Artist Is Present, invites museum goers to sit across from her and stare into her eyes) or
Rirkrit Tiravanija (whose project Pad Thai invites gallery visitors to enjoy a Thai meal
that he cooks in front of them); however, the physical site of engagement remains the
museum or gallery space, and the restrictions that such spaces pose in terms of cultural
hierarchies, openness and inclusivity are significant parameters that limit widespread
engagement.
Finally, the literature from cultural studies models a more collective
conceptualization of participation, which other fields could benefit from. In art, for
instance, the locus of empowerment and agency is most often identified at the individual
!26
level. Perhaps this is because art is a comparatively more intimate experience, and there
is an inherent individualism embedded in both the process of artistic production (the
paradigm of the artist as lone genius) and reception (the paradigm of the viewer silently
contemplating the work in front of him, almost as if in a vacuum). Such an understanding
of participation can, therefore, be highly enhanced by Jenkins’ focus on the collective
dimensions of participatory cultures. Jenkins (2013) writes, “while, certainly, individuals
do gain some greater degree of communication capacity as a result of their access to new
tools of production and circulation, my primary emphasis is on ‘collective agency’, that
is, the capacity of networked publics to work together towards shared goals and common
interests (p. 18). Interestingly, as a counterpoint to this collective model, the members of
the Janissary Collective (2012), led by Mark Deuze, claim that participation has an
individualizing effect. They argue that “participatory culture spotlights the individual in
ways a traditional, premodern culture could not” (p. 260) and that “we participate by
shouldering individual responsibility for what used to be collective endeavors” (p. 259).
Although I disagree, I find their argument and their examples to be very intriguing and
worth further philosophical exploration.
!27
Art
Definitions and key debates
Within the fields of art history and criticism, participation is understood most
often through the prism of relational aesthetics. In what is considered to be one of the
3
defining texts of contemporary art criticism, Bourriaud (2002) introduced the concept of
relational aesthetics as an artistic form that values the social above the material. In other
words, the artist in relational art projects functions as the facilitator of a social experience
rather than as the creator of a tangible object. Such a perspective generally implies a
micro-utopian ethos; indeed, for these artists, “it seems more pressing to invent possible
relations with our neighbors in the present than to bet on happier tomorrows” (Bourriaud,
2002, p. 45). While Bourriaud wrote about relational art as limited to the realm of face-
to-face interactions, my research has extended the concept of relational aesthetics to
apply to digital art as well, where the social interaction between participants is mediated
by technology (Literat, 2012). Chapter 3 provides a detailed description of crowdsourced
art and its contextualization within the relational aesthetics scholarship.
In an interesting testament to the divergence and conflict that characterize
participatory art criticism within contemporary art history, Bishop (2006a) enumerates
the variety of terms used to describe the participatory impetus identified in recent art:
such projects are interchangeably dubbed “socially engaged art, community-based art,
experimental communities, dialogic art, littoral art, participatory, interventionist,
In light of my comments above, my understanding of participation in art is in the active sense of
3
collaboration and co-creation. Therefore, my discussion below does not include a consideration of
participation as mere attendance, consumption, or interpretation.
!28
research-based, or collaborative art” (180). Although the concept of relational aesthetics
is used almost universally when discussing such works, Bishop remarks that the artists
themselves are less interested in the aesthetics aspect (the form of the work), than in the
creative rewards that come from opening up the artistic process. The aim of participatory
art, and of relational aesthetics in particular, is thus to “collapse the distinction between
performer and audience, professional and amateur, production and reception” (Bishop,
2006b, p. 10).
With a more cynical eye, Beech (2008) criticizes contemporary art for illustrating
a “declining ambition for the politics of participation” (p. 1). Relying on recent examples
from relational art projects, wherein the mere act of having dinner or drinking beer is
seen as artistic participation, he wonders whether we are witnessing an erosion in the
political prerogative of participatory art. He warns that, in art just like in politics,
“participation cannot deliver what participation promises,” since it represents, first and
foremost, an utopian ideal of social reconciliation, and not a mechanism to achieve this
ideal (p. 2). What is more, the participation of non-artists (or publics) in the production of
artwork does not actually challenge the cultural distinctions between artist and audience;
rather, it merely “re-enacts that relationship in an ethnographic fashion” (p. 2).
Degrees of participation
Beech (2008) makes the excellent argument that participation in art is most often
very limited: entering a participatory art project is akin to accepting the parameters of a
contract that one cannot modify, critique, or subvert. My own research on participatory
!29
art in online contexts supports this view, highlighting the consequential distinction
between participation and collaboration in terms of the ability to affect conceptual or
structural change (see Chapter 3). In a sense, this distinction also highlights the
dichotomy between agency and choice. Participatory art projects present choices of
action and engagement, but when the design of the project is pre-established by the
project initiator (the so-called “alpha artist”) prior to any public involvement, the
potential contributors are not given the prerogative to exercise their creative agency.
Thus, “although any type of structure – artistic of otherwise – rests on a certain
combination of available choices, it is the ability to affect or provide these choices that
constitutes the mark of structural agency” (Literat, 2012, p. 2979).
A key conclusion that emerged my research is that participation should not be
used as a blanket term; such a conception is erroneous and unproductive, since it does not
account for the complexities of creative agency, artistic hierarchies, access, and capital. In
Chapter 3, I therefore suggest a more nuanced model of understanding the various levels
of engagement, which implies a breakdown of this concept into receptive, executory, and
structural modes of participation. Receptive participation refers to the process whereby
the viewer or audience receives a finished artistic product. Executory participation is the
task-based participation in a pre-designed artistic project: it is analogous to entering a
contract whereby the contributor accepts the parameters of participation, but has no
power to challenge or modify these parameters. According to their level of transparency
and the space allowed for personal expression, executory crowdsourced art projects can
be tokenistic, engaged, or creative. Finally, with structural participation, participants have
!30
a say in the conceptual and artistic design of the project; in other words, they are allowed
to demonstrate structural agency. This distinction between executory and structural
modes of public input has crucial implications in terms of the ethics of engagement,
marking the difference between participation (a task-based contribution to a pre-designed
project) and collaboration (structural input in the form of co-design).
Strengths and opportunities
Art is understood and written about in terms of the interplay between form and
content; this aesthetic bifurcation could be a valuable mode of inquiry when considering
the participatory opportunities that art can facilitate. Specifically, a conceptualization of
form versus content in regards to participation in other domains would allow for a more
nuanced interrogation of participatory opportunities, by attempting to locate them in both
of these dimensions. Scholars writing about participatory cultures should thus evaluate
both the quality of the process (form) of participation and the quality of the final product
(content) of participation. Indeed, a focus on quality is often missing in discussions of
participation in political philosophy and cultural studies. The quality of the consensus
arising from participatory political processes is rarely questioned (and indeed, can be
difficult to operationalize). The quality of the piece of fan fiction or remix is rarely
investigated, with scholarly and popular accounts often focusing on the best and most
impressive of these cases.
In fact, as several scholars note, one of the principal problems with the discourses
on participation is that it is almost always seen as a positive and empowering force
!31
(Carpentier, 2009; Fish et al., 2011; Jenkins & Carpentier, 2013; Kelty, 2012; Schaefer,
2011). This notion appears in the art world as well, but, given the unique aesthetic
prerogative of fine art, issues of aesthetic quality and, increasingly, of participatory
ethics, are adamantly - sometimes even aggressively - brought to the fore (Bishop, 2004;
Foster, 2006). In the realm of participatory art, critics are calling for the development of
aesthetic criteria by which to judge socially engaged art (Bishop, 2006a; Kester, 2004,
2011). There is an acute awareness that the task of aesthetic evaluation has been
overshadowed - or, in same cases, rendered insignificant - by the emphasis on the
political goals of such art, in the form of de-alienation, community-building and the
restoration of social bonds. I believe that a similar awareness in other participatory
domains, and a meticulous qualitative examination of both form and content (separately
and together) can help participation scholars arrive at conclusions that are more nuanced,
more grounded, and ultimately more pragmatic.
Last but not least, the arts can provide a blueprint for alternative ways to enable
participation, while reflecting the civic and creative benefits that such models can
facilitate. Artists can take risks in a way that political organizations and commercially-
minded media institutions cannot. In political systems, constraints to participation are
relatively stable and often difficult to challenge (Dahlgren, 2011). In media cultures,
commercial interests often keep participation at bay and deliberately oppose change and
(r)evolution (Jenkins et al., 2013). In art, by comparison, change is not only possible, but
indeed it represents a desirable feature of the system: art is made to be challenged,
questioned, reinvented. In art, change is applauded as a revitalizing force - and,
!32
significantly, this revitalization is increasingly framed in terms of opportunities for
participation and collaboration.
Conclusion
In closing, it is worth considering whether participatory culture is, perhaps, an
unreachable ideal. In my view, this question represented one of the most intriguing topics
of discussion in the conversation between Jenkins and Carpentier (2013), recently
published in Convergence. In this dialogue, Jenkins acknowledges that “participatory
culture, in any absolute sense, may be a utopian goal, meaningful in the ways that it
motivates our struggles to achieve it and provides yardsticks to measure what we’ve
achieved” (Jenkins & Carpentier, 2013, p. 2). Nonetheless, this realization does not
decrease the value of the conversation to be had around participation; rather, it makes this
dialogue more crucial and more pragmatic. The past couple of years have produced some
promising examples of interdisciplinary approaches to understanding participation (see
the conversations in the International Journal of Communication Forum: Participations,
issue 8/2014; also Delwiche & Henderson, 2012; Fish et al., 2011; Jenkins & Carpentier,
2013). The moment is ripe to continue inquiring about ways to maximize participation,
and we should strive to make this dialogue as interdisciplinary as possible.
!33
CHAPTER 3:
CROWDSOURCED ART AND COLLECTIVE CREATIVITY
This chapter lays a theoretical foundation for crowdsourced art, discussing key
examples and analyzing this artistic practice within multiple theoretical frameworks.
Significantly, I aim to fill a critical gap in the study of contemporary art and participatory
culture by developing a typology of online crowdsourced art and exploring the levels of
artistic participation. The chapter begins with an overview of crowdsourcing as a content-
generation strategy, then looks at illustrative cases of crowdsourced art across a variety of
artistic forms, and finally discusses the shifting roles of the artist and the public in the
digital environment. A key theoretical takeaway is the “pyramid of participation” which
reconceptualizes the notion of creative participation as passive, executory or structural
modes of engagement.
A Primer on Crowdsourcing
In order to fully understand the phenomenon of crowdsourced art, it is necessary
to start with a general discussion of crowdsourcing, tracing its roots and its principal
features as a cultural and commercial practice. In the words of Jeff Howe (2006b), the
Wired columnist who coined this term in June 2006, “crowdsourcing represents the act of
a company or institution taking a function once performed by employees and outsourcing
it to an undefined (and generally large) network of people in the form of an open
call” (n.p.). The vital elements that qualify an outreach strategy as crowdsourcing are,
!34
according to Howe, the use of the open call format and, respectively, the reliance on a
large network of potential workers (Howe, 2006b). Although in some cases there is a
material reward for the contributions deemed by the company or institution as most
valuable, the existence of financial incentives is not a required feature in crowdsourcing.
Crowdsourcing began as a corporate strategy of engaging customers in the
generation of creative content, innovation and brand development. Companies like Pepsi
or Doritos, for instance, asked users to film their own commercials, while Lego’s
DesignByMe campaign invited fans to create innovative Lego models, the best of which
then went into fabrication. Even the US Military launched an open call for the design of
an “experimental crowd-derived combat-support vehicle,” where participants have the
opportunity to “contribute to the future of military vehicle manufacturing” by submitting
their own designs, as well as voting on others’ submissions (DARPA, 2011, n.p.).
Online ventures – both commercial and non-commercial – built around the
practice of crowdsourcing have also abounded. Wikipedia has developed an impressive
knowledge database made up solely of user contributions; Threadless invites T-shirt
designers to submit their ideas, with users voting on the best designs to be produced and
sold on the site; iStockPhoto builds a database of user-submitted stock photographs and
pays the photographers each time their picture is bought and downloaded; Kickstarter is
based on the concept of crowdfunding, where artists and entrepreneurs submit their ideas,
and users pledge donations to their projects of choice. In addition to these fully-
crowdsourced web initiatives, there are also websites that are enhanced by
crowdsourcing, in the form of user-generated reviews and ratings. Some examples here
!35
are Netflix, Amazon, the Internet Movie Database (imdb.com), and other ratings-based
platforms, which Howard Rheingold (2002) calls “reputation systems”. Crowdsourcing is
also the foundation of multiple public databases, and has seen a particularly useful
application in the aftermath of natural disasters such as the earthquake in Haiti or the
recent tsunami in Japan, when locals built ad-hoc databases with information on found
survivors, emergency shelters or local radiation levels. Indeed, thanks to instrumental
websites such as Ushahidi.com, crowdsourced crisis-mapping has become a standard
element in earthquake response campaigns.
Because crowdsourcing provides companies with free or very cheap labor and
innovation but profits generally do not get distributed among all the contributors, this
practice has often been criticized as a form of labor exploitation (Schenk & Guittard,
2011; Brabham, 2008; Chandler & Kapelner, 2010). However, studies show that
contributors are equally motivated by incentives other than financial rewards. According
to management researchers at MIT’s Center for Collective Intelligence, the incentives
that motivate participants can be broadly classified as money, love, and glory, or a
combination of the three (Malone, Laubacher & Dellarocas, 2010). Researching the
community at Threadless.com, Brabham (2010) similarly found four principal motivators
for participation: “the opportunity to make money, the opportunity to improve one’s
creative skills, the opportunity for eventual freelance design work, and the love of
community at Threadless” (p. 1124). In a highly interesting study of the relationship
between task meaningfulness and motivation to participate on Amazon’s Mechanical
Turk, it was found that US workers were more inclined to perform a task if they
!36
perceived it as socially meaningful (in this case, labeling tumor cells to aid a
groundbreaking cancer treatment), but Indian workers were not influenced by the
perceived social utility of the task (Chandler & Kapelner, 2010).
The concept of crowdsourcing is too often used in a generic sense to account for a
wide range of phenomena and modes of interaction, so it is worth looking more closely at
the various types of crowdsourcing processes. Jeff Howe, the originator of the term,
initially suggested the categorization of crowdsourcing practices into four main types,
according to their primary function: crowd wisdom, crowd creation, crowd voting, and
crowd funding (Howe, 2008). Conceptualizing crowdsourcing as a model for problem-
solving, Brabham (2013) proposes a different typology, centered around four dominant
crowdsourcing types: the knowledge-discovery and management approach, the
broadcast-search approach, the peer-vetted creative-production approach, and the
distributed-human-intelligence tasking approach. A categorization that I find particularly
useful and straightforward is that suggested by Malone, Laubacher & Dellarocas (2010).
In their seminal Sloan Management Review article, the authors divide crowdsourcing
practices into three main types, according to the kinds of processes involved: collection
(when the activity can be divided into small pieces, which contributors work on
independently of each other), contest (when only one or a few good solutions are
needed), and collaboration (when an activity cannot be divided into small independent
pieces, and there is a feasible system of coordination among the contributors).
The contrast between collection and collaboration processes is particularly useful
in regards to crowdsourcing, as it marks the crucial difference between aggregative and
!37
deliberative modes of collective input. The collection model may include instances where
contributors are not aware of the final purpose or uses of their data (as in Schaefer’s
concept of implicit participation) and does not depend on communication or coordination
between the participants. Surowiecki’s notion’s of “the wisdom of crowds” (2004) is a
good representation of this approach, illuminating the power of aggregating large
amounts of data from diverse contributors working independently from each other. In
contrast, the collaboration model requires a level of coordination among the contributors,
who share their knowledge and insights with one another. The value of collaboration
models lies in the collective intelligence of the contributors, which Pierre Levy (1997)
describes as “a form of universally distributed intelligence, constantly enhanced,
coordinated in real time, and resulting in the effective mobilization of skills” (p. 13). The
distinction between collection- and collaboration-based crowdsourcing models is highly
applicable in creative or artistic contexts too; its usefulness will become apparent when
introducing the typology of crowdsourced art and the different levels of creative
participation.
Nonetheless, if we admit that crowds can be intelligent, can they also be creative?
Artistic creativity is different from innovation and intelligence, and requires a unique set
of skills and sensibilities. Joseph Beuys, originator of the social sculpture theory,
famously claimed that everybody is an artist, but this artistic potential gets stifled by
social norms and pressures. Far from advocating – as some mistakenly interpret his well-
known adage – that everyone should be involved in the creation of art, Beuys believed in
the creative abilities of people in their everyday life, and advocated for a more extensive
!38
and inclusive reconceptualization of art (Mesch, 2007). However, other critics are more
skeptical, doubting the creative potential of the crowd. Dwight McDonald (1952), in his
discussion of “masscult" and “midcult,” claims that “a mass society, like a crowd, is
inchoate and uncreative” (p. 9). Nick Carr, a noted social technology dystopian, provides
a more nuanced, yet equally doubtful, view: “I think what we're discovering is that big
online groups are very good at performing time-consuming, fairly routinized tasks that
can be broken up into many discrete units of work…But if you're looking for the new, the
creative, the moment of blazing insight, you're still going to have to look not to a crowd
but to an individual human mind” (Carr, quoted in PBS Frontline, 2010, n.p.).
Howard Becker (1974), writing on art as collective activity, provides a more
practical viewpoint that is perhaps the most relevant observation within this current
discussion of crowdsourced art. In capitalist societies, Becker suggests, “people keep the
idea that artists have a special gift but do not believe that there is any way to tell who has
it outside of letting everyone try and then inspecting the results” (Becker, 1982, p. 16).
And that is as good of a place as any to begin our foray into the realm of online
crowdsourced art.
Online Crowdsourced Art: Cases and Typology
Online crowdsourced art is the practice of using the Internet as a participatory
platform to directly engage the public in the creation of art. Although creative
participation is enabled by – and dependent upon – web access, this artistic strategy is
significantly different from other types of digital art such as individually-created net art
!39
or data visualization, since it relies on the collective input of a large pool of contributors
in the form of an open call – the same characteristics which represent the two principal
features of crowdsourcing.
To better understand the scope of online crowdsourced art, let us turn to a few
paradigmatic examples of this artistic practice. Within the field of visual arts, digital artist
Aaron Koblin – who is also Creative Director of Google’s Data Arts Team in San
Francisco – is, arguably, the lead innovator in online crowdsourced art. Explicitly
referencing the connection between art and crowdsourcing, Koblin often uses Amazon’s
crowdsourcing marketplace Mechanical Turk (www.mturk.com) to post creative “tasks”
that online workers can perform in exchange for a small fee. His most famous project,
The Sheep Market (www.thesheepmarket.com), paid workers on Mechanical Turk $0.02
to “draw a sheep facing to the left”, aggregating 10,000 examples of user-drawn sheep.
Similarly, Koblin’s Ten Thousand Cents (www.tenthousandcents.com) paid Mechanical
Turk users 1 cent to paint a tiny part of a $100 bill, working in isolation from each other
and without knowledge of the final task. The total labor cost to create the image of the
bill, as well as the reproductions available for purchase, are all $100. Koblin has also
ventured in the collective creation of an animated music video. His excellent Johnny
Cash Project (www.thejohnnycashproject.com), a collaboration with director Chris Milk,
invites participants to create a drawing that is then woven into a collective musical tribute
to Johnny Cash, set to his final studio recording, “Ain’t No Grave.” Made using a custom
drawing tool on the project website, each participant’s drawing represents one frame in
the animated music video. The interactive website allows visitors to see the collective
!40
end-product, but also to inspect individual contributions, frame by frame. The number
and diversity of contributions so far has been impressive, and the result is a moving
global homage to the beloved artist.
Koblin’s most recent crowdsourced art project is the aptly named This Exquisite
Forest (www.exquisiteforest.com), a collaboration with Chris Milk, The Tate Modern,
and Google Creative Lab. The title is a reference to the “exquisite corpse” technique, an
artistic practice developed by the Surrealists in the 1920s, which involved different
artists contributing to a larger piece of work in a game-like fashion, without seeing
previous additions. Similarly, Koblin and Milk’s project invites online visitors to create
short animations that build off one another, resulting in a collection of branching
narratives that – visually and poetically – resemble trees in a forest. An example of the
increasing penetration of online crowdsourced art projects within more traditional and
institutionalized spaces of art exhibition, This Exquisite Forest has been displayed at the
Tate Modern in London for six months in 2012.
Other notable visual arts projects in the field of online crowdsourced art have
been SwarmSketch, Wikipainting and Learning to Love You More (LTLYM). Each week,
SwarmSketch (www.swarmsketch.com) randomly chooses a popular internet search term,
which becomes the sketch subject for the week. Visitors can contribute one line to the
group illustration, and also vote on which lines will remain and which will be eliminated.
Wikipainting (www.wikipainting.free.fr) is similarly based on the idea of collaborative
painting, but gives participants free rein in terms of content, and also allows anyone to
modify the publicly-available source code, and thus improve the painting applet.
!41
Learning to Love You More (www.learningtoloveyoumore.com) is a collaboration
between artists Miranda July and Harrell Fletcher. Every week, the artists issued a call for
participation in the form of a school-type “assignment” like “Take a picture of your
parents kissing”, “Draw the news”, “Recreate an object from someone’s past”, “Draw a
constellation from someone’s freckles”, or even the meta-assignment “Make a LTLYM
assignment.” The participants submit their “reports” in the form of writing, drawings,
photograph, or video, and all the contributions are uploaded to the project site, under the
heading of each weekly assignment.
The recent trend of crowdsourced art has also found productive and ingenious
applications in cinema. Casey Pugh’s acclaimed mash-up Star Wars Uncut
(www.starwarsuncut.com) used fan-submitted media to crowdsource a remake of the
classic film Star Wars IV: A New Hope. Almost a thousand people participated in this
project, recreating the famous world of Luke Skywalker and Princess Leia with
everything from cats to fingers to costumed children to salt and pepper shakers.
Garnering very positive reviews, Star Wars Uncut won a 2010 Emmy Award for
Outstanding Creative Achievement in Interactive Media. Following the success of this
initial experiment, the creators of the project are now accepting submissions for The
Empire Strikes Back, the next episode in the Star Wars saga.
Cinematic crowdsourced art is not limited to independent or emerging
filmmakers; recently, even well-established producers and directors have capitalized on
the opportunity to engage their audiences and benefit from free user-generated content. In
2012, acclaimed Danish director Lars von Trier launched an open online invitation to
!42
contribute to his new film, Gesamt (www.gesamt.org), by submitting media clips inspired
by six famous works of art that von Trier selected. “What happens when a master
challenges the people? When everyone is invited to reinterpret some of the greatest art
pieces of our time? Is it within everyone to make art? And do we stand out the most when
we stand together?”, the director asks on the project website, commenting on the
aesthetic mission of Gesamt. Similarly, Life in a Day (http://www.youtube.com/user/
lifeinaday), a documentary project envisioned and produced by renowned Oscar-winning
filmmakers Ridley Scott and Kevin MacDonald, invited participants from all over the
world to film, on July 24, 2010, a brief video of their life on that particular day. The
5000+ hours of video submitted by users was then edited and condensed into a feature-
length documentary, which premiered at the Sundance Film Festival. Released in theaters
nationwide one year later (precisely on July 24, 2011), the film is now available for free
on YouTube.
Beyond the field of visual arts, there have also been several exciting applications
of artistic crowdsourcing in the production of sound and music. Aaron Koblin is a pioneer
in this area as well: his crowdsourced musical project, Bicycle Built for Two Thousand
(www.bicyclebuiltfortwothousand.com) is a collection of 2,088 voice recordings from
online workers assembled into the song “Daisy Bell.” Each participant was prompted to
listen to a short sound clip, and then record themselves imitating that sound, without
knowledge of the final task; put together, the sounds came together as “Daisy Bell.” The
Virtual Choir (www.ericwhitacre.com/the-virtual-choir), a similar project conceived by
acclaimed modern composer Eric Whiteacre, collected 2052 self-made recordings of
!43
people performing the same song. Whiteacre then recombined these sound clips into one
collective rendition, and uploaded the result back to YouTube.
Darren Solomon’s collaborative music and spoken word project In B Flat
(www.inbflat.net) further incorporates an element of interactivity, by allowing visitors to
the site to build their own symphony by play any or all of the user-submitted music jams,
at the same time and in any order. In addition, recently, a few famous music artists have
also joined the crowdsourcing trend, most notably Radiohead
(www.radioheadremix.com), who offer raw tracks for download and invite fans to remix
them for a new album.
Finally, there have also been a few notable applications of the crowdsourcing
process in the realm of literary arts. In comparison to other artistic forms, crowdsourced
literary projects are significantly more rare, due to doubts regarding the suitability of this
technique for the process of creative writing, and the recent failure of a series of
collaborative journalism projects, particularly the high-profile Assignment Zero. Among
these few projects, the online writing community Protagonize (www.protagonize.com)
has been one of the most successful endeavors. Protagonize allows members to work on
their own stories, as well as contribute to other’s narratives, but full membership is
reserved for paying subscribers. Another interesting example is filmmaker Tim Burton’s
Cadavre Exquis (www.burtonstory.com), an experiment in crowdsourcing a movie script
via Twitter. The project, whose title also referenced the “exquisite corpse” technique of
collective art-making, invited fans to contribute to the development of a collective plot
using the hashtag #BurtonStory. Every day, Burton chose one tweet that became the
!44
continuation of the story. The opening line? “Stainboy, using his obvious expertise, was
called in to investigate mysterious glowing goo on the gallery floor.”
In view of the diversity of online crowdsourced art projects, as illustrated by the
examples cited so far, it is useful to map out this artistic trend by developing a
comprehensive and multidimensional typology of online crowdsourced art. The following
table therefore represents an attempt to organize this classification according to a set of
multiple criteria.
Table 1: A Typology of Online Crowdsourced Art
Criterion Categories Examples
By medium ! Visual (drawing,
photography, video, etc)
The Sheep Market,
LTLYM, SwarmSketch,
Life in a Day
! Musical (music and acoustic
art)
Bicycle Built for Two
Thousand, The Virtual
Choir, In B Flat
! Literary (poetry and creative
writing)
Tim Burton’ s Cadavre
Exquis, Protagonize
By the role of
the alpha artist
! Vertical (giving specific
assignments)
LTLYM, The Johnny Cash
Project, The Sheep Market
! Horizontal (open-ended) This Exquisite Forest, Life
in a Day
By conceptual design ! Transparent (participants
know what the end product
will be)
SwarmSketch, Star Wars
Uncut, Life in a Day, The
Virtual Choir
! Opaque (participants
contribute without knowledge
of the final product)
Ten Thousand Cents, the
Sheep Market, Bicycle
Built for Two Thousand
!45
Crowdsourced Art and Relational Aesthetics
As critics have pointed out over the past two decades, contemporary art is
undergoing a crucial change, in that “the art object is no longer necessarily the primary
focus of the encounter with art” (Beech, 2010, p. 20). The demotion of the passive,
receptive viewer – rooted in conceptualism and a critique of opticality – has given rise to
participatory artistic practices that seek to blur the boundaries between artist and
By degree of
interrelation
! Dialogic (individual
contributions are in dialogue
with each other, and
participants get to see others’
contributions and build on
them)
SwarmSketch, This
Exquisite Forest, Tim
Burton’ s Cadavre Exquis
! Independent (individual
contributions are independent
from each other, and
participants do not get to see
others’ contributions until the
project is finalized)
Life in a Day, Ten
Thousand Cents, The
Sheep Market, Bicycle
Built for Two Thousand
By end product ! Single (a single collective
artwork made of small
contributions from numerous
users)
Ten Thousand Cents, Life
in a Day, The Johnny
Cash Project,
SwarmSketch, Star Wars
Uncut
! Multiple (individual user-
submitted artworks prompted
by a common assignment)
LTLYM, The Sheep
Market
By financial reward ! Free (unrewarded) LTLYM, Life in a Day
! Paid (participants are paid for
their contributions)
Ten Thousand Cents, The
Sheep Market
! Fee-based participation
(participants have to pay to
contribute)
Radiohead Remix Project
!46
audience, producer and consumer, actor and subject. On a socio-political level, the act of
questioning or challenging these boundaries is a fundamental prerogative of
contemporary art, and positions participatory art as part of “the current enthusiasm for
revisited spaces of conviviality and crucibles where heterogeneous modes of sociability
can be worked out.” (Bourriaud, 2006, p. 164). According to Claire Bishop (2004), the
emergence of participatory art mirrors social and cultural shifts in late capitalism, and is
thus a response to the perceived erosion of social relations, the commodification of
experience, and similar social and political factors. Artists working under this model
generally invoke three main motivations for the practice of participatory art: activation
(by stimulating empowerment, as well as individual and collective agency), authorship
(or the cessation of authorship to make the artistic process more democratic and
egalitarian), and community (via the restoration of lost or weakened social bonds)
(Bishop, 2006b). Jacques Ranciere (2006) echoes Bishop’s view. Noting the same
perceived deterioration of social relations in late capitalism, he argues that “yesterday’s
distance towards commodities is now inverted to propose a new proximity between
entities, the institution of new forms of social relations.” In this sense, “art no longer
wants to respond to the excess of commodities and signs, but to a lack of connections” (p.
90).
Finally, no discussion of participatory art is complete without considering
Bourriaud’s Relational Aesthetics (2002), a critical treatise on the spectrum of
contemporary art. In this important series of essays, Bourriaud defines relational art as
“an art that takes as its theoretical horizon the sphere of human interactions and its social
!47
context, rather than the assertion of an independent and private symbolic space” (p. 14).
While he admits that art has always been relational to a certain extent, he argues that
relational aesthetics came to fruition in the 1990s, when artists like Rirkit Tiravanija,
Angela Bulloch, and Angus Fairhurst began creating ephemeral social encounters as art.
Building on Marx’s use of the term “interstice” to denote trading communities that escape
the capitalist framework (barter, autarky, and so on), Bourriad advances the notion of the
artwork as “social interstice”: “a space in social relations which, although it fits more or
less harmoniously and openly into the overall system, suggests possibilities for exchanges
other than those that prevail within the system” (Bourriaud, 2006, p.161). For Bourriaud,
the basic hypothesis of relational art is “the sphere of human relations as site for the
artwork,” where “form takes priority over things, and flows over categories: the
production of gestures is more important than the production of material
things” (Bourriaud, 2006, pp. 165, 170).
Crowdsourced art, as a digital subspecies of participatory art, is rooted in the
latter’s artistic heritage of collaboration and participation. The “exquisite corpse” art
developed by the Surrealists in the early 1920sis an early example; however, this mode of
artistic production was not open to the public at large, but only to a few invited artists in
the Surrealist circle (Adamowicz, 1998). The participatory art of the 1960s and 70s – best
exemplified by movements such as Fluxus or Happenings – extended the invitation to
contribute to the artistic process to a wider public, but this was dependent on their
physical presence at the site of production or exhibition. Subsequent decades have seen a
proliferation of artists working in the participatory vein, and key figures like Thomas
!48
Hirschorn, Rirkit Tiravanija and Marina Abramovic continue to shape the directions of
contemporary art. Their projects, illustrating the social and authorial concerns of
relational art, invite participants to step into constructed situations that highlight the
social dimension of experience, and challenge traditional notions of authorship and
artistic distance.
With the rise of the Internet, artists interested in collaborative or participatory art
found an ideal platform to reach an infinitely wider and more diverse pool of potential
contributors. As curator Andrea Grover notes, “having the audience become co-creators is
not a new impulse”; the Internet simply offered a new platform to accomplish this goal
(Strickland, 2011, para. 5). Nevertheless, due to the particularities of the digital medium
as a facilitating platform, and given the lack of physical co-presence which lies at the
core of encounter-based participatory art like that of Tiravanija or Abramovic, there is a
need to differentiate between online and offline participatory art, and to engage in a
tailored discussion of online crowdsourced art that takes into account the fundamental
characteristics of its platform and mode of interaction.
Although several crucial paradigms of participatory art – particularly those of
Bishop, Bourriaud and Kester – are highly valuable for a deeper socio-cultural and
political understanding of this genre, they fall short of fully explicating the phenomenon
of online crowdsourced art due to the their omission of the Internet’s function and
mediating effect as a technological forum for collaborative artmaking. Significantly,
because there is no physical encounter in online crowdsourced art, the social aspect of
artistic co-production is to be found in the formal (digital) fiber of the art project; form
!49
and socialization are necessarily interlinked, jointly exposing the technological structure
that facilitates and mediates this digital encounter. Speaking of Rirkit Tiravanija’s
participatory art practices, Bishop (2004) acknowledges the conditioning factor of co-
presence, and aptly identifies cultural capital as a further source of identification and
social interrelation:
Despite Tiravanija’s rhetoric of open-endedness and viewer emancipation, the
structure of his work circumscribes the outcome in advance, and relies on its
presence within a gallery to differentiate it from entertainment. Tiravanija’s
microtopia gives up on the idea of transformation in public culture and reduces its
scope to the pleasures of a private group who identify with one another as gallery-
goers (pp. 68-9).
This sense of community-building and identification, however conditional and
however brief, is a core prerogative of participatory or relational art, and is much harder
to attain in online crowdsourced artistic projects. Whereas in relational art, according to
Bishop, the audience or “viewers are actually given the wherewithal to create a
community” (p. 54), the lack of social encounter and face-to-face communication makes
the creation of this community more problematic in online situations. In crowdsourced
art, the crowd is still a crowd, not yet a community.
Redefining the Role of the Artist: The Alpha Artist in Crowdsourced Art
Since crowdsourced art relies quintessentially on the artistic contributions of an
open pool of participants, this radical democratization of the artistic process comes into
!50
conflict with the traditional way of making art. The longstanding notion of art as the
individual expression of one person’s vision and artistic sensibility is questioned within
the inclusive, participatory modus operandi of crowdsourced art. This ideological conflict
has significant implications for the role of the artist, as well as the role of the audience. Is
the artist still the main author of the artwork if they completely crowdsource the
generation of artistic content? And are the contributors – as producers of artistic content –
participants, collaborators or artists themselves? While this section will focus on the role
of the project originator – let us call him “alpha artist” – in the process of crowdsourced
art, the next section will discuss the part played by the contributors, and the varying
degrees of participation that characterize such projects.
In his canonic sociological study, Art Worlds (1984), Howard Becker devotes
considerable attention to the cultural and ideological perception of the artist’s role in
society, noting the complex factors behind bestowing the title of “artist” on a cultural
producer. Since the aesthetic standards that people use to label something as art, or
someone as an artist, are not perceived in their minds as “arbitrary and conventional”, but
rather as “natural, proper and moral”, it follows that “an attack on a convention and its
aesthetic is also an attack on morality” (Becker, 1984, p. 305). In this vein, the manner in
which crowdsourced art challenges existing standards of artistic merit and questions the
cultural privilege of the professional artist in society can be seen as an attack on an entire
morality, and not just on a cultural tradition of artistic production.
The question then becomes, as Becker (1984) aptly puts it, “how little of the core
activity can a person do and still claim to be an artist?” (p.19). Given the unique
!51
characteristics of crowdsourced art, it is often difficult to delineate exactly what this
“core activity” consists of. If we consider, on the one hand, the alpha artist’s act of
conceiving the art project and providing a platform for user participation, and,
respectively, on the other hand, the participants’ act of producing the actual artistic
content, arguments can be made in favor of both these practices as “core activities.”
Without the work of the alpha artist, there would be no concept and no platform for
participation; without the work of the contributors, there would be no artistic content to
realize the artist’s vision. In either case, the project would not materialize –both the alpha
artist and the contributors are mutually dependent and can thus, arguably, both be
considered as agents of “core activities.”
However, the relative significance of the alpha artist’s input is also determined by
the vertical or horizontal nature of the crowdsourced art project, as per our previous
typology. Surely, there is a fundamental difference in the artistic engagement of Peter
Edmunds, the originator of the rather horizontal project SwarmSketch, and, respectively,
Miranda July, the artist behind Learning to Love You More, which is a more vertical
artistic endeavor. While Edmunds simply designed the drawing interface of
SwarmSketch.com and programmed it to select a new word by random from the top-
ranked Google searches of each week, July conceived of new creative assignments each
week, communicated them to the public, reviewed all submissions received and then
posted them to the website on a weekly basis. Thus, although they are both alpha artists
in their respective projects, the levels of engagement and their artistic merits are
considerably different.
!52
In a sense, in projects like July’s Learning to Love You More or MacDonald’s
Life in a Day, the alpha artist is akin to a curator. As the manager of the project, the alpha
artist – just like a curator – judges, selects, and arranges the participant-submitted
contributions into a final product which is not unlike an online gallery or museum
exhibition. The involvement of the alpha artists thus facilitates the presentation of the
project in a feasible and appealing fashion, and most often showcases the contributions
that are deemed of a superior quality.
In terms of the recognition of artistic merit, the position of the alpha artists in
crowdsourced art brings up important questions regarding formal authorship and aesthetic
accountability. Given the way that art worlds operate – and since art is rarely anonymous
– someone inevitably gets the credit for being the “artist”; moreover, as Foucault (1979)
argued in relation to literature, the author also fulfills a functional need. In music, for
instance, it is the conductor or the composer who gets the credit, not the orchestra
members; in cinema, it is the director, not the writer, the producer or the actors; in
fashion, it is the designer, not the manufacturer, the stylist or the model. However, just as
the conductor is responsible for ensuring the musical cohesion of his orchestra, so too
does the alpha artist play a fundamental role in coordinating the art project, and
establishing certain conventions that need to be in place for a meaningful participation.
These conventions, which are necessary for the functioning of any art world (Becker,
1974), must be rigid enough to guarantee a sense of cohesion in the work of art, but
flexible enough to allow for the expression of individual creativity. The establishment of
these conventions, which gives the artwork its needed structure, is a key activity that the
!53
alpha user performs, further enhancing the significance of their role and underscoring the
need for their involvement.
Redefining the Role of the Public: Qualifying Participation
While crowdsourced art challenges the traditional role of the artist, as discussed
above, it simultaneously redefines the conventional function of the public, turning them
from passive receivers into engaged producers. In spite of the value of this inclusive
trend, however, there is a critical set of questions concerning the creative agency of the
crowd in such artistic projects, and the overarching moral economy of crowdsourced art.
Specifically, in terms of crowd contributions, what constitutes meaningful artistic
participation? And is crowdsourced art guilty of the same ethical critiques attributed to
commercial crowdsourcing?
As Dave Beech (2008) aptly notes,
there is a temptation, within this earnest tradition of participation, to treat it as a
solution to the problems endemic to the whole range of established forms of
cultural engagement, from the elitism of the aesthete to the passivity of the
spectator, and from the compliance of the observer to the distance of the onlooker
(p. 315).
Artists and curators working in the tradition of relational art tend to see
participation and sociability through an optimistic, micro-utopian lens, and their
discourse most often omits conflict or hierarchization (Bishop, 2004; Foster, 2004). The
trend towards inclusion and participation, and the call to re-conceptualize art as an
!54
egalitarian cultural practice is indeed reminiscent of the Marxist impulse to revolutionize
society by abolishing class distinctions: “participation is thought of as a form of cultural
engagement that does away with all previous problematic forms of cultural engagement
by eradicating the distinction between all of the previous cultural types and all cultural
relations between them” (Beech, 2008, p. 315). From a formal or aesthetic perspective,
Bourriaud (2006) even calls for equating aesthetic judgment with an ethical or political
judgment of the relationships produced by the artwork; however, as Foster (2004)
observes, the quality of these relationships is never questioned or investigated.
In reality, there are subtle degrees of participation in art. Participation cannot be
used as a blanket term or as a panacea, since it does not account for the complexities of
creative agency, artistic hierarchies, access and capital. A close analysis of online
crowdsourced art reveals a complex ladder of participation, and I therefore suggest a
more nuanced model of understanding the various levels of engagement, highlighting the
different affordances of receptive, executory, and structural participation.
!55
Table 2: The Levels of Creative Participation in Crowdsourced Art
Type Defining Features Examples
Receptive
Consumption-based:
audience receives a
finished artistic product
Traditional artistic
modes: painting,
opera, etc
Executory
Tokenistic
Opaque and non-dialogic
micro-participation in a
pre-designed project
Bicycle Built for
Two Thousand,
Sheep Market
Engaged
Transparent but highly
structured participation in
a pre-designed project
Johnny Cash
Project,
SwarmSketch
Creative
Transparent, reflective and
expressive participation in
a pre-designed project
Life in a Day,
LTLYM, This
Exquisite Forest
Structural
Co-design
Participants are invited to
weigh in on the design or
structure of the project
Wikipainting
Co-authorship
Participants’ structural
contributions are formally
recognized and rewarded
Only offline
examples
!56
! Figure 1. The pyramid of participation: a schematic depiction of the levels of creative
participation
As illustrated by the diagram above, as the intensity of participation increases, so
does the creative agency of the contributors. In the same time, artistic projects that allow
for a greater level of participation also evidence a higher degree of conceptual
transparency, as well as an increased interdependence between the individual
contributions.
Receptive participation requires the lowest degree of creative engagement under
this model, and refers to the process whereby the viewer or audience receives a finished
artistic product. Many crucial voices in cultural studies and media scholarship have
pointed to the active elements of consumption and spectatorship (most prominently,
Barthes (1979), DeCerteau (1984), Eco (1989), Hall (2006), and Ranciere (2006)), noting
!57
that the spectator is more than a passive consumer, and takes an active role in interpreting
or decoding the artwork. The notion of receptive participation, as employed here,
accommodates this view, acknowledging the space for active interpretation and creative
consumption. Therefore, the receptive nature of this type of engagement refers to the
situation where the viewer may interpret and reflect upon the artwork, but,
fundamentally, they not invited to take an active part in the generative process.
Executory participation is the task-based, generative participation in a pre-
designed artistic project. It is analogous to entering a contract by agreeing to its specific
terms and conditions: when deciding to participate in the artistic process, the contributor
accepts the parameters of participation, but has no structural agency or, in other words,
has no power to challenge or modify these parameters. Oftentimes these types of projects
are experiments in conflating art with labor, research, or sociological experiments, as
exemplified by Aaron Koblin’s Mechanical Turk art. According to their level of
transparency and the space allowed for personal expression, executory crowdsourced art
projects can be tokenistic, engaged, or creative.
Tokenistic projects are characterized by micro-level participation in projects that
are usually opaque and non-dialogic, meaning, as per our previous typology, that the
contributor has no knowledge of the final product, nor can they see other’s contribution
and build on them. These tasks, such as recording a sound bite in Bicycle Built for Two
Thousand, or reproducing a minuscule part of a pre-drawn dollar bill in Ten Thousand
Cents, are generally uncreative and leave little space for meaningful creative expression.
Such tokenistic projects are in the most acute danger of being criticized for objectifying
!58
the crowd for the benefit of the artist and of the artwork. Grant Kester (2004) argues that
all participatory art is a way for the artist to profit from their social privilege, but
tokenistic art, in view of its lack of transparency (since contributors are kept in the dark
as to the purpose of their contribution and the scope of the project), is particularly
concerning in this respect. As Koblin (2010) anecdotally mentions in his talk at the TED
conference, out of the thousands of participants that blindly followed his instructions to
draw a sheep on Mechanical Turk, only one of them used the drawing interface to
scribble down the words “Why? Why are you doing this???” Arguably, the analogy
between the participants and sheep – as unquestioning animals drawn by herd mentality –
is present both in the concept and the execution of Koblin’s project.
The second variety of executory participation – which I have called engaged – is
transparent but highly structured participation in a pre-designed project. Engaged
participation is different from tokenistic participation in that the former refers to
transparent, and not opaque, projects. For instance, Koblin’s Johnny Cash Project is an
illustration of engaged participation due to the fact that contributors know the purpose of
their involvement, and the scope of the project; SwarmSketch similarly lets users know
the precise function of the project. Therefore, while the space for personal expression and
creative agency is still limited (for instance, a reproduction of a frame in The Johnny
Cash Project, or the generation of a single line in SwarmSketch), these two projects are
examples of engaged, and not tokenistic, participation due to their quintessential
transparency.
!59
Creative participation, while still an executory type of engagement in a pre-
designed project, refers to transparent and dialogic projects allowing for reflective,
creative personal expression. While the contributors still do not have structural agency,
the parameters of artistic participation are less strict, and the emphasis is on individual
creativity and the diversity of artistic expression. Projects where the alpha artist acts as a
curator, like Learning to Love You More or Life in a Day, are the most representative
illustrations of creative executory participation. A potential pitfall to be avoided in this
type of creative projects relates to the artistic quality of the received contributions.
Specifically, how can the alpha artist ensure that the contributions received from an
unknown and non-professional pool of participants will be of a satisfactory aesthetic or
artistic quality? Therefore, in cases like these, the role of the alpha artist in the process of
conceiving and curating the project is all the more critical, as the parameters of
engagement have to allow for enhanced creativity while simultaneously keeping the
barriers of entry – both aesthetical and technical – sufficiently low. Miranda July and
Harrell Fletcher, with Learning to Love You More, provide an admirable example of
striking a right balance in this respect, as their weekly assignment-based project allowed
for multiple points of entry, and the use of a wide variety of artistic and multimedia skills.
Finally, with structural participation, participants have a say in the conceptual and
artistic design of the project; in other words, they are allowed to demonstrate structural
agency. In co-design, participants are invited to weigh in on the structure of the project,
while in the case of co-authorship, in addition to their conceptual input, their
contributions are formally recognized and rewarded. Both co-design and co-authorship
!60
reflect a more genuine desire for openness, and they tend to function as statements on
alternative modes of artmaking.
Wikipainting is the paradigmatic example of co-design in online crowdsourced
art. This initiative provides a space for collaborative painting (with users being unlimited
in the number and type of additions they can make to any new or extant drawing), but the
feature that allows for structural agency and really qualifies this project as an exemplar of
co-design is the ability to modify the freely-available source code, as well as to suggest
improvements to the structure of the applet. The Wikipainting homepage invites
participants with coding expertise to submit upgrades or help fix lingering bugs, and
everyone else is asked to add suggestions to a “to-do list,” which the coders use to
improve the painting applet. The to-do list includes suggestions like the ability to undo
more than one brush stroke, to add text boxes, or to paint in full-screen mode.
The case for co-authorship is more complicated, and so far there have been no
cases of formally sharing authorial rights in online crowdsourced art. Indeed, participants
cannot even claim legal authorship of their individual contribution, given the copyright
and intellectual property policies of such online crowdsourced projects. Further
complications arise when the alpha artist accrues revenue from the exhibition and sale of
crowdsourced artworks, but profits are not shared with the original contributors – a
criticism similarly attributed to commercial crowdsourcing, which is often condemned as
a form of labor exploitation (Brabham, 2008; Chandler & Kapelner, 2010; Schenk &
Guittard, 2011). In the case of The Sheep Market, the participants were paid 2 cents for
their contributions, and Koblin sold each of the 10,000 generated sheep for $20 a piece.
!61
Upon hearing this, the original contributors that had drawn the sheep were outraged, and
passionately discussed this topic in a series of message-board conversations on the
Mechanical Turk forum. They created a special discussion thread titled “They’re selling
our sheep!!!”, which tackled the ethical, legal, and aesthetic aspects of the issue and
included posts like ““Does anyone remember signing over the rights to the drawings?”
and “Someone should contact them and see how much they'd charge you to buy back the
rights to one of your own sheep” (Koblin, 2006, p. 29-30).
A comparison between the executory and structural modes of participation in
crowdsourced art highlights a consequential distinction between participation and
collaboration. This fundamental contrast is the lack of influence that participants – unlike
collaborators – have over the structural design of the artwork, and the claim of authorial
rights (Browne, 2008). Thus, participants, as opposed to collaborators, do not have the
agency to affect conceptual or structural changes in the artistic project; when deciding to
contribute to the artwork, they automatically accept the parameters of the project and
agree to “enter a pre-established social environment that casts [them] in a very specific
role” (Beech, 2008, p. 315).
Another way to look at the contrast between executory and structural participation
is in terms of the dichotomy between agency and choice. The existence of a pre-
established structure or design for the artistic project offers choices of action, but falls
short of allowing for the ability to affect these choices through creative agency. In fact,
one critic goes so far as comparing, quite convincingly, the tokenistic type of
contributions in participatory art to “the participation offered in an elected democracy, or
!62
in public consultation methods where residents are given the opportunity to choose from
a fixed number of designs” (Abbott, 2008, p 24). Although any type of structure – artistic
of otherwise – rests on a certain combination of available choices, it is the ability to affect
or provide these choices that constitutes the mark of structural agency.
Conclusion: Agency and Structure in Crowdsourced Art
As evidenced by the complexities of participation outlined so far, a critical
analysis of crowdsourced art points to a larger discussion of the tension between structure
and creativity – one that has been heavily debated in both sociology and art criticism.
Janet Wolff suggests that this tension is unnatural, arguing that structure and agency are
in fact mutually interdependent, and should not be conceived of as one taking primacy
over the other (Wolff, 1981). Anthony Giddens (1976) advanced the equally
reconciliatory notion of the “duality of structure.” He suggests that structures are both the
product and the conditions of human agency, and “must not be conceptualized as simply
placing constraints upon human agency, but as enabling” as well (p. 161).
Online crowdsourced art can be seen as the product of two specific structures: the
conceptual or aesthetic structure of the project itself, and the technological structure of
the Internet as a facilitating platform. Fundamentally, the two are highly interrelated,
even though they each present their own particular affordances and pitfalls. While our
discussion so far has focused on the aesthetic structure of the project – and the type of
relationships this structure produces – it is also necessary to briefly examine the nature of
the Internet as a conditioning structure in online crowdsourced art.
!63
As Giddens suggests in regards to the duality of structure, the web, in acting like a
structure for these artistic projects, is both enabling and constraining. On the one hand,
the radical democratization of artistic involvement, by means of the web-enabled open
call, extends the opportunities of participation to a potentially global community, and thus
removes significant economic, cultural and social barriers that might have otherwise
prevented people from participating in a conventional, encounter-based artistic process.
As illustrated by our survey of the roots of crowdsourced art, while participatory art
projects existed prior to the rise of the Internet, it is the unique affordances of the web as
a platform for diversity and interconnection that enabled such initiatives to come into full
fruition. In this sense, online crowdsourced art is a timely practice that capitalizes on the
rise of participatory cultures in our contemporary society, harnessing the diversity and
creative potential of global networked communities, and providing a channel for the
collective expression of human experience. Moreover, the interactive nature of Web 2.0
and its emphasis on user-generated content (Brabham, 2008) makes it a particularly
suitable channel for crowdsourced creative contributions, turning self-expression into
digital play.
However, it is important to acknowledge that, although the invitation to
participate in crowdsourced art projects is, in theory, open to anyone, there are certain
barriers that make universal participation implausible and, in this sense, the web acts as a
constraining structure. The participation gap (Jenkins et al., 2006) – a revision of the
rather problematic term “digital divide” – is one of the principal obstacles: in spite of the
rising worldwide popularity of the web, a vast portion of the global population does not
!64
benefit from Internet access. What is more, even those who do have an Internet
connection might not be able to contribute to these crowdsourced art projects, due to low
bandwidths and the lack of sophisticated Flash and Java plugins required by the complex
interfaces of such websites. Furthermore, just like the invention of printing created a
different audience for literature (Wolff, 1981), the web as a technology produces both a
specific kind of contributor pool and a specific kind of audience. The creation – or
definition – of this audience is significant, because it affects the nature and message of
the artwork. I would argue that the desire to participate in crowdsourced art projects is
strongly related to a certain type of cultural capital, which makes one recognize the
artistic innovation and aesthetic significance of such initiatives within a larger tradition of
art and culture. This is less applicable in the case of opaque projects, where contributors
do not have knowledge of the final product or the purpose of their creations, but it is
certainly a major factor determining one’s desire to participate in projects like Learning
to Love You More or Life in a Day. Finally, for a minority of such projects, contributors
also need economic capital in order to participate: for Radiohead’s remix project, for
instance, users who wanted to try their hand at remixing the tracks had to buy the song
“stems” – the components of the track – from iTunes, thus raising important questions
about the function of this initiative as a mechanism for authentic fan engagement or,
alternatively, a profit-generating strategy.
Therefore, while crowdsourced art represents a welcomed progressive
development in the cultural tradition of art-making and a significant step towards a more
inclusive type of creative production, it is important to acknowledge that public
!65
participation in such projects is not structural, nor truly democratic. Certainly, further
research is needed in this area, particularly in regards to the motivations for participation
in crowdsourced art. Future research should attempt to look more closely at the economic
aspects of participation, from both an aesthetic and an ethical perspective. For instance,
insightful connections could be drawn, in this respect, between online crowdsourced art
such as Koblin’s Mechanical Turk projects and paid involvement in offline participatory
art – I am thinking primarily of the controversial art of Santiago Sierra, who pays illegal
immigrants, prostitutes, and homeless people minimum wage to sit in boxes in museums,
or tattoo a line across their backs.
The practice of crowdsourced art has consequential implications for the study of
online engagement, participatory culture, collective creativity, and the future directions of
modern art. The emergence of online crowdsourced art is a significant development from
both a socio-cultural and an aesthetic perspective. “Art was intended to prepare and
announce a future world: today it is modeling possible universes” (Bourriaud, 2001, p.
13). However, until participants can truly become collaborators and claim structural
agency and authorial rights, their contribution does not fundamentally challenge art’s
social marginalization, and is in danger of merely refashioning the artist-audience
relationship in an ethnographic manner (Beech, 2008). In view of these criticisms, while
crowdsourced art certainly enhances the social and cultural significance of both the final
artwork and the artistic process, it is still far from triggering the aesthetic revolution that
Joseph Beuys envisioned in his writings more than 30 years ago. While the distinction
between artist and crowd continues to represent a conditioning factor in the structure and
!66
cultural function of art in society, perhaps the next consequential question to be explored
is not whether the crowd can become artists, but whether artists can ever truly come to be
perceived as part of the crowd.
!67
CHAPTER 4:
METHODOLOGY
This chapter provides the methodological foundation for the principal study
conducted on Amazon’s Mechanical Turk. An experimental research design was used to
test the hypotheses regarding the impact of specific incentive structures (i.e. financial
rewards, bonuses, specification of project purpose, attribution of authorship credit) on the
outcomes of creative participation online (quantity of submissions, quality of
submissions, time spent on task, and number of rebellious contributions). Please note that
this chapter refers exclusively to the principal study conducted on Mechanical Turk;
Chapter 6 provides a detailed discussion of the attempts to implement this project on two
alternative platforms (hitRECord and Peer 2 Peer University), which failed to produce the
desired effects and to sustain collaboration, but which offer important insights
nonetheless.
Participants
The study was conducted on the Mechanical Turk crowdsourcing platform.
Developed by Amazon in 2005, the Mechanical Turk website allows individual requesters
to post “human intelligence tasks” (“HITs”), which online workers (“turkers”) complete
in exchange for a fee. Common HITs require workers to tag visual or written content,
transcribe audio, or answer surveys. Generally, Mechanical Turk tasks are fast, easy and
often repetitive; the monetary rewards for workers are low, most tasks being worth only a
!68
few cents. Due to its large user base and the low costs required to elicit participation, the
platform has become a popular tool for researchers, especially in the social sciences.
For the purpose of the current research, this particular platform was chosen for
two key reasons. First, after two failed attempts on alternative platforms (see Chapter 6),
I concluded that relying on Mechanical Turk was the most feasible way to facilitate large-
scale participation and to guarantee that all my tasks would be accepted and completed.
4
Second, the structure of Mechanical Turk as a vast collection of self-contained micro-
tasks made it easy to post a variety of HITs while modifying the key factors (independent
variables) that I was interested in.
5
According to Amazon, the Mechanical Turk platform has over 500,000 registered
workers (Marvit, 2014); realistically, however, the number of active turkers regularly
signing in and working on tasks must be much lower. Given the scope of this study, no
demographic data was collected about the participants (although many chose to self-
disclose information about their gender and occupation in the two open-ended questions
in my HITs; see Chapter 5). Research on the general demographics of Mechanical Turk
An interesting observation here is that the relative fame or reputation of the alpha artist is a
4
crucial factor that determines whether a particular project is likely to attract sufficient
contributions. If I were Aaron Koblin or Tim Burton, for instance, I would already have a strong
following and plenty of publicizing options, so receiving sufficient submissions for a
participatory project would not be a concern. In such cases, crowdsourced art projects can be
implemented on a designated website (or any platform of the artists’ choosing) and would not
have to rely on the paid user base facilitated by Mechanical Turk.
In January-February 2014, a pilot study was conducted to assess the feasibility of this project on
5
the Mechanical Turk platform. In particular, my concern was twofold: that workers would be
reticent to accept such unusual tasks (especially when the reward amounts were so low) and,
respectively, that the quality of their submissions would be disappointing. After launching 6 HITs
about Hashtag the Snail (3 written and 3 visual), I realized I could not have been more wrong on
both accounts. The tasks were accepted and completed quickly, and the results were truly
impressive. In addition to confirming the feasibility of this study on Mechanical Turk, running
this pilot test with different reward amounts also gave me a better idea of the ideal range of
monetary incentives that I should be offering for this type of tasks.
!69
indicates that most workers are from the US (46.8%) and India (34%). In terms of
6
gender, the majority of American workers are female, while in India the situation is
reversed. When compared to the general population of Internet users, Mechanical Turk
workers tend to be younger and better educated, while income levels are, roughly,
similarly distributed (Ipeirotis, 2010).
Materials and Procedure
This study used an experimental design whereby 42 different tasks (“HITs”, in the
Mechanical Turk terminology) were posted on Amazon’s Mechanical Turk. Each HIT
allowed for 100 responses (“assignments”); thus, a HIT was considered completed
(“expired”) when all 100 responses were submitted.
Two basic task templates were created: one for writing tasks and one for
illustration tasks. Both tasks mentioned that the workers were participating in writing - or,
respectively, illustrating - a children’s book about a snail called Hashtag (see Appendix
A).
The writing task provided an initial setup (“On his way home, Hashtag the Snail
stumbled upon another snail’s shell. He looked around, wondering who this mysterious
shell belonged to, but there was no other snail in sight…”) and asked participants to
continue the story, exquisite-corpse-style, by providing the next sentence. It was specified
The large contingent of Indian workers can be explained in part by Amazon’s decision to allow
6
for the direct payment of Indian nationals in rupees (in the form of bank checks), while all other
international workers can only disburse their earnings to an Amazon.com gift certificate
(Amazon.com, n.d. )
!70
that, in terms of the storyline, “the only requirement is that Hashtag must somehow use
the Internet to accomplish his goal.”
Figure 2. A screenshot of a writing task (control condition)
The illustration task provided the same initial sentence and asked workers to draw
and then upload an image that accurately represents that narrative situation. The HIT
further specified all types of visual depiction were welcome: digital illustrations,
drawings made by hand and then scanned or photographed, and even collages, ASCII art
or any other way they would choose to depict Hashtag’s story. The range of accepted
media was deliberately broad, so as not to exclude workers who do not have experience
with digital illustration software.
!71
Figure 3. A screenshot of an illustration task (control condition)

In addition to the field where participants wrote or uploaded their submissions,
both writing and illustration tasks included two additional open-ended questions, which
were optional (i.e. participants got paid whether or not they answered the two optional
questions). The first question asked respondents why they chose to accept that particular
task; the purpose of this question was to facilitate a better understanding of participants’
motivations in completing creative online tasks, which represented a secondary interest of
this study. The format of this question was deliberately open-ended, rather than multiple-
choice, so as not to precondition or limit the range of responses. The second question
asked participants for any feedback or comments about the task or about Hashtag’s story,
in order to provide a space for them to voice their opinions, as well as to gather valuable
feedback for the subsequent process of crowdsourcing the entire children’s book on
Mechanical Turk.
!72
Building on these two basic task templates (i.e. writing and illustration), for each
experimental condition I modified the following factors (independent variables):
Amount of financial reward. A third of the tasks (14 HITs, 1400 assignments)
paid 5 cents, a third paid 10 cents, and a third paid 20 cents. In order to keep all my
controls identical, I used the same amounts for both writing and illustration.
Provision of supplemental bonus. 6 HITs (600 assignments) mentioned a $1
bonus for the best submission, while another 6 HITs (600 assignments) promised a larger
$5 bonus. The bonus was mentioned in the title of the HIT, as well as in the description
which users can preview before deciding whether or not to accept a HIT. In addition, the
word “bonus” was also included as a keyword for users that might search for tasks in this
way. All other HITs made no mention of a potential bonus.
Stated purpose of the project. 6 HITs (600 assignments) specified - truthfully -
that the purpose of the project was charity, with the proceeds from the final book being
donated to a nonprofit organization that teaches digital literacy to impoverished youth in
India. Conversely, another 6 HITs (600 assignments) claimed - untruthfully - that the
book will be sold commercially, online and in stores, for the profit of the author. The
purpose (charity or commercial) was mentioned in the title of the HIT, in the description
and as a keyword. All other HITs made no mention of the purpose of this project.
Attribution of authorship. 6 HITs (600 assignments) mentioned that participants
will be credited as co-authors in the finalized book, while another 6 HITs (600
assignments) specified that participants would not be getting authorship credit when the
book is published. Just like in the previous cases, the attribution of authorship (or lack
!73
thereof) was mentioned in the title of the HIT, in the description and as a keyword. All
other HITs made no mention of whether participants would receive authorship credit or
not.
In order to understand the interaction between factors, a total of 42 tasks were
created, accounting for all possible combinations between type (written or visual),
amount of reward (5, 10 or 20 cents) and the specific incentives being investigated in this
study (bonus, purpose, attribution of credit). Table 3 conveys the full list of tasks that
were created and posted on Mechanical Turk.
In total, 4200 submissions were collected (2100 written and 2100 visual),
corresponding to 42 tasks with 100 responses each.
It is also important to note that I used the same initial sentence for all writing and
illustration tasks. This strategy allowed for a more accurate comparison between the
conditions and facilitated the development of a universal coding scheme to assess the
quality of all submissions. Importantly, it also helped avoid any differences caused by the
inherent characteristics of the given narrative situation; for example, one sentence might
be more narratively closed than another and therefore harder to build off of, or one
narrative situation might be more difficult to draw than another.
!74
Table 3. List of Mechanical Turk Tasks (Experimental Conditions)
Main Category Experimental Conditions
CONTROL 1. writing – 5c – no mention of other variables
2. writing – 10c – no mention of other variables
3. writing – 20c – no mention of other variables
4. illustration – 5c – no mention of other variables
5. illustration – 10c – no mention of other variables
6. illustration – 20c – no mention of other variables
BONUS 7. writing – 5c + $1 bonus
8. writing – 10c + $1 bonus
9. writing – 20c + $1 bonus
13. illustration – 5c + $1 bonus
14. illustration – 10c + $1 bonus
15. illustration – 20c + $1 bonus
16. writing – 5c + $5 bonus
17. writing – 10c + $5 bonus
18. writing – 20c + $5 bonus
16. illustration – 5c + $5 bonus
17. illustration – 10c + $5 bonus
18. illustration – 20c + $5 bonus
PURPOSE 19. writing – 5c – charity
20. writing – 10c – charity
21. writing – 20c – charity
22. illustration – 5c – charity
23. illustration – 10c – charity
24. illustration – 20c – charity
25. writing – 5c – commercial
26. writing – 10c – commercial
27. writing – 20c – commercial
28. illustration – 5c – commercial
29. illustration – 10c – commercial
30. illustration – 20c – commercial
CREDIT 31. writing – 5c – credit
32. writing – 10c – credit
33. writing – 20c – credit
34. illustration – 5c – credit
35. illustration – 10c – credit
36. illustration – 20c – credit
37. writing – 5c – no credit
38. writing – 10c – no credit
!75
The tasks were posted over a period of 4 months, from July to October 2014. To
minimize the chance that participants could come across different tasks from this same
study, the tasks were spaced out in terms of their time of posting. When scrolling through
the list of available HITs on Mechanical Turk, participants could see the title and brief
description of each task (see Table A1 in Appendix 1 for a detailed list of the titles and
descriptions that participants saw). If they clicked on a HIT of interest, they could request
to see a preview which included the text containing the task instructions. The standard
instructions for both writing and illustration tasks (control condition) are included in
Appendix 1. In addition to these instructions, an initial paragraph was added, where
applicable, to expand on the context of the children’s book project. For instance, the task
instructions for the charity condition provided further details about the educational
project that the money would be donated to, while the instructions for the commercial
condition mentioned that the finalized book would be sold commercially in stores and
online.
Although this mode of implementation implied a significant challenge - i.e. the
fact that participants could see and participate in related tasks - this strategy was chosen
because it allowed for the investigation of all research questions in a feasible manner.
Specifically, for my first research question, the goal was to assess which tasks
participants would be most likely to accept, depending to the different incentive
CREDIT (cont’d) 39. writing – 20c – no credit
40. illustration – 5c – no credit
41. illustration – 10c – no credit
42. illustration – 20c – no credit
!76
conditions. In this case, obscuring the context of the task and placing participants in a
random experimental condition would not have been a good solution, because it would
not have enabled me to observe their selection process. Furthermore, because of the sheer
number of HITs posted every hour, and the large number of potential participants on
Mechanical Turk, the possibility that people would participate in more than one HIT was
not deemed to be an insurmountable challenge in the context of this study. Nonetheless,
this is an important methodological challenge, and one that is common to studies
undertaken on Mechanical turk and similar task-based online platforms. In Chapter 5, I
address this limitation in more detail and suggest alternative strategies for the setup of
similar experimental studies on crowdsourcing platforms.

Measures
Based on comparative analyses between these experimental conditions, the
principal goal of the study was to assess the impact of the above-mentioned independent
variables on the following measures (dependent variables):
Quantity of submissions. The quantity of submissions was operationalized as the
amount of time (measured in hours) that it took for all 100 assignments to be completed
within each task. In other words, how easy was it to reach the desired number of
submissions in each experimental condition? This measure was computed by calculating
the difference, in hours, between the time a task was posted and, respectively, the time
that the last assignment pertaining to that task was submitted.
!77
Quality of submissions. For the purpose of this study, the highest quality
submissions were those that were most creative. The most widely accepted
conceptualization of creativity has been as a combination of usefulness and novelty
(Amabile, 1983; Amabile & Pillemer, 2012). Therefore, in assessing the quality of the
submissions in this study, a coding rubric was developed to measure those two
dimensions of creativity: usefulness and novelty. Both written and visual submissions
were assessed for usefulness (ranging from 0 to 3 points) and novelty (also 0-3 points), as
per the rubric in Appendix B.
To ensure maximum applicability and relevance, the coding rubric for measuring
the quality of submissions was developed in an emerging fashion, after a preliminary
examination of the data. Unitizing was intuitive in this case, as each submission
represented one coding unit. In developing the scheme, a key goal was to create
consistency - as much as possible - across the two types of tasks: writing and illustration.
Thus, for example, looking at the criteria for novelty, both writing and illustration tasks
are assessed according to the same essential features: exceeding the requirements of a
basic sentence/illustration, using unusual (narrative or visual) elements, and employing
humor or whimsy. For usefulness, maintaining consistency across task type - writing and
illustration - was comparatively more difficult, given that the elements that make a
written sentence useful are inherently different than those that qualify an illustration as
useful. I did, nonetheless, manage to establish one common criteria - that of fit (narrative
or visual) - while the remaining two criteria were based on the submission’s adherence to
the specific task instructions.
!78
After developing an initial draft of the coding scheme and training another coder,
we both coded independently a small sample of submissions (200 written submissions
and 200 illustrations) in order to determine the accuracy of the coding scheme. When we
compared results, the percent agreement across all categories (writing-usefulness,
writing-novelty, illustration-usefulness, illustration-novelty) was higher than 85%, but we
noticed that the most frequent disagreements occurred in regards to the assessment of
novelty. Specifically, we disagreed as to what counted as “exceeding the basic
requirements” of a sentence/illustration and what counted as humorous or whimsical.
Based on these initial observations, the coding scheme was refined and the criteria for
assessing those two novelty points were made more specific (see Appendix B).
With the finalized coding scheme in place, we randomly selected a subset of 10
tasks (5 writing, 5 illustration) to code, which represented 23.8% of the entire sample.
The selected tasks were representative of the full spectrum of experimental conditions
(i.e. all reward amounts and all incentive structures). Intercoder reliability was assessed
with the aid of ReCal2, an online software developed by Dr. Deen Freelon of American
University. The results were very satisfactory and are reproduced in the table below.
Table 4. Intercoder Reliability Results for Writing Tasks
Measure
Percent
Agreement
Scott's
Pi
Krippendorff's
Alpha
Usefulness 99.20 0.98 0.98
Novelty 96.40 0.93 0.93
!79
Time spent on task. This dependent variable aimed to assess the relative effort
that participants put into their work, by measuring the amount of time, in seconds, that
was spent on completing each submission. This measure was computed automatically by
Mechanical Turk and exported as metadata.
Number of rebellious contributions. Any submission that had nothing to do with
the given assignment (specifically, with the snail’s story) was considered rebellious. Such
submissions were identified during the coding process and were marked as rebellious in a
special column, in addition to receiving 0 points for usefulness and 0 points for novelty.
Most rebellious contributions consisted of random text or images copied and pasted from
the Internet, in order to receive the payment for the task without actually creating an
original submission specifically for this assignment. For an overview of the content and
scope of the rebellious contributions submitted by participants, see the discussion in
Chapter 5.
As mentioned previously, each task on Mechanical Turk also contained an
optional open-ended question about the participants’ motivations for selecting that
particular task. These responses were also coded and categorized, but this task was done
by the main researcher alone, without help from a second coder. The reasons for this
choice are twofold: one, the question regarding participant motivations was secondary in
Table 5. Intercoder Reliability Results for Illustration Tasks
Measure
Percent
Agreement
Scott's
Pi
Krippendorff's
Alpha
Usefulness 96.92 0.95 0.95
Novelty 93.28 0.90 0.90
!80
importance, and two, the categorization of answers in this case was relatively
straightforward and did not require complex decision-making and interpretation.
The development of the coding categories for participant motivations was done ex
post facto, after reading through all the responses provided. Six categories emerged as
most prominent:
• For fun
• For the money
• I enjoy being creative
• The task seemed unique / unusual
• I enjoy children’s books
• I wanted to contribute to something bigger
In addition, for the tasks that tested the impact of the incentive structures (bonus,
purpose, credit), a seventh category was added, corresponding to the relevant incentive in
each case (“for the potential bonus”, “for charity”, “for authorship credit”).
The coding scheme for participant motivations, along with representative
examples, is provided in Appendix C.
Data analysis
The data was first cleaned up by removing compromised, incomplete and
duplicate contributions. Next, the variables were coded numerically and inputted into
SPSS 22, where the data set was analyzed via means comparison (in regards to the
quantity of submissions), ANOV As (for usefulness, novelty, and time spent) and chi
!81
square tests (regarding number of rebellious contributions). When there was homogeneity
of variances, as assessed by Levene's test for equality of variances, ANOV As were
followed by post-hoc Turkey tests to determine statistically significant contrasts between
the groups. When variances were found to be not homogeneous, Welch’s ANOV As were
used, followed by post-hoc Games-Howell tests.
The results of these statistical tests are presented in the next chapter, along with a
graphical depiction of the differences between the main experimental conditions in terms
of quantity, quality, time spent and rebellious contributions.
!82
CHAPTER 5:
RESULTS
On an empirical level, the principal goal of this study was to investigate, through
an experimental design, the conditions that most effectively foster creative participation
in online spaces. Specifically, using Amazon’s Mechanical Turk as my main experimental
platform, I wanted to assess the impact of four key factors (possibility of bonus payment;
stated purpose of the project; attribution of authorship credit; amount of reward) on the
quantity and quality of the resulting submissions, the time spent by participants on task,
and the number of rebellious (i.e. off-task) contributions. This chapter outlines my
findings in relation to my hypotheses, and then provides a discussion of the most
insightful results, as well as a consideration of the limitations of the current study.
Quantity of Submissions
My first hypothesis (H1) stated that the quantity of submissions will increase
when:
a. there is a potential bonus payment.
b. the stated purpose of the task is charity, and will decrease when the
stated purpose is commercial.
c. there is the possibility of gaining credit for one’s work, and will
decrease when there is no attribution of authorship credit.
d. the amount of the financial reward is greater, and will decrease when
the reward is lower.
!83
The quantity of submissions was operationalized as speed of completion, meaning
the time (in hours) it took for all 100 assignments to be completed within each condition.
Looking at the results for each experimental condition (Table 6), the hypotheses
regarding bonus payments and reward amount were supported. The hypotheses
7
regarding purpose and authorship credit were contradicted, as both of these variables had
an opposite effect than had been expected. Rather than encouraging potential workers to
accept the task, the mention of a charitable purpose and, respectively, the possibility of
gaining credit dissuaded them from participating.
Table 6, below, represents the average number of hours to completion for writing
tasks, in each of the 7 incentive conditions (aggregated across all reward amounts) and
for each of the 3 reward amounts (aggregated across all incentive conditions).
Table 6. Quantity of Submissions: Summary of Results
Hypothesis Group
# of hours till 100
submissions
Hypothesis
supported?
H1: The quantity of submissions will
increase when:
a. there is a potential bonus payment
control 358
YES $1 bonus 322
$5 bonus 294
b. the purpose of the task is charity, and
will decrease when the purpose is
commercial
control 358
NO charity 380
comm. 328
c. authorship credit is offered, and will
decrease when there is no credit offered
control 358
NO credit 553
no credit 356
Further analyses of statistical significance could not be performed in this case, since the value of
7
the dependent variable (i.e. number of hours) refers to the entire group, and not to individual
cases within the groups.
!84
For a clearer picture of the impact of reward amounts on the speed of task
completion, please see Appendix D, which depicts the distribution of values according to
reward amounts for writing tasks (Figure A1) and for illustration tasks (Figure A2).
Quality of Submissions
According to the rationale outlined in Chapter 4, quality was conceptualized along
two dimensions: usefulness and novelty. For the purpose of clarity, I will treat the two
concepts - usefulness and, respectively, novelty - separately in the following analysis.
Usefulness
Hypothesis H2 predicted that the usefulness of the submissions will increase
when:
a. there is a potential bonus payment.
b. the stated purpose of the task is charity, and will decrease when the
stated purpose is commercial.
c. there is the possibility of gaining credit for one’s work, and will
decrease when there is no attribution of authorship credit.
d. the amount of the financial reward is greater, and will decrease when
the reward is lower.
d. the amount of the reward is greater,
and will decrease when the reward is lower
5 cents 491
YES 10 cents 378
20 cents 241
Hypothesis Group
# of hours till 100
submissions
Hypothesis
supported?
!85
First, a series of one-way ANOV As were conducted to assess - in accordance with
these hypotheses - whether the difference in usefulness scores between the experimental
conditions was statistically significant.
In terms of the impact of bonuses, although usefulness scores increased from no
bonus to $1 bonus to $5 bonus, in that order, the difference between groups was not
statistically significant.
A similar outcome occurred in regards to purpose as well. Although, as predicted,
the average usefulness score for charity was higher than in the control condition, while
the score for commercial was lower than in the control condition, the difference was not
statistically significant. However, if we analyze the illustration submissions separately,
that ANOV A is indeed significant, F(2, 674) = 4.828, p = .008. The post-hoc Tukey test
identified a statistically significant contrast between the charity and commercial groups:
for illustration tasks, usefulness scores for the charity condition (M = 1.40, SD = 1.24)
were much higher than for the commercial condition (M = 1.04, SD = 1.27), resulting in a
mean increase of .36, 95% CI [0.086, 0.637], p = .006.
In regards to authorship credit, just like in the cases above, the difference in
scores fits the hypothesis - compared to the control condition, usefulness scores are
higher when credit is offered and lower when there is no attribution of credit - but once
again, the difference is not large enough to be statistically significant.
Finally, looking at the impact of the reward amounts, there was a statistically
significant difference in usefulness scores between the three conditions (5 cents, 10 cents,
and 20 cents): Welch’s F(2, 2435.396) = 8.937, p < .001. According to the Games-Howell
!86
post-hoc test, the contrast between the lowest and the highest paid groups was statistically
significant at p < .001, with a mean increase of .20 (95% CI [.0891, .3109]) in usefulness
scores between the 5-cent condition and the 20-cent condition.
Therefore, we can conclude that only Hypothesis 2d (regarding the amount of the
financial rewards) was supported to a statistically significant degree, while Hypothesis 2b
(regarding the stated purpose of the project) was true for illustration tasks, but not for
writing tasks. Hypotheses 2a and 2c were not supported: offering a bonus payment or
providing authorship credit made little difference in terms of the usefulness of the
submissions across both types of tasks.
A summary of the results is presented in the Table 7, below.
Table 7. Usefulness of Submissions: Summary of Results
Hypothesis
Group means and
standard deviations
ANOV A
F value
Hypothesis
supported?
Group M SD
H2: Usefulness will increase when:
a. there is a potential bonus
payment
control 1.97 1.15
0.378 NO $1 bonus 2.02 1.14
$5 bonus 2.03 1.18
b. the purpose of the task is charity,
and will decrease when the purpose
is commercial
control 1.97 1.15
2.020
PARTIALLY:
Only for
illustration
tasks
charity 2.07 1.09
comm. 1.93 1.22
c. authorship credit is offered, and
will decrease when there is no credit
offered
control 1.97 1.15
1.221 NO credit 1.99 1.16
no credit 1.89 1.19
d. the amount of the reward is
greater, and will decrease when the
reward is lower
5 cents 1.89 1.20
8.937** YES 10 cents 1.99 1.15
20 cents 2.08 1.13
Note: ** p < .001
!87
In addition to these one-way tests, a series of factorial ANOV As were conducted
in order to assess the potential three-way interaction between incentive condition, reward
amount and type of task. Specifically, each of the ANOV As looked at a different incentive
structure: bonus (control vs $1 bonus vs $5 bonus), purpose (control vs charity vs
commercial) and attribution of credit (control vs credit vs no credit). There were no
significant three-way interactions between incentive, reward amount and type of task for
any of the incentive conditions: bonus, purpose, or credit. However, the two-way
interaction between type of task and reward amount was significant in all cases: for bonus
(F (2, 1567) = 16.47, p < .001), purpose (F (2, 1559) = 7.51, p < .001) and credit (F (2,
1562) = 11.79, p < .001). In addition, there was a significant two-way interaction
between purpose and type of task (F (2, 1559) = 3.74, p < .024).
For a more detailed view of the distribution of values according to reward
amounts, please see Figures A3 and A4 in Appendix D.
Novelty
In regards to the other dimension of quality, hypothesis H3 predicted that the
novelty of the submissions will increase when:
a. there is a potential bonus payment.
b. the stated purpose of the task is charity, and will decrease when the
stated purpose is commercial.
c. there is the possibility of gaining credit for one’s work, and will
decrease when there is no attribution of authorship credit.
!88
d. the amount of the financial reward is greater, and will decrease when
the reward is lower.
The impact of bonus was statistically significant, F(2, 1582) = 6.484, p = .002.
The Tukey test further identified a statistically significant difference between the control
condition and the $5 bonus condition, which amounted to a mean increase of .196, 95%
CI [.0680, .3240], p = .001.
In terms of the stated purpose of the project, the difference between the three
conditions (control, charity and commercial) was also statistically significant, Welch’s
F(2, 1047.543) = 21.159, p < .001. The Games-Howell test revealed two significant
contrasts (both at p < .001): there was a mean increase of .22 (95% CI [.1043, .3419])
between the control group and the charity group, and an even bigger increase of .33 (95%
CI [.2081, .4479]) between the commercial and the charity conditions.
The difference between credit conditions (control, credit, no credit) was also
statistically significant, Welch’s F(2, 1047.077) = 20.497, p < .001. Novelty scores
increased from the no credit group to the control group to the credit group, in that order.
A Games-Howell post-hoc analysis revealed that all these increases were statistically
significant: the mean increase from no credit to control (.16, 95% CI [.0459, .2686], p = .
003) from control to credit (.15, 95% CI [.0276, .2678], p = .011) and from no credit to
credit (.30, 95% CI [.1926, .4174], p < .001).
Looking at the impact of the reward amounts, the difference between the three
conditions was significant, F(2, 3671) = 8.200, p < .001. According to the post-hoc Tukey
test, there was a statistically significant contrast between the 5-cent and the 20-cent
!89
groups (a mean increase of .13, 95% CI [.0492, .2097], p < .001), and between the 10-
cent and 20-cent groups (a mean increase of .11, 95% CI [.0258, .1846], p=.005).
In conclusion, all hypotheses about novelty were supported (and with very high
statistical significance levels), suggesting that the novelty of creative contributions can
indeed be boosted by a bonus payment (Hypothesis 2a), a charitable purpose (Hypothesis
2b), the attribution of authorship credit (Hypothesis 2c) or a higher financial rewards
(Hypothesis 2d).
A summary of the results from the statistical tests is presented below, in Table 8.
Finally, a three-way ANOV A indicated that there were no statistically significant
three-way interactions between incentive, type of task and reward amount for any of the
Table 8. Novelty of Submissions: Summary of Results
Hypothesis
Group means and
standard deviations
ANOV A
F value
Hypothesis
supported?
Group M SD
H3: Novelty will increase when:
a. there is a potential bonus
payment
control 0.63 0.83
6.484* YES $1 bonus 0.74 0.89
$5 bonus 0.83 0.93
b. the purpose of the task is charity,
and will decrease when the purpose
is commercial
control 0.63 0.83
21.159** YES charity 0.85 0.88
comm. 0.52 0.76
c. authorship credit is offered, and
will decrease when there is no credit
offered
control 0.63 0.83
20.497** YES credit 0.77 0.84
no credit 0.47 0.71
d. the amount of the reward is
greater, and will decrease when the
reward is lower
5 cents 0.64 0.83
8.200** YES 10 cents 0.66 0.81
20 cents 0.77 0.88
Note: *p < .05; ** p < .001
!90
incentive conditions. There was, however, a significant two-way interaction between
bonus and task type (F (2, 1567) = 4.58, p = .01) and between purpose and task type (F
(2, 1559) = 3.03, p = .49). Additionally the interaction between task type and reward
amount was significant in regards to the impact of bonuses (F (2, 1567) = 7.03, p = .001).
For the distribution of values according to reward amounts, please see Figures A5
and A6 in Appendix D.

Time Spent on Task
According to H4, I predicted that the time spent by participants on completing the
task will increase when:
a. there is a potential bonus payment.
b. the stated purpose of the task is charity, and will decrease when the
stated purpose is commercial.
c. there is the possibility of gaining credit for one’s work, and will
decrease when there is no attribution of authorship credit.
d. the amount of the financial reward is greater, and will decrease when
the reward is lower.
In regards to bonus payments, the difference between groups was statistically
significant, Welch’s F (2, 1013.039) = 7.779, p < .001. Games-Howell tests found that
both increases in time spent were statistically significant: between no bonus and $1 bonus
(a mean increase of 125.25, 95% CI [44.8483, 205.6539], p = .001) and between no
bonus and $5 bonus (a mean increase of 85.50, 95% CI [7.2861, 163.7111], p = .028).
!91
The differences in terms of project purpose (control vs charity vs commercial)
were also significant, Welch’s F (2, 972.770) = 3.484, p = .031. There was a statistically
significant contrast between the commercial and charity conditions: a mean increase of
131.04, 95% CI [13.8820, 248.2046], p = .024.
In terms of credit, although the time spent by participants on the task increased
from the no credit condition to the control condition to the credit condition, in that order,
the difference was not statistically significant.
Finally, looking at the amount of the financial rewards, the difference between the
three groups (5 cents, 10 cents and 20 cents) was significant, Welch’s F (2, 2301.900) =
22.489, p < .001. The post-hoc tests revealed that all the contrasts between the conditions
were statistically significant: between 5 and 10 cents (a mean increase of 54.13, 95% CI
[11.9854, 96.2838], p = .007), between 10 and 20 cents (a mean increase of 129.97, 95%
CI [60.7303, 199.2020], p < .001) and between 5 and 20 cents (a mean increase of
184.10, 95% CI [118.0893, 250.1121], p < .001).
Table 9 presents a summary of the findings in relation to the hypotheses. As a
reminder, the time spent by participants on completing the task is measured in seconds.
!92
A factorial ANOV A indicated that there was a statistically significant three-way
interaction between bonus, reward amount and task type, F (4, 1567) = 5.929, p < .001.
and between credit, reward amount and task type, F (4, 1562) = 3.238, p = .012. No
significant interaction was found between purpose, reward amount and task type. Each of
the significant three-way interactions were followed up with simple two-way interactions,
then simple simple main effects, and finally simple simple pairwise comparisons.
For bonus, there was a statistically significant two-way interaction between bonus
condition and reward amount for illustration tasks, F (4, 1567) = 11.160, p < .001, but not
for writing tasks. Next, I found a statistically significant simple simple main effect of
bonus condition for illustration tasks that paid 20 cents, (F (2, 1567) = 28.664, p < .001),
Table 9. Time Spent on Task: Summary of Results
Hypothesis
Group means and
standard deviations
ANOV A
F value
Hypothesis
supported?
Group M SD
H4: Time spent on task will
increase when:
a. there is a potential bonus
payment
control 231.20 439.90
7.779** YES $1 bonus 356.45 659.21
$5 bonus 316.70 620.53
b. the purpose of the task is charity,
and will decrease when the purpose
is commercial
control 231.20 439.90
3.484* YES charity 340.05 1075.84
comm. 209.01 390.25
c. authorship credit is offered, and
will decrease when there is no credit
offered
control 231.20 439.90
1.621 NO credit 287.26 587.85
no credit 244.86 547.84
d. the amount of the reward is
greater, and will decrease when the
reward is lower
5 cents 202.40 378.91
22.489** YES 10 cents 256.54 497.10
20 cents 386.50 921.44
Note: *p < .05; ** p < .001
!93
but not for ones that paid 5 cents or 10 cents. All simple simple pairwise comparisons
were run for illustration tasks at the 20-cent level with a Bonferroni adjustment applied.
The mean time spent by participants was 362.68 seconds (SD = 59.85) when no bonus
was offered, 979.16 seconds (SD = 55.70) when a $1 bonus was offered, and 642.19
seconds (SD = 61.71) when a $5 bonus was offered. There was a statistically significant
mean difference between the no bonus and $1 bonus conditions (616.48 seconds, 95% CI
[420.539, 812.413], p < .001), between the no bonus and $5 bonus conditions (279.51
seconds, 95% CI [73.475, 485.547], p = .004), and between the $1 bonus and $5 bonus
conditions (336.97 seconds, 95% CI [137.733, 536.197], p < .001).
Similarly, looking at authorship credit, there was a statistically significant simple
two-way interaction between credit condition and reward amount for illustration tasks, F
(4, 1562) = 6.859, p < .001, but not for writing tasks. There was a statistically significant
simple simple main effect of authorship credit for illustration tasks that paid 20 cents, (F
(2, 1562) = 12.406, p < .001), but not for ones that paid 5 cents or 10 cents. All simple
simple pairwise comparisons were run for illustration tasks at the 20-cent level with a
Bonferroni adjustment applied. The mean time spent by participants was 362.68 seconds
(SD = 59.85) in the control condition, 758.33 seconds (SD = 57.44) when authorship
credit was offered, and 511.92 seconds (SD = 58.56) when credit was explicitly not
offered. There was a statistically significant mean difference between the control
condition and the credit condition (395.65 seconds, 95% CI [203.898, 587.403], p < .001)
and between the credit and no credit conditions (246.40 seconds, 95% CI [49.825,
442.991], p = .008).
!94
The above interactions - regarding bonus and credit - are confirmed by a look at
the distribution of time spent values in relation to reward amounts (see Figures A7 and
A8 in Appendix D).
Rebellious Contributions
According to hypothesis H5, I expected the number of rebellious submissions to
increase when:
a. there is a potential bonus payment.
b. the stated purpose of the task is commercial, and will decrease when the
stated purpose is charity.
c. there is no attribution of authorship credit, and will decrease when there
is the possibility of gaining credit for one’s work.
d. the amount of the financial reward is lower.
Although a comparative look at the number of rebellious contributions across
incentive conditions indicated that all these statements were true (with the exception of
the $1 bonus condition, which had fewer rebellious submissions than the control
condition), the relationships were not statistically significant, according to a series of chi-
square tests for association. For increased clarity, the chi square tests were also performed
separately for writing and illustration tasks. All expected cell frequencies were greater
than 5. None of the associations between incentive conditions and number of rebellious
contributions were found to be statistically significant, for either writing or illustration
tasks. However, for illustration tasks, there was a statistically significant association
!95
between the amount of the financial reward and the number of rebellious contributions
(χ
2
(2) = 17.120, p < 0.001), with far more rebellious submissions in the 5-cent condition
as compared to 10 cents and 20 cents, across all incentive structures. This association was
not statistically significant for writing tasks.
Therefore, we can conclude that only Hypothesis 5d (regarding the amount of the
financial reward) was partially supported for illustration tasks, but not for writing tasks.
None of the other hypotheses were supported at a statistically significant level.
For the distribution of values according to reward amounts, please see Figures A9
and A10 in Appendix D.
Participants’ Motivations
Another important issue that the current study helped illuminate concerns the self-
stated motivations that drive creative participation online. This data was based on the
responses that participants provided for the open-ended optional question (“Why did you
accept this particular HIT?”) included within every task. All answers were coded as per
the coding scheme detailed in Appendix C.
The overall response rate was 54%, which is very impressive, given that the
questions was optional and that Mechanical Turk workers are looking to finish tasks as
quickly as possible, in order to complete as many HITs as they can and thus maximize
their hourly rate. Under these conditions, the fact that more than half of the participants
took the time to answer an optional question exceeded my expectations regarding the
response rate and, I believe, is a further sign that they genuinely enjoyed the task and felt
!96
more connected to this HIT in comparison to other tasks on the Mechanical Turk
platform.
Figure 4. Participants’ motivations for accepting the HITs about Hashtag the Snail
Averaging the responses across all reward amounts and incentive conditions, the
primary motivation that participants invoked was a desire to have fun (42%), which is
closely related to the second most popular response: the enjoyment of creative tasks
(17%). For this latter category, many respondents chose to disclose their profession
(“writer”, “freelance journalist”, “illustrator”, “graphic designer”) or their education
(“English Lit major”, “art school”) if they deemed it relevant to their enjoyment of, or
experience with, creative work. The environment where they encountered the task
mattered significantly too: 9% said they accepted the HIT because it seemed unique and
unusual compared to the dull and non-creative tasks typically posted on Mechanical Turk
(most of which being lengthy surveys and requests to tag, label or organize written and
visual information).
!97
Another 9% chose to work on the task because they enjoy children’s books; these
responses were the most personal ones, where people self-disclosed information about
their childhood memories, their favorite children’s books, or about their own children,
nieces, nephews or grandchildren. Some responses were downright adorable. One
participant said that her submission was written by her young daughter because “writing
is something we enjoy doing together”, while another talked about sharing his illustration
with his daughter on social media: “once I finished, I posted the image on Facebook &
tagged my youngest kiddo. Her smile alone was worth taking this HIT.”
On average, only 7% of participants invoked the financial reward as their reason
for accepting this HIT. For 5-cent tasks, this rate was lowest (only 5% invoked money as
their primary motivation), and it grew slightly for 10-cent tasks (9%) and 20-cent tasks
(8%).
Beyond such minor contrasts, there were few differences in terms of participant
motivations across reward amounts and incentive conditions. The percentages remained
strikingly similar across all tasks. A noteworthy conclusion is that fun and creativity still
outweighed in importance the experimental incentive structures (bonus, purpose and
credit), at least in terms of the participants’ explicit self-disclosure of their motivations. In
their respective experimental conditions, only 4% chose the task because of the $1 bonus,
7% for the $5 bonus, 13% because of the charity aspect, and only 4% for the authorship
credit. Appendix E includes a more detailed look at the motivations invoked across
reward amounts and incentives, with a pie chart for each experimental condition.
!98
Discussion
This study makes a significant contribution towards a better understanding of
creative crowdsourcing practices. In particular, the process of modifying these key factors
(financial reward, bonus, purpose and authorship credit) and assessing the resulting
submissions helps shape a more nuanced view of the strategies that work best when
soliciting creative contributions online. These conclusions are valuable not only within
the realm of crowdsourcing and participatory culture scholarship, but also on an applied
level, in terms of informing the practice of individuals and organizations that seek to rely
on large-scale public contributions for their creative projects.
When the goal is to gather as many contributions as possible in a relatively short
time span, the results of this study indicate that financial rewards work best to achieve the
desired result. In other words, offering a higher reward and/or an additional bonus will
lead to the timely completion of tasks, but does not always ensure the best quality and
greatest effort on the part of the contributors. A surprising result that contradicted my
hypotheses was that neither a prosocial purpose nor the attribution of credit work to
precipitate the completion of tasks. In particular, the attribution of credit, which I had
envisioned as a significant incentive, proved to have the opposite effect, dissuading
participants from accepting the tasks in a timely manner. This effect was observed for
both writing and illustration tasks (although it was significant more pronounced for the
former) and is most likely explained by the fact that people are reticent to attach their
name to a project or task unless they are confident in their skills and certain that their
contribution will be well received. In the case of this children’s book, the fact that
!99
contributors did not have full knowledge of the final content of the book - nor its public
framing and publication venues - could have made some of them less eager to contribute
and be credited as co-authors on a final product that is largely outside of their control.
Therefore, if authorship credit is an essential element of a collaborative creative project,
sufficient details must be provided regarding the final outcome of the project, including -
when applicable - legal and ethical considerations surrounding collaborative authorship.
If, however, the goal is not quantity and speed of completion, but quality -
especially in terms of maximum creativity and diversity of submissions - the best strategy
is to offer a bonus, emphasize the purpose of the project (if it is a charitable/prosocial
one), and offer authorship credit. Indeed, for both written and visual contributions, a
bonus, a charitable purpose or an attribution of authorship all promise to increase the
novelty of the responses - thus increasing the overall diversity of the pool of submissions
- but not necessarily their usefulness. The time spent by participants on completing the
task also increases in these conditions.
A key conclusion that emerged from this study is that, when the task is perceived
as interesting or pleasurable, internal incentives (fun, relaxation, creative effort) trump
external rewards. This observation is supported by the participants’ self-stated
motivations in the first optional question - fun and creativity being the predominant
reasons for completing the task - as well as by the feedback that they provided on the
second optional question. Furthermore, this conclusion is in line with the literature on
intrinsic versus extrinsic motivation, which suggests that intrinsic incentives are stronger
than extrinsic ones, and that in some cases the latter can “crowd out” the former,
!100
decreasing performance and enjoyment (Deci, 1975; Deci, Koestner & Ryan, 1999; Frey,
1997). In the current study, the amount of the financial reward - whether the task paid 5,
10 or 20 cents - had little impact on the quality of the written submissions and the time
spend by the participants on the task. For illustration tasks, on the other hand, quality and
effort were indeed higher when the reward was greater, presumably because the pay for
illustration tasks was deemed to be too low, given the amount of time it took to complete
these drawings and the comparative difficulty of this task versus the written tasks.
Indeed, the fact that the reward amounts for illustration tasks were considered to
be too low - while the writing tasks were deemed appropriately priced - could have had
an impact on the accuracy of the results in regards to illustration submissions. The
decision to offer the same rewards (5, 10 and 20 cents) for both writing and illustration
tasks was in order to keep all controls identical and avoid skewing the results. However,
the feedback from participants in the post-task optional question made it became clear
that they considered the pay for illustration HITs to be insufficient for the time they
devoted to the task. Interestingly enough, the respondents did not seem upset about the
unfair pay, but rather pointed it out in a positive manner (“Took 30 minutes, but I enjoyed
it!”; “I figured that for the price on this HIT you'd have trouble finding any takers. I
figured what the heck.”). Beyond impacting the relationship between money and quality/
effort - discussed in the paragraph above - the relatively low reward for illustration tasks
also resulted in many rebellious contributions and rampant cheating, which did not occur
!101
for writing HITs. Participants submitted duplicate images from different accounts ,
8
uploaded images of snails from the Internet and - even more interestingly - submitted a
wide variety of rebellious contributions. The content of the rebellious contributions was
9
nonetheless very amusing if oftentimes perplexing; their analysis could be fodder for an
entirely new study! Most rebellious contributions were pictures of animals and babies
(including a couple of babies with snail shells on their backs, taken from the Internet).
Others were personal pictures, screenshots and even a couple of comics. In one of the
more interesting cases, someone uploaded a 3D digital model of a house that they had
created, noting in the feedback space that they could not draw a snail but they are very
talented at drawing buildings if I am ever in need of those skills. A selection of rebellious
contributions is included in Appendix F.
Although the results in regards to the study hypotheses are very insightful, a
statistical analysis cannot begin to convey the extraordinary creativity, diversity and
humor that characterized the responses for both written and visual responses.
For writing tasks, most participants continued the initial sentence by having
Hashtag post a “lost and found” ad for the missing shell on various websites, online
Although it is against the rules of the site, this practice is increasingly common on Mechanical
8
Turk, because it allows workers to multiply their income by participating in the same task and
submitting similar or identical responses.
I could have rejected these submissions on a case-by-case basis, but I chose to auto-approve all
9
submissions because:
1) the sheer volume of responses (i.e. 4200) made it very unfeasible to inspect each one in turn
and approve payment in a timely manner; if I did not approve payment in a timely manner
after task completion, my requester reputation on Turkopticon would plummet, dissuading
users from accepting my tasks.
2) I had already auto-approved writing tasks so I wanted to be consistent across task types.
3) I thought it was interesting and instructive to see what participants choose to submit and
under what circumstances they are tempted to cheat or rebel.
!102
community boards and social media (Snailbook, Snitter, Snailslist, etc), usually after
snapping a picture of the shell with his smartphone (amusingly dubbed slimePhone,
iSnail, iSlime, shellular phone, etc.). Hashtag found many friends in his adventures:
characters like Trending the Slug, Underscore the Worm, Ampersand the Snake, Emoji
the Turtle, Tweeter the Bird, Wifi the Walrus, Instagram the Bee, Grandmother Google,
Google the Frog, Google the Goat, Google the Groundhog, Escargoogle, and many other
snails that went by names like Selfie, Retweet, Backslash, Dotcom, Websurf, Asterisk,
Barcode, SlowPoke, Shell Script, and Cyber. The submissions included lots of clever
puns (Looking at the empty shell, Hashtag wondered “where did Eskar go?!?”) and even
rhyming (An empty shell? What a fright! / For surely this shell is another snails's
delight…/ Oh! I know what to do, I'll make it right. / I’ll make it right with this tweet I
write!).
Figure 5. A word cloud of the words most commonly included in written submissions
!103
The images uploaded for the illustration tasks are perhaps even more impressive
in their creativity and whimsy, and in the tremendous effort that participants evidently put
into their work. There was a wide range of visual styles represented, as well as a
multitude of visual media, including digital illustrations, 3D renderings, hand drawing,
acrylics, watercolors, collages, and found objects. Appendix G includes a selection of
these images, although their full breadth and diversity is impossible to convey with only a
few examples.
The optional question asking participants for open-ended feedback also elicited an
impressive number of responses, with approximately half of the participants choosing to
fill it in. However, these responses were not coded for in-depth analysis, since the
question was included primarily for informational purposes and for improving the task
format and/or instructions when crowdsourcing the actual book, sentence by sentence.
The feedback, nonetheless, proved to be quite interesting and does warrant a brief
discussion.
Most responses expressed the participants’ appreciation of the HIT; in fact, many
workers just used this space to thank me for an enjoyable task. Some mentioned the effect
that completing the Hit had on them, often in ebullient terms (“Feeling very relaxed after
this joyful thing!”; “I was in a bad mood, and drawing Mr. Hashtag cheered me up!”).
Participants also liked the idea of collectively writing and illustrating a story, one
sentence at a time, and some stated that they wanted to try out this idea with their kids or
students (“I’m a preschool teacher and love to create and teach kids stories and crafts.
This HIT made me think to do some similar sort of project with my kids”).
!104
Some respondents used the feedback opportunity in order to convey their thoughts
on the value of writing a children's book about the Internet. The views were conflicting,
yet rather vehement. While some participants lauded the initiative, others did not see the
social or educational value of the idea, and some feared that it would expose children to
the Internet - and the dangers thereof - too early. Many respondents cautioned me that a
snail cannot use the Internet; several mentioned the very practical problem of typing
when you have no hands. However, the choice of Hashtag’s name was unanimously liked
and praised, although one person made the very rational point that “hashtags won't be
called that forever. As they were pound and number and comment before, so shall they
change again.”
A large percentage of respondents used the feedback opportunity to ask for more
HITs like this, some even providing their email addresses in addition to their Mechanical
Turk usernames. Many also specified that they wanted to read the full story when it
would be complete and asked where they could access it in its entirety. This is important
feedback and it made me realize that, when crowdsourcing the actual story, it would be
more fair and ethical to make the finished story available online and provide participants
with the URL so they can see how it all turned out.
Only a minority of respondents used the feedback space to provide constructive
criticism about the HITs. Of these, the most common observations concerned the need to
pay more for illustration tasks, to let them write as much as they want for the writing
HITs (this was implied in my instructions which only stipulated a minimum length, i.e. a
sentence, but perhaps that was not entirely clear) and to provide more information about
!105
the visual characterization of Hashtag the Snail. On this last point, respondents advised
that “you should have an illustration of your character, so any future artist contributing to
your project have a reference to work from.” Another, who identified as a “graphic
designer / illustrator” agreed: “For something like this the illustrator will need as much
data as you can possibly provide about the character you have in mind. Personality is one
of the most important traits in order to design Hashtag.”
Limitations
This study also presents several limitations that must be acknowledged. Firstly,
the fact that a series of related tasks were posted on the same platform - albeit spaced out
over a period of time - meant that participants could have seen and even participated in
multiple tasks. This is problematic for two reasons: one, if participants saw multiple
related tasks, they could have realized that certain key variables were being modified; and
two, if workers participated in more than one conditions, that violates the assumption of
independence of samples. As mentioned in Chapter 4, however, this strategy of
implementation was necessary in order to answer all my research questions and, given the
available options when creating and posting tasks on Mechanical Turk, there was no easy
way to avoid this challenge. However, this is a significant challenge that merits further
discussion and investigation. A recent study by Chandler, Mueller and Paolacci (2014)
has shown that researchers using Mechanical Turk are largely unaware of the possibility
that workers might participate in related experiments. The authors caution that, although
the Mechanical Turk worker pool can seem almost inexhaustible (especially in
!106
comparison to participant pools used in traditional research studies), duplicate workers
are more common than researchers assume. Beyond post-hoc data cleaning - which is
very common but sometimes problematic - the authors suggest a few strategies that
researchers can rely on to avoid this problem. A simple solution would be running
multiple related experiments through one single link within the same HIT, but this is not
always feasible - as in the case of my Mechanical Turk study - because such a strategy
does not allow the researchers to modify key variables like reward amounts and task
details. Other potential strategies, depending on a study’s research design, are to assign
Qualifications to workers who are prescreened (within the Mechanical Turk platform), or,
for those researchers with significant programming experience, to use the Mechanical
Turk API (Application Program Interface) in order to modify the HIT parameters and
exclude certain workers (Chandler, Mueller and Paolacci, 2014). Finally, and also
depending on the research design, an alternative option would be to use an external
research platform, such as Qualtrics, in order to prescreen Mechanical Turk workers or to
randomly assign them to different experimental conditions within the same study.
Another limitation of this study pertains to the accuracy of the observations
regarding quantity (operationalized as speed of completion) for writing tasks. With the
exception of three HITs (20c control, 5c $1 bonus, 10c no credit), all writing HITs were
posted on HitsWorthTurkingFor, a popular subreddit (message board) for Mechanical
Turk workers on Reddit. Once the HITs were posted, their rate of acceptance rose quickly
and precipitated the completion of the tasks, thus affecting final completion times.
However, since all but two of the writing HITs were featured on this platform, the effect
!107
was leveled across most of the tasks, somewhat alleviating the problem. This issue did
not come up for illustration tasks, since none of those HITs were posted on the forum,
presumably because the pay was deemed to be too low so they were not considered "HITs
worth turking for."
A third limitation has to do with the accuracy of the time spent variable. This
variable was computed automatically by Mechanical Turk based on the time elapsed
between task acceptance and submission. However, there is no guarantee that the users
spent all that time actively working on the task: they could have been multitasking or
could have even stepped away from their computer.
In regards to participant motivations, it is likely that the workers who enjoyed the
tasks most are the ones who voluntarily provided informations about their motivations in
the optional question. This could have skewed the results in favor of the “fun” factor, thus
explaining the large proportion of respondents who said they accepted the HIT for
pleasure and/or out of a desire to be creative. Furthermore, it must also be acknowledged
that participants also self-select based on interest and skills, meaning that people looking
to take a fun break or to be creative are the ones who accepted these HITs. In the same
time, however, a certain degree of self-selection is unavoidable given the invitational
rhetoric of crowdsourced art.
Finally, another important limitation pertains to the generalizability of the
findings beyond the Mechanical Turk population. In addition to the demographical
particularities of Mechanical Turk workers - addressed in Chapter 4 - there are other
important differences to take into account when comparing Mechanical Turk samples to
!108
traditional research samples. In a 2013 study, Goodman, Cryder & Cheema found that
turkers are less likely to pay attention to experimental materials, and more likely to rely
on the Internet to find answers, even when there is no incentive to submit a correct
response. Both of these considerations are relevant to the current study: the former, as the
authors observe, can reduce the statical power of the experimental research, while the
latter can help explain why so many workers cheated on the illustration task by uploading
pictures from the Internet. Finally, studies show that Mechanical Turk participants also
have idiosyncratic attitudes about money, that are not representative to those of a normal
population but are in fact similar to the attitudes of student populations (Goodman,
Cryder & Cheema, 2013). This is a very interesting observation, which could play a
significant role in terms of the impact that financial rewards had on the present study’s
dependent variables.
!109
CHAPTER 6:
CROWDSOURCING CREATIVE TASKS
ON ALTERNATIVE PLATFORMS
In the initial version of my research design, in addition to the empirical testing of
the effects of various incentives structures, I also aimed to investigate the impact that a
platform’s social and technical features might have on the process of creative
participation. In order to answer this research question, I attempted to crowdsource the
same children’s book on two other online platforms beyond Mechanical Turk: as a public
collaboration on HitRECord, and, respectively, as a participatory online course on Peer 2
Peer University (P2PU). By comparing the process of creative participation across the
three platforms (Mechanical Turk, hitRECord, and P2PU), my goal was to pinpoint the
specific features that support – or, alternatively, hinder – effective participation and
collaboration online.
On the one hand, participation in online communities is shaped by the social
characteristics of that group, such as the relationships between its members, their
perceived common purpose or collective identity, and the group communication and
collaboration protocols that determine their interactions. On the other hand, creative
participation is simultaneously influenced by the technical features available in these
online communities. Thus, the nature and quality of participation in online spaces can be
significantly shaped by the communication tools available to participants, the interface of
the platform, or the design of communication spaces. By launching and studying the
!110
same creative participation project on three different online platforms, I hoped to achieve
an insightful comparative analysis of the social and technical features that determine the
nature and quality of participation. From a research design standpoint, I used the same
children’s book as a starting point on all three platforms in order to minimize differences
between the three implementations.
Unfortunately, in spite of a concerted effort to promote the projects and engender
participation, the attempts to crowdsource the book on hitRECord and on P2PU were not
successful, failing to elicit sustainable participation. This chapter will describe both
implementations and will discuss some key takeaways that can be derived from these
experiences. Although these attempts failed to produce the desired effects - and thus were
not able to facilitate an effective comparative analysis across platforms - it is worth
dissecting them here, as the lessons learned are both instructive and insightful.
Importantly, these unsuccessful implementations also brought up several interesting
questions that I hope to explore in my future research, especially in regards to group
communication in online creative communities, the role of momentum in collaborative
projects, and the emergence of leadership in online groups.
hitRECord
Platform Description
Founded by actor Joseph Gordon-Levitt in 2005, hitRECord is officially
described as “an open collaborative production company. Think of it as a studio where
artists collaborate on projects together and remix each others work with the potential to
!111
contribute to money-making productions” (hitRECord, n.d.). The site invites its members
to submit original audio, video, text, or images, with the understanding that their work
may be remixed by other members of the community.
Under the guidance of Gordon-Levitt and his production team, the hitRECord
community has released many successful media products such as short films, anthology
books and music albums. Most recently, the site has also facilitated the collaborative
production of a TV series, hitRECord on TV, a variety show hosted by Gordon-Levitt and
featuring the multimedia contributions of hitRECord community members.
One of the most intriguing and unique features of hitRECord is their participatory
approach to profit sharing. For every monetized project produced via the community
(including hitRECord on TV), hitRECord retains 50% of the profits; however, in order to
determine the distribution of the other 50%, staff members put together a proposal for
splitting the profits, making sure to credit all contributing artists. The profit sharing
proposals are then posted on the site, so that they can be reviewed and openly discussed
by the community. After 2 weeks of commenting time (4 weeks for hitRECord on TV),
the proposal is amended as per the community members’ input and a revised version is
posted, giving users an additional week to review and respond to the revised proposal.
Once this last version is approved, the contributing artists are sent their remuneration in
the form of a check. While this approach to profit-sharing is by no means perfect -
especially considering the fact that hitRECord keeps 50% of the revenue by default - its
is nonetheless laudable for its transparency and for the participatory nature of the
decision-making process.
!112
Project Implementation
After familiarizing myself with the hitRECord community and participating in
other members’ creative initiatives, I launched the children’s book project in the form of a
collaboration , starting with the initial sentence about Hashtag the Snail finding another
10
snail’s shell. The title of the collaboration explicitly indicated the types of contributions
that were needed (“Add a sentence or an illustration to a children's book about Hashtag
the Snail!”). These were also reiterated within the “Request List” on the main project
page, which specified that I was looking for writing and illustration contributions. The
full text of my task instructions (i.e. the description of the collaboration) is reproduced in
Appendix H.
Once the collaboration was posted, I promoted it internally, within the hitRECord
community, by cross-mentioning it in other users’ related collaborations. As I will discuss
towards the end of this section, the platform does not offer other means to publicize or
promote one’s work within the community; collaborations can only be found through
keyword searches or by their popularity ratings. Externally, outside the hitRECord
platform, I promoted my collaboration via social media, on Facebook, Twitter & Reddit
(in the subreddit devoted to hitRECord membership).
"In"hitRECord"terminology,"a"“collaboration”"represents"a"series"of"“records”"(audio,"video,"
10
text"or"image"Biles)"that"relate"to"the"same"project."
!113
Results
Within the span of a month, the collaboration elicited 14 total contributions: 6
written submissions and 8 illustrations (see Fig. 6). The written contributions were
submitted by different community members, but 7 out of the 8 illustrations were
submitted by the same artist, a young woman going by the username of RandomShyGirl.

Figure 6. A screenshot of the hitRECord collaboration, featuring some of the submitted
records
!114
However, participation subsided after the first two weeks and the last contribution
was posted 28 days after the launch of the project. Trying to keep the momentum going, I
was very active in the hitRECord community, participating in others members’
collaborations (especially ones that were related to the Hashtag the Snail collaboration,
by theme or by format) and commenting on others’ records and projects. Unfortunately,
in spite of these efforts, participation did not pick up and no new records were posted on
the collaboration.
Key Takeaways
An important conclusion that came out of this experience concerns the crucial role
played by a platform’s technical features in supporting grassroots participation. In my
view, the limited sustainability of participation and collaboration on the hitRECord
platform can be explained - to a significant degree - by the lack of certain key features,
especially in regards to member-to-member communication and project visibility. For
being an “open collaborative production company” - as the site’s founders claim -
hitRECord does not actually offer the proper tools and interface to facilitate the
collaborative development of grassroots projects launched by members of the
community.
Looking at the platform’s design and affordances, the most significant problem is
the lack of communication tools that would enable site members to talk to one another
and engage collaboratively in the creative process. There is no support for private
(member-to-member) messages, no group messaging capabilities and no other
!115
communication channels like forums or discussion threads. In fact, the only way to
communicate with other members is by posting a comments on submitted records.
However, users do not have the option to enable notifications for posted comments
(either via email or on the site dashboard), so unless you check the comment thread for
every individual record in a regular manner, you would not know whether a comment has
been posted. I only had 14 records posted on my collaboration (not a large number at all,
compared to the more successful collaborations on the site) and it was still extremely
frustrating to have to check each one of them individually, in order to determine if
someone had been trying to reach me. For example, RandomShyGirl, who illustrated
many of my sentences, had tried to get in touch with me to let me know that she would be
delayed in illustrating the next sentence: her brother’s computer, which she was using,
had stopped working. She posted this message as a comment on one of her previous
illustrations, but even though I was logging into the site daily, I did not see the comment
until much later, when a response would have been irrelevant and unhelpful.
The lack of notification options - coupled with the fact that comments are posted
on individual records (i.e. submissions) and not on the main collaboration page - is, I
believe, a significant obstacle preventing efficient communication between members. In
fact, when I was trying to establish a presence in the community and, later, to stimulate
participation in my own project by commenting on others’ collaborations, it was very
seldom that I received replies to the comments I posted. In all likelihood, this was
because members had not actually seen my comments, especially if they were posted on
older projects or records. Furthermore, as the poster of the comments, I had to keep
!116
checking back on every single comment I had posted in order to see if I had received a
reply (I actually made a spreadsheet keeping track of all records I commented on, with
the relevant URLs!).
Having looked, to no avail, for other communication options between site
members, I turned to the hitRECord FAQs, where one of the questions was “Where can I
communicate with other members of the hitRECord community?”. In spite of the glaring
inefficiency of the commenting feature as a means of communication, the site insisted
that “the best way to communicate with other members of hitRECord is to comment on
their RECords” (hitRECord, n.d.). The FAQs further specified that “there is also a
CONVERSATION PAGE you can participate in here: http://www.hitrecord.org/forum/
recent”. “Conversation” sounded good, but my hopes were soon dashed when I realized
that the conversation page was merely a list of the most recent comments posted on any
and all records on the site. While there is an option to sort these comments by “most
active” and “most popular” in addition to “most recent”, there is no way to search for a
particular record, collaboration or member that you interested in, thus rendering the
conversation page effectively useless.
!117
Figure 7. The hitRECord conversation page
Beyond the dearth of effective communication options - which is, indeed, a
crucial problem for a platform focused on collaboration and creative dialogue - another
design- related shortcoming is the lack of visibility for grassroots collaborations.
Specifically, collaborations are sorted by popularity, so the first few pages (as well as all
“Featured” collaborations) are those launched by Joseph Gordon-Levitt and the
hitRECord staff members. This makes it extremely difficult for regular users like me -
especially those that are just starting up on the site - to have their collaborations noticed
!118
and to generate participation. In response to the question of “How do I get my work
noticed?”, in the hitRECord FAQs, the site admins suggest two courses of action:
1) Contribute to active collaborations. Maybe one or more of your RECords are a
good fit for a collaboration that the hitRECord community is already working on?
To find active collaborations, check out our Featured page ( http://
www.hitrecord.org/featured ). Take a look around and see if any active
collaborations stand out to you! There are all sorts of collaborations happening at
any given time… 2) Remix the work of other artists on the site. Remixing is a
huge part of our creative process - it’s how we make things together. (hitRECord,
n.d.)
What is remarkable about this answer is the fact that starting your own project
(i.e. your own collaboration) is not even presented as an option. Rather, this approach
encourages users to contribute to active collaborations rather than launch new ones. This
is concerning for two primary reasons: one, it discourages the grassroots development of
new projects that are meaningful to its users and two, it once again showcases the
collaborations started by Gordon-Levitt and his team (“To find active collaborations,
check out our Featured page”, which showcases all of their projects), thus coopting
community members’ creative efforts for the benefit of the hitRECord production
company rather than encouraging members to pursue their own interests and ideas.
This last conclusion is a crucial one, illustrating the role of celebrity status in
attracting and sustaining creative participation online. As mentioned elsewhere within
this dissertation, the public profile (and subsequent following) of the alpha artist in
!119
crowdsourced art projects is not to be underestimated. Oftentimes, as we have seen in
artistic initiatives started by Ridley Scott, Miranda July or Radiohead, it is this celebrity
status that guarantees the projects’ success and widespread participation. Joseph Gordon-
Levitt’s hitRECord is no exception. The actor has “unshakable cred" among the 18-to-34-
year-old demographic (Wiegand, 2014) - to which the majority of hitRECord members
belong - and, furthermore, enjoys huge popularity on social media, where most of his
posts are hitRECord-related. While Gordon-Levitt’s star appeal has certainly paved the
way for the success of the platform, it also conditions participation on the site in often
problematic ways. Specifically, because of the design issues mentioned previously,
visibility for new, user-launched projects is low; the projects that are consistently
spotlighted and that generate - by far - the most participation are those started by Gordon-
Levitt and the hitRECord staff members. As discussed earlier, the technical features that
would support a more bottom-up, equitable and collaborative production process are
missing, which lends further credence to the observation that the platform is primarily
designed to serve the production needs of Gordon-Levitt and his team. In fact, the site
was initially envisioned by the actor and his brother, Dan, as a place to post their own
videos, music and short stories and get feedback on works in progress; only later did they
open it up for community input and collaboration (Rosenberg, 2014). In many ways, the
platform still feels like its initial incarnation, where participation revolves around the
creative needs of the actor and his staff.
This imbalance of power is problematic in terms of the ethics of participation and
collaboration. In particular, given the fact that Gordon-Levitt’s projects are monetized
!120
and he keeps at least 50% of the profits - as per the hitRECord terms of use - one could
leverage the accusation that the platform is designed to give the actor and his production
team access to a large supply of talented and cheap labor, while simultaneously
cementing his image as a “magnanimous spirit” and a “polymath” (Caramanica, 2014).
The critique of crowdsourcing as labor exploitation, which is often wielded in the context
of commercial crowdsourcing projects (Brabham, 2013), is one that Gordon-Levitt is
decidedly trying to avoid. His rhetoric - on social media and in press interviews - and
even his hitRECord username (RegularJoe) can be seen as attempts to downplay his
celebrity status and emphasize the open, democratic and collaborative image of
hitRECord. Nonetheless, in spite of these efforts, and given the site’s lack of support for
communication and bottom-up collaborations, the platform often feels - in the words of a
New York Times columnist reviewing hitRECord on TV - as a “vanity project masking as
generosity” (Caramanica, 2014).
Peer 2 Peer University (P2PU)
Platform Description
For the third iteration of my children’s book project, I chose to implement it
in the form of a participatory and not-so-massive MOOC (massive open online
course). A key rationale in selecting this particular vehicle - beyond its potential to
facilitate a more nuanced and diversified comparison of the role of social and
technical features in shaping online participation - was my desire to test the
feasibility of an open online course as a site for creative participation and
!121
collaboration. In a sense, the aspirations of MOOCs and crowdsourced art are not
unrelated, as they embody the same inclusive ethos of large-scale and global
engagement by means of online technologies. But beyond this shared ethos, I
thought a MOOC would be an appropriate platform for implementing this project
because it could help create the enabling conditions for meaningful participation:
the contextual facilitation of a participatory culture, and the provision of an
adequate platform for interaction and collaborative activity. Furthermore, the
choice of an educational experience as a vehicle for creative participation is
congruent with the ethical aspirations of creative participation, as it equips
contributors with the relevant knowledge and understanding needed to become
fully invested collaborators, rather than ingenuous contributors.
After much research into online course platforms, I opted to launch my
course on Peer 2 Peer University (P2PU). Founded in 2012, P2PU is a nonprofit
open education project that allows anyone to both teach and take classes, creating
“a model for lifelong learning alongside traditional formal higher
education” (Peer 2 Peer University, “About”, n.d.). Beyond their nonprofit status
and their admirable Creative Commons copyright model, P2PU was appealing
because of their values, which were perfectly in line with my ambitions for this
project: openness, community, and peer learning (Peer 2 Peer University, “What
We’re All About”, n.d.). Significantly, their focus on connectivism and non-
hierarchical learning represented a good fit for my pedagogical style. And finally,
choosing P2PU instead of for-profit giants Blackboard or Coursera allowed me to
!122
distance my work from the more concerning aspects of MOOCs, especially as
they relate to copyright, monetization, and the hierarchical “sage on stage”
pedagogical model; P2PU, on the other hand, is affiliated with the Digital Media
and Learning (DML) Research Hub - a community I am familiar with, and whose
work I sincerely respect.
Project Implementation
After participating - as a learner - in several P2PU courses (including a very
helpful one on best practices for designing P2PU courses), I designed, drafted and
published my course, titled “Crowdsourced Art: A Participatory Exercise in Collaboration
and Collective Creativity.”
The course was structured as four modules, implemented in consecutive weeks. In
its initial draft, my course was longer and more intensive, requiring more effort on the
part of the learners; I had planned to include weekly readings (with response prompts)
and scaffolded introductory assignments. However, after participating in other successful
courses on the site and consulting with the P2PU Learning Lead, Vanessa Gennarelli, I
decided to simplify my course in order to encourage more participation and lower the
threshold of engagement. The four-week, four-module design of the course thus
represented a simplified and abridged version.
The first module introduced the the concept of crowdsourced art (with links to
key examples and further reading on the topic), and described the children’s book project,
as well as its rationale and charitable purpose. As “homework” for this module,
!123
participants were asked to introduce themselves to the group and to talk about why they
were interested in this project and/or in the topic of crowdsourced art, and what they
hoped to get out of it.
The second module consisted of the collaborative writing of the narrative, part by
part, exquisite-corpse style, on a shared Google Drive document (accessible and editable
without the need to sign into a Google account). The instructions for this module
reminded participants of the target age group and approximate length of the storybook,
and encouraged them to use the dedicated P2PU comment board as a forum to exchange
thoughts and concerns about the overall plot or the writing process.
Module #3 facilitated the collaborative illustration of the story, asking participants
to sign up for illustrating one or more sentences on the shared Google Drive document
and to then upload the resulting drawings onto the comment forum. This module also
included a discussion of the importance of visual consistency, encouraging participants to
be mindful of this issue and to build on the contributions posted by others on the forum.
Finally, the fourth module represented a wrap-up of the project and a discussion
of final details, such as the title of the book, cover illustration and design, and best
options for publishing and publicizing the book online. Significantly, this last module
also included a request to complete a brief reflection survey - hosted on Qualtrics - about
participants’ experiences in the course. The survey contained four questions:
1. What motivated you to enroll in this course?
2. What did you enjoy most about this experience?
3. What could be further improved?
!124
4. Any other comments or feedback?
Preceding these modules was an “About” page (Appendix I), which included an
introduction to the course and to myself as the facilitator, as well as laying out a few
community guidelines that I hoped would encourage polite, ethical and meaningful
engagement for all participants in the course. The “About” page conveyed the rationale
behind my interest in crowdsourced art and, consequently, the genesis of the project
itself:
This initiative was born out of my belief that creativity should not be - and was
never meant to be - the prerogative of a few chosen individuals. Nor should
creativity be limited to the institutional space of museums, galleries, and art fairs.
The Internet provides a unique platform to engage a potentially global community
in the creation of collaborative art, and I think there’s something pretty special
about that.
I hoped that the explicit framing of the project in these terms - as well as the
emphasis on community and collaboration, which I purposefully insisted upon, when
introducing the guidelines of participation - would help facilitate a vigorous participatory
culture among the course members. I also hoped that by sharing my passion for this topic,
and my sincere belief in the significance of online creative collaboration, participants
would embrace that passion and become more invested.
Recruitment and promotion efforts were varied and concerted; in fact, due to the
support of P2PU staff, recruitment efforts for this course were much more effective than
my independent promotion of the hitRECord collaboration. As a learning site affiliated
!125
with the Digital Media and Learning Research Hub, P2PU is very supportive of research
initiatives, stating that they “ will actively promote and support research, especially
action research, on the P2PU model and make it a priority to share information and
reflections with the larger community on a regular basis” (Peer 2 Peer University,
“About”, n.d.). My experience confirmed this commitment and I am very grateful for the
promotional help I received from P2PU, and especially from Bekka Kahn, the P2PU
Community Manager, and Vanessa Gennarelli, the site’s Learning Lead. They invited me
to write a blog post about the course, which they published on the P2PU blog, they
showcased my course in the “Community Picks” area of the site, and they spread word
about it via their social media accounts. I also used my own social media accounts to
promote the course. In particular, as a result of the tweeting activity of P2PU staff, Henry
Jenkins and myself, the initiative achieved significant visibility on Twitter.

Results
A total of 40 participants signed up for the course. Most of them, however, did not
actively engage with the course material or discussion: only 6 learners commented in the
course forums, 6 contributed to the story on the shared Google Drive document, and 3
participated in both commenting and writing the narrative. There were 13 contributions
on the Google Drive document (from 6 individual users), although - on a admittedly
subjective level - the half-finished story that emerged was badly written, full of
contradictions and repetitions, and overall rather lackluster. Only one illustration was
!126
posted, but it represented an image concept for the main character rather than the
illustration of a scene from the narrative.
The course was originally envisioned to run as a 4-week-long experience, running
from September 15 to October 12. In view of its reliance on the synchronous creative
contributions of multiple people, its scheduling within a pre-established timeframe was
crucial. Moreover, I had hoped that framing the course as a relatively easy and quick 4-
week commitment would help alleviate the issue of attrition, which is a known problem
for free open online courses, and one that I certainly feared. However, after the first 3
weeks passed with very little participation - and, consequently, very little progress being
made on the collective story - I decided to increase the time allocated to the writing
process by another two weeks and I sent an announcement to the entire class, letting them
know about the extra time and encouraging them to jump in on the story. I did not want
potential contributors to be discouraged by the limited time frame and, in making this
decision, I was further motivated by the following comment on the forum, posted by user
dirkcuys:
I feel a bit of anxiety contributing to the story. I want to contribute ideas, but I
don't want to highjack the story or completely derail it. Maybe we could have two
passes through the story so that the stakes don't feel that high the first time and
that we can make it more coherent when we do the second pass?
Dirkcuys’ idea was a good one, and it was seconded by another participant on the
forum. It required more time, which we now had, but it also required a lot more
participation, which we did not have. With such low participation rates, it was difficult to
!127
get the first draft of the story going, so unfortunately there was no opportunity for a more
refined editing stage.
On my end, as the course facilitator, I did all I could to encourage participation:
framed the entire experience as an exercise in participation, community-building and
collective creativity; replied promptly to all messages posted in the course forums;
extended the time allotted to story development; communicated with the entire roster on a
regular basis, twice a week; and, finally, sent three pleas for participation, keeping it in
very positive terms (“we just need one big push”, “we’re halfway done with the
writing”). But since participation did not pick up at all after my messages, I had to stop
reaching out to the group because I did not want to come across as pushy or irritating.
Key Takeaways
The limited success of this initiative brings up important questions about the
sustainability of participation in online courses, and particularly in online courses that
rely on creative collaboration. Attrition is a known problem for MOOCs; the high dropout
rates are often invoked by MOOC opponents as proof of the inefficiency and unfeasibility
of such courses as true educational vehicles (Onah, Sinclair & Boyatt, 2014). Although
my P2PU course was not massive per se, I was indeed worried about the issue of attrition
and, in designing the course, I took preemptive measures to try to prevent this situation
from occurring. I abbreviated the course down to four weeks, in order to maintain
momentum and lower the perceived threshold for participants’ time commitment. I asked
participants to work on one overarching project rather than separate mini-projects for
!128
each module, to try to maintain their interest and deepen their commitment to the course.
I even included a heartfelt message about the importance of ongoing participation in the
community guidelines on the course’s About page (Appendix I): “Please don’t leave the
course before it’s over, because we’re counting on you for the collaborative children's
book project! And we would miss you, really.)”
Nonetheless, these measures were not sufficiently effective in combating attrition
- although, in this case, attrition did not translate to people dropping out of the course, but
rather to a significant decline in participation after the initial week. Because the course
had only 40 enrolled students, this decline in participation was highly significant and
detrimental to the success of the initiative. However, as I have argued elsewhere (Literat,
2015), it does not always make sense to compare MOOC attrition levels with the attrition
levels of traditional college courses where students have clear incentives to complete
coursework. Since most MOOCs are free and do not offer credit, lurking or auditing is
more accepted than in a traditional classroom environment, and participation is not
subject to the same normative and financial pressures to finish and do well in a course.
Indeed, research shows that many MOOC participants sign up out of curiosity or
enjoyment, and do not even intend to complete quizzes or assignments (Rivard, 2013).
Thus, it is important to acknowledge that learners engage with MOOCs in a
variety of ways, and for different reasons. The P2PU course illustrated a variety of
engagement levels. While most learners who signed up for the course chose to participate
by lurking (or not at all), there were a few that were very involved. For instance, the most
active participant, who went by the username of simonrae, was extremely involved on the
!129
course forums and in building the story, offering structural input and encouraging others
to participate as well. This wide spectrum of engagement modes echoes my overarching
conclusion about the need for a more nuanced framework of participation that
distinguishes between different degrees or levels of engagement. The P2PU environment,
for example, offers opportunities for lurking but also for deliberation and structural
participation (as illustrated, for instance, by dirkcuys’ suggestion about adding a second
editing phase to the writing process), thus showcasing the crucial distinctions between
receptive, executory and structural modes of participation.
Another interesting dynamic that this experience highlighted is the role of
leadership in creative collaborations, and the fine balance between collaboration and
coordination. On the one hand, there is an aspiration - in terms of aesthetics but also
ethics - to create an environment that is utmost collaborative, participatory and inclusive.
But on the other hand, there is a practical need to have some sort of control or
coordination regarding the collaboration process, thereby creating leadership structures
and, in a way, countering the aforementioned ambitions of inclusive, nonhierarchical and
quintessentially collaborative participation. As the facilitator of the course, but also as a
scholar who genuinely believes in the empowering potential of nonhierarchical creative
collaboration, I felt torn between the practical need to assert control over the creative
process and, respectively, a desire to maintain a horizontal and open creative environment
that is deeply respectful of participants’ ideas and contributions. In light of my own
pedagogical beliefs and P2PU’s instructional model which is based on peer-learning, I
saw myself as the facilitator of a collective and collaborative learning experience, rather
!130
than a transmitter of knowledge. This approach, moreover, seemed to be a natural fit for a
participatory course on collaborative art. In fact, I explicitly outlined my position in the
community guidelines on the About page: “Our course is designed to reflect the P2PU
values: openness, community and peer learning. I am here to guide you along and to
provide you with support and feedback if you need it, but the most valuable insights and
experiences will come from your interactions, your collaboration, and your collective
development as a creative community.”
At times, however, I questioned the effectiveness of this approach, particularly
when I felt that the participants themselves were expecting me to take full control of the
creative writing process (beyond facilitating their participation and providing feedback).
When participation stalled, simonrae, the most active course participant, posted the
following comment on the forum:
Really very sad about this ... Hashtag seems to have run out of slime! Just
withdrawn into his shell and gone to sleep. Or maybe, she’s lurking, waiting for
someone to do something! That would be nice ... for someone to take manage the process. As I can see it, the story needs to be finished, then edited, then 'signed-off'
before we choose which line to illustrate. Then we all get down and dirty with the art-work. Then someone puts it all together and gets it published! Easy.
!131
While I certainly appreciated his perspective, I was also taken aback by the
implication that I was not a successful facilitator, especially since I had been extremely
active in communicating with the group, responding to comments and trying to
encourage participation. Furthermore, I had made it clear, in the course description and
in individual modules, that I would take charge of the editing and publishing process. I
just did not understand what more I could do, given the lack of actual contributions to
the story; due to the nature of this collaborative experiment, I could not write and
illustrate the story myself without others’ contributions. From an analytical perspective, I
really appreciated simonrae's comment, specifically because it speaks to the above-
mentioned tension between collaboration and coordination, between the horizontal and
the vertical. This tension is particularly pronounced in a learning-centered environment
like P2PU (albeit a very open and less hierarchical learning environment), where certain
lingering expectations regarding leadership and teacher-student dynamics might
complicate attempts to create a truly participatory and collaborative environment. As
Elizabeth Losh (2014) reminds us, the tension between contrasting pedagogical models -
primarily, “sage on stage” versus “guide on the side” - becomes all the more prominent
in the case of online courses, and MOOCs in particular. Facilitators of online courses
must therefore consider “how the interface of the Internet delivery system itself
functions and how their online audiences are constituted when designing modular user
experiences” (Losh, 2014, p. 114). This is a valuable observation that is highly
applicable to my own experience with the P2PU course, where I attempted to emulate
the “guide on the side” model, in accordance with the values of participatory and non-
!132
hierarchical project-based learning, whereas course participants were most likely
expecting (and looking for?) a more traditional “sage on stage” learning experience.
Seen from this perspective, my experience on P2PU carries significant
implications for current educational debates regarding the potential of online courses to
stimulate deep engagement, interaction and collaboration. While some commentators
praise MOOCs for extending educational opportunities in a more open and accessible
fashion (Friedman, 2013; Tabarrok, 2012), others criticize this trend as a threat to current
models of higher education and a low-quality substitute for traditional learning
(Crawford, 2013; Kolowich, 2013; Vaidhyanathan, 2012). Nonetheless, both critics and
supporters can agree that the MOOC phenomenon is an incredibly significant
development, with crucial implications for the future of higher education; therefore, it is
vital to know more about the specific kinds of learning and engagement that MOOCs best
facilitate. So far, MOOCs have been used mostly for knowledge delivery at a massive
scale, relying on filmed lectures and allowing for very little interaction between the
students or between the instructor and the students – indeed, this lack of opportunities for
participation and collaboration, especially when compared to the embodied classroom
experience, is also the root of most common criticisms of MOOCs. Given this context,
there is a need to look beyond the knowledge delivery model and assess the potential of
open online courses - massive or not - as platforms for participatory learning and creative
co-production. Can open online courses facilitate a form of participation and
collaboration that learners will find meaningful, empowering and worthwhile, or is such
!133
engagement unfeasible within the current incarnation of MOOCs and open online
learning?
!134
CHAPTER 7:
CONCLUSION
Situated at the intersection of communication research, Internet studies and art
history, the present research makes a significant contribution to our understanding of
crowdsourced art and collective creativity online. From a theoretical perspective, this
inquiry lays the groundwork for a more nuanced view of online creative participation, as
well as establishing a much-needed bridge between the study of participatory art - which
had focused exclusively on offline interactions - and, respectively, of online creativity.
Additionally, on a practical level, the results of this study shed light on the conditions that
most effectively foster creative participation online (as well as those that fail to do so),
investigating both the inputs and the outputs of creative participation. In view of the
versatility and growing popularity of creative crowdsourcing projects, these findings will
hopefully be useful not only to scholars, but also to companies, artists and practitioners
who would like to rely on open public participation for their creative projects.
This chapter will discuss the key conclusions - both empirical and theoretical -
that emerged from this work, as well as suggesting a few vital directions for future
research. While the observations presented herein arose from an investigation of the
cultural domain, they can nonetheless productively inform our understanding of
participation across domains and should therefore be considered from an interdisciplinary
standpoint, beyond the specific context of art and culture.
!135
Crowdsourcing Creative Tasks: What Works and What Doesn’t?
A primary empirical thrust of the current study was a desire to better understand
the impact that several key factors - specifically, financial rewards, bonuses, task purpose,
and attribution of credit - have on the process of creative participation online. As
discussed in Chapter 5, the findings showed that financial / extrinsic rewards (i.e. higher
pay and bonuses) are effective stimuli in encouraging participants to accept a given task,
whereas the strategies that work best to boost the quality of the submissions are offering a
bonus, mentioning the charitable or prosocial nature of the project, and giving
contributors authorship credit.
These last two conditions - purpose and, respectively, credit - are, in particular,
novel findings worthy of deeper consideration in future research. In terms of the larger
purpose of participatory projects, more research is needed on the impact of task
meaningfulness on the quality of participation. Interestingly, the relationship between
purpose and quality observed in this study - specifically, that the quality of the
submissions was higher when the stated purpose of the project was charitable - stands in
contrast to the findings of Chandler and Kapelner (2010), who concluded, also based on a
Mechanical Turk study, that the framing of a task as meaningful (in this case, labeling
tumor cells for cancer research) does not boost the quality of the resulting submissions.
There is a significant gap in the literature as to the role of authorship credit in
crowdsourcing projects - creative or otherwise - so this is a key area where future
scholarship is needed. The present study found that attributions of authorship deepen
participants’ investment in the context of creative projects, but it is uncertain whether the
!136
same conclusion would hold true in other contexts. In the same time, it is also important
to note that, as illustrated by the rise of crowdsourced art and participatory cultures,
notions of authorship are in flux; consequently, there is a need to account for new forms
of authorship, especially ones that are quintessentially collective or distributed. For
instance, Jonathan Gray’s discussion of distributed authorship in the context of cultural
paratexts (Banet-Weiser et al., 2014) is a strong example in support of this argument.
Another vital conclusion that came out of this study is that, when the task is
inherently interesting, internal incentives trump external ones. This finding - which
echoes researchers’ conclusions regarding intrinsic versus extrinsic motivations in the
field of psychology (Deci, Koestner & Ryan, 1999; Frey, 1997) and also more
specifically in crowdsourcing (Rogstadius et al., 2011; Zheng, Li & Hou, 2011) -
emerged both from the analysis of the statistical results and from reviewing participants’
self-stated motivations in the optional open-ended question. Upon a comparative
assessment of all three attempts to implement the children’s book project, the fact that
Mechanical Turk (the only platform where participation was paid) was the only
successful implementation might make it tempting to reach the opposite conclusion: that
extrinsic rewards are more powerful, being the only ones that effectively guaranteed the
success of the project. However, I want to caution against overestimating the role of the
financial reward in such initiatives. The fact that hitRECord and P2PU failed to elicit a
sufficient amount of participation should by no means lead to the conclusion that people
participate only when you pay them. Rather, the reasons behind the limited success of
these implementations are highly specific to the social and technical features of each
!137
platform and community; these observations have been discussed in detail in Chapter 6.
In addition, to pursue this comparative approach further, Mechanical Turk - from a purely
practical standpoint - provides a much larger pool of potential participants than either
hitRECord or P2PU. Therefore, strictly from the perspective of size, the chances that a
creative task would engender sufficient participation are much higher on Mechanical
Turk, the largest crowdsourcing platform on the Internet, than on other sites. Finally,
based on reviewing the responses to the optional question about participant motivations,
it is likely that, in many cases, the extrinsic reward - albeit a modest fee of a few cents -
encouraged participants to accept the task but then acted as a trigger for a much more
powerful set of intrinsic stimuli, such as pleasure, flow and the exercise of creativity. In
other words, you might accept an illustration assignment about Hashtag the Snail because
you are on Mechanical Turk looking to make extra income, but, once you start working
on it, it is not the 5 cents that drive you to meticulously paint every blade of grass and
color every cloud and bird in the sky: you do it because you enjoy it, as participants
themselves have stated in the optional questions.
Attaining a better understanding of participant motivations in crowdsourcing
projects should be a key priority for future research in this field (Brabham, 2013),
especially given the rising popularity of crowdsourcing platforms in both commercial and
academic contexts. Significantly, striving to better understand these motivations should
involve - perhaps first and foremost - listening to the participants themselves and getting
a clearer picture of how crowdsourcing fits into their daily lives. Too many times, in
research on crowdsourcing and digital labor, these voices are critically absent from the
!138
scholarship written about their very activities. In particular, there is a need for
ethnographic research that would facilitate a more human perspective on these labor
practices.
Several researchers have taken up this task, producing promising examples of a
more qualitative approach to crowdsourcing research. For Face in the Crowd, a new
project from Microsoft Research, an international and interdisciplinary team of
researchers led by Mary L. Gray are investigating workers’ experiences on three popular
crowdsourcing platforms: Amazon's Mechanical Turk, Microsoft’s Universal Human
Relevance System (UHRS), and MobileWorks, a startup with a social and entrepreneurial
mission. Their research focuses on US and Indian workers and employs a mixed methods
approach, combining ethnographic and quantitative research (Microsoft Research, n.d.).
While the Microsoft Research project is the most prominent and ambitious such
initiative, other researchers are also gathering rich ethnographic data on smaller and more
specific populations, such as full-time Mechanical Turk workers in the United States
(Williamson, 2014) and in India (Gupta, Martin, Hanrahan & O’Neill, 2014).
While the collection of ethnographic data about crowdsourcing workers is useful
in helping both researchers and employers learn more about these participants, what is the
outcome of such an investigation for the workers themselves? I believe this is the next
crucial question that needs to be addressed. In other words, “how does knowing more
about the lives of crowdsourcing platform workers … better equip us to consider the
political and ethical stakes of the very material labor digital capitalism
produces?” (Microsoft Research, n.d.).
!139
One possible answer is that gaining a more human perspective on the lives and
livelihoods of crowdsourcing workers paves the way for reform. For instance,
Williamson’s ethnographic research on Mechanical Turk workers in the United States
(2014) found that most of these laborers are not hobbyists; rather, they are trying to make
ends meet by working full time on Mechanical Turk, usually constrained by
circumstances such as short- or long-term disabilities, unemployment or the need to stay
home and care for children. Having a better understanding of these workers’
circumstances can - Williamson argues - spearhead the need for reform in terms of
minimum wage requirements on crowdsourcing platforms, and the establishment of
ethical policies meant to protect the workers used as samples in crowdsourced research.
Last but not least, ethnographic research - particularly in the vein of participatory
action research - can enable workers themselves to supply recommendations that would
facilitate ethical reform. Based on recent activity, it seems that workers are more than
happy to join the conversation and make suggestions that would improve their
experience. In addition to multiple message boards and online forums (such as
turkernation.com, mturkgrind.com, mturkforum.com and many active channels on
Reddit), Mechanical Turk workers have been participating in significant ways on
Dynamo (wearedynamo.org), an online community, launched in 2014, where crowd
laborers can mobilize for the achievement of collective goals. The platform features, for
11
instance, a letter-writing campaign to Amazon CEO Jeff Bezos, a document outlining
It is interesting to note that, although the content on Dynamo is contributed by Mechanical
11
Turk workers, the platform is designed and run by a team of researchers from Stanford University
and the University of California, San Diego.
!140
ethical guidelines for academic researchers (which has been signed, so far, by 159 MTurk
workers and 55 academics), a list of campaign ideas that can be uprooted or downvoted
by community members, and a forum where these ideas can be discussed in length, and
which currently includes talks of unionizing. While it is too soon to assess the outcomes
of such initiatives, Dynamo - and platforms like it - are certainly an exciting development
and a step in the right direction.
!
Figure 8. Introduction to the letter-writing campaign to Jeff Bezos, as posted on Dynamo.
Workers ask to be treated as human beings, to be presented as skilled - not cheap - labor,
and to have systems of recourse in dealing with unfair requesters.

Reassessing Participation and Its Outcomes
A conclusion that I keep referring back to, throughout my work, is that
participation is not a blanket term and should not be conceived and talked about as such;
instead, there is a need for a more nuanced view, emphasizing agency and empowerment.
!141
This, in my view, is the most significant theoretical conclusion that emerged from my
work on crowdsourced art, and one that has shaped both my current and my future
research.
In the same time, having implemented these three creative projects as part of my
empirical study, I now realize that the distinctions between the different modes of
participation are less clear-cut in practice than in theory, and that the design choices
available to initiators of creative participation projects often hide an important struggle
between, on the one hand, aesthetic and ethical wants and, on the other hand, practical
needs. Specifically, as mentioned in Chapter 6, I found that there is a significant tension
in crowdsourced art projects between a desire for openness and, respectively, a practical
need for control and coordination. My own experience as designer and facilitator of these
projects illustrates this tension. On Mechanical Turk, my tight control over the project
and very specific task instructions resulted in a successful implementation, but the mode
of participation was quintessentially executory and I lamented not being able to offer
contributors opportunities for structural participation and not allowing for deeper, more
meaningful forms of engagement. On P2PU, on the other hand, openness and
transparency were key priorities and I aimed to set up an environment that would
facilitate structural participation, but, in all likelihood, it was this openness that ultimately
contributed to the failure of the project; I had invited P2PU learners to participate
structurally, when in fact they were looking to participate in an executory fashion.
In light of these observations, I find myself pondering: is structural participation -
in the form of co-design and/or co-authorship - really feasible in crowdsourced art
!142
projects? Or, in order for a project to be successful, must the alpha artist always maintain
a certain degree of authority and control, thus rendering the project de facto executory?
Looking at the most impressive examples of crowdsourced art projects, it seems like the
most successful ones (e.g., Life in a Day, Learning to Love You More, The Johnny Cash
Project) are those where participation is executory - with very specific, narrowly defined
and pre-designed tasks - and the alpha artist is very involved in the creative process.
12
Quality is an important consideration here as well, since the alpha artist’s
involvement is most often necessary to uphold the standards of aesthetic quality. As
mentioned in the theoretical overview of crowdsourced art, the alpha artist often
functions as a curator or editor, wading through all the submissions and picking out the
ones that are most impressive and that fit best within the larger design of the project. Life
in a Day is a great example in this regard: the production team sorted through more than
80,000 user-submitted video clips to choose the footage that would make it into the final
95-minute film (National Geographic, n.d.). Then, through this lens, perhaps a better
question to ask is whether structural participation is feasible when the quality of the final
product matters, and when the standards of aesthetic quality are high. For many
participatory art projects, it is the quality of the participatory process that matters more
than the final output (if any) - although, even in the art world, such assessment criteria are
increasingly questioned by critics (Beech, 2008; Bishop, 2006a; Kester, 2004, 2011) -
Of course, beyond project design, another factor that most likely contributed to the success of
12
these initiatives - but that is very difficult to quantify - is the fame of the alpha artists involved. It
is not a coincidence that the projects exemplified here as most successful benefitted from the
strong fan following of the alpha artists (Ridley Scott & Kevin Macdonald, Miranda July; Aaron
Koblin & Chris Milk). This observation supports my earlier conclusion that, oftentimes, factors
that are not directly related to the design of the project - in this case, the artist’s public persona -
can have a profound influence on the ultimate success of the initiative.
!143
while in the more commercial sphere, such as that of crowdsourced filmmaking, the
quality of the final product is of an utmost importance. When the quality of the output
does matter, it is often difficult being in the position of the alpha artist and not
intervening in a hierarchical manner. In my experience with P2PU, I must admit it was
hard seeing the story go in a direction that I did not like. I found it difficult to stick to my
commitment of relinquishing creative control and I had to actively keep myself from
intervening in the story and altering its course according to what I believed was a higher
standard of quality.
13
But beyond the ways in which it refines our vocabulary and understanding of
participatory modes, it is important to note that theorizing the nature of digital
participation across socio-cultural spheres is more than a semantic or conceptual project.
Indeed, this theoretical redefinition is important for the ways it draws our attention back
to the stakes of participation, and to the outcomes of participation for the participants
themselves. Consequently, agency and empowerment – whether it is social, creative,
civic or educational empowerment – should represent core aspects of the discussion
around participatory cultures.
Furthermore, shining a spotlight on the outcomes of participation also implies that
these outcomes must be reevaluated and more critically assessed. In other words, we
must ask ourselves: is participation always empowering or do we just like to think so?
Several scholars have noted that a key problem with the discourses on participation - both
The P2PU story - at least up to the point where participation ceased - contained many narrative
13
elements that did not make sense (such as terrestrial snails swimming, snail shells being used for
clothes and shoes) or references that were not appropriate for the target age group (mentions of
cyber-security and obscure details about the history of computer-mediated communication).
!144
scholarly and popular - is that they are almost always seeped in utopia and idealism, with
participation being presented as an invariably empowering practice (Carpentier, 2009;
Fish et al., 2011; Jenkins & Carpentier, 2013; Kelty, 2012; Schaefer, 2011). Relying on an
interdisciplinary analysis, Chapter 2 suggested a few ways to address this issue by
capitalizing on the lessons learned in different theoretical domains. Political philosophy,
for instance, can remind scholars of participatory cultures to maintain a key focus on
power and on the dynamics of power (re)distribution. As Carpentier and Dahlgren (2011)
observe, it is important to keep in mind that the concept of participation is and has always
been inextricably linked to “the power logics of decision making” (p. 8). This emphasis
on power redistribution and decision making can shed new light on the outcomes of
participation, facilitating a more informed assessment of agency and empowerment.
Similarly, a key insight that can be gleaned from the scholarship on participatory art is a
dual awareness of form and content. In the art world, both form and content - as well as
the interplay between the two - are critically assessed. As argued in Chapter 2, I believe
that such an awareness would be extremely useful in evaluating, beyond artistic contexts,
the outcomes of participation as they pertain to both form (process) and content (end
results). Finally, as modeled by the literature on participation in politics and cultural
studies, it is crucial to assess these outcomes - agency, empowerment, efficacy - at both
the individual and the collective level. Indeed, it is only by considering both of these
layers that we can come to fully understand the overall impact of participatory practices
and the significance they hold for individuals and communities.
!145
Assessing the outcomes of participation for the participants themselves becomes
particularly important in the case of crowdsourcing and paid micro-labor platforms like
Mechanical Turk. As argued previously, the extremely low wages and limited protection
against exploitation creates a thorny ethical situation. Labor on Mechanical Turk has been
described as “intellectual piecework” (Schneider, 2015), and scholars have made linkages
between crowdsourcing and the problematic history of piecework in terms of gendered
and regional power imbalances (Microsoft Research, n.d.). Moreover, just like the piece
workers of earlier times, crowdsourcing participants also lack the employment
protections that apply to regular workers and lack official channels to mobilize and
demand better labor standards (Williamson, 2014). In a sense, as Trebor Scholz recently
concluded, “the crowdsourcing industry is wiping away a century of labor
struggles” (Scholz, quoted in Schneider, 2015).
On the positive side, researchers and activists have begun thinking of ways to
address these imbalances of power. In addition to giving crowd workers a voice through
communities like Dynamo and Turkopticon, new crowdsourcing platforms are being
designed with ethics as a key priority. For example, a new website called Prolific
Academic, from the University of Oxford, provides workers a set minimum wage of $7
per hour, as well as the ability to rate requesters; the crucial obstacle will be, of course,
attracting a critical mass of participants (Schneider, 2015). Another exciting possibility -
and one that has been bolstered by many an upvote on Dynamo - is the prospect of crowd
workers creating and running their own platform, that would address their grievances and
better respond to their needs. Furthermore, in the realm of academic research, there are
!146
several institutional structures that could be activated to help ensure the ethical treatment
of crowd workers. For instance, as Williamson (2014) suggests, journals can commit to
publishing only those articles that pays respondents at ethical rates; grant makers can
include stipulations about ethical pay and ethical treatment of workers; and, last but not
least, institutional review boards (IRBs) can create specialized guidelines for work that
involves crowdsourcing samples.
14
Finally, a discussion of the utopian, idealized view of participation would not be
complete without considering the participation gap. As part of the positive rhetoric
surrounding participation, we like to hold on to the belief that participation is
quintessentially inclusive, and that everyone and anyone can participate, should they
choose to do so. In the cultural domain in particular, there is an assumption that anyone
can participate in online art projects or write fan fiction or contribute footage to a
crowdsourced video, but that only some choose to accept this invitation (Banet-Weiser et
al., 2014). In reality, though, the participation gap remains a crucial issue, limiting the
scope and diversity of participation in creative projects. There is a need for a more
objective examination of the positive rhetoric surrounding technology as an equalizing
factor in enabling creative participation in crowdsourced art and beyond. But,
significantly, the participation gap is more than just access to technology, and
encompasses the social and cultural competencies and skills needed for meaningful
Human subject regulations upheld by IRBs are often out of touch with the specific issues raised
14
by Internet research in general and crowdsourcing-based research in particular. When they do
address such contexts, they can sometimes do more harm than good. The IRB at Pomona College,
for instance, explicitly waives its minimum rate of $15/hour for Mechanical Turk research, so as
to avoid participant coercion, which can occur as a result of paying participants at a rate that is
much higher than the platform average (Scheider, 2015).
!147
engagement in today’s world (Jenkins et al., 2006). We need to pay more attention to who
is included and who is excluded when it comes to creative participation online; although
it was outside the scope of this current work, further research is needed in regards to the
demographics of crowdsourced art participants. An important direction for future
research on this topic also concerns the cultural identity of the participants in
crowdsourced art projects. Recent empirical work in the field of crowdsourcing indicates
that oftentimes the crowd is not comprised of amateurs, as we would like to assume, but
rather of self-selected groups of professionals and semi-professionals (Brabham, 2010,
2011). It would therefore be significant to find out whether the participants in
crowdsourced art projects similarly belong to a self-selected group of artists and
professionals affiliated with the art world.
An Interdisciplinary Approach to Participation
As Christopher Kelty (2012, p. 23) observes, recent scholarship has popularized a
wide array of terms and concepts to account for the impact of the Internet and new media
on participation:
terms such as ‘peer production’ (Benkler 2006), ‘produsage’ (Bruns 2008), ‘the
wisdom of crowds’ (Surowiecki 2004), “prosumers/prosumption” (Toffler 1980;
Ritzer &Jurgenson 2010), the ‘network society’ (Castells 1996; 2001), ‘user-led
innovation’ (von Hippel 2005), ‘recursive publics’ (Kelty 2008), ‘creation
capitalism’ (Boellstorff 2010), ‘convergence culture’ (Jenkins 2006), ‘organized
!148
networks’ (Rossiter 2006 and Lovink 2005), ‘wikinomics’ (Tapscott and Williams
2006) or ‘networked publics’ (Varnelis 2008; boyd, 2009).
While the growing interest in participatory culture studies is certainly a welcome
development, it is a problem that “all of these terms refer to social media, the Internet,
software, fan cultures and ‘knowledge societies,’ but not to each other: they are the
scholarly equivalent of proprietary formats, they lack compatibility, interoperability or
convertibility” (Kelty, 2012, p. 23). Furthermore, Kelty is right in suggesting that the
reason for this lack of integration is - at least in part - the firm rooting of these terms and
concepts within specific disciplines. As such, each theoretical framework is shaped by its
respective disciplinary allegiances, but these disciplines are not being put in productive
dialogue with each other.
Given this context, I believe that there is a significant need for stronger
interdisciplinary integration in the field of participatory culture studies and I am eager to
take part in this endeavor. In my next research project, I aim to strengthen cross-
disciplinary bridges by working towards a better integration of the concept of digital
participation across domains of knowledge. I have already taken several steps towards
this goal. Chapter 2 lays the foundation for this type of comparative approach,
emphasizing the strengths of each perspective and suggesting a few opportunities for
cross-pollination and symbiosis; in the next stage, I plan to build on this work by
including a discussion of the educational domain as well and examining the relationship
between participation and learning in each of these fields. In addition, with my writing on
crowdsourced art and online creativity I have tried to illustrate the benefits of adopting a
!149
more interdisciplinary perspective, exemplifying the ways in which scholars and theorists
could build cross-disciplinary bridges (here, between new media studies and art history).
In fact, as testament to the potential of this interdisciplinary approach, my theoretical
framework - the pyramid of participation - has been getting a lot of attention from fellow
scholars and is now being used to examine participatory processes in a variety of
domains, such civic activism, education, and commercial crowdsourcing.
Finally, although I envision the next phase of my research as predominantly
theoretical, it is very important to me, as a scholar and an educator, to practice what I
preach. Given my personal outlook and my background in development and social
justice, I find it important to not only study participation, but to actually create
opportunities for meaningful, equitable and ethical public participation. While the
concept of a participatory society may be unattainable in its utopian, idealized form, but a
more participatory society is certainly attainable, and represents a worthy ideal to strive
towards.
!150
EPILOGUE:
A CHILDREN’S BOOK ABOUT THE INTERNET,
BY THE INTERNET
For the purpose of the empirical study on Mechanical Turk, it was necessary to
solicit contributions for the same sentence and the same illustration across all tasks, in
order to maintain all other conditions identical while modifying the independent variables
of interest. However, once the experimental phase of the study was complete, I found it
important to follow through with crowdsourcing the entire book, like I had attempted to
do on hitRECord and P2PU as well.
When crowdsourcing the actual book, paragraph by paragraph and illustration by
illustration, I made sure to implement everything I had found in the empirical study on
Mechanical Turk. Specifically, in order to boost the quality of the resulting contributions,
I offered a bonus for the best submission, I mentioned the charitable purpose of the
project, and I offered participants the opportunity to provide their name or nickname in
order to be credited in the final book. I also incorporated the feedback that was provided
in the optional question within the empirical study: principally, respondents had said that
they wanted to write more than a sentence (so this time I explicitly mentioned that they
can write as much as they wanted, with one sentence being the minimum) and that they
wanted more details about the project (so I provided background information as well as
more details about the purpose of this initiative).
!151
In addition, because the contributions had to build on each other in a logical
manner, I made the task instructions more specific, in order to ensure narrative and visual
continuity. For example, for each writing task, I provided a summary of the story up to
that point, as well as any details that would be relevant to continuing the narrative. For
illustration tasks, in order to facilitate visual continuity, I included an image of Hashtag
the Snail (which I had selected from among the illustrations received for the experimental
study), asking participants to try to adhere to this character design - as best as they could -
and to maintain the same color scheme, depicting Hashtag with a green body and a dark
pink shell.
When conducting the empirical work on Mechanical Turk, I had been concerned
about the ethics of my engagement with the workers, especially since I was paying them
so little and I felt that they were giving a lot in return and putting in a tremendous amount
of effort into the task. Of course, these feelings were also exacerbated by my deep
familiarity with the criticisms of Mechanical Turk as an exploitative platform and a
digital sweatshop. Furthermore, as discussed previously, I had wanted to extend
opportunities for more meaningful participation, but the design of the project - and of the
platform itself - limited the modes of engagement to executory participation. Therefore,
when crowdsourcing the entire book, in order to approach the process as ethically as
possible, I was very transparent about the parameters of the project; I offered everyone
the option of being credited in the final book; I increased all monetary rewards to the
highest amount I could afford (20 cents); and I paid bonuses to all participants whose
writing or illustrations I included in the final book. In addition, because Mechanical Turk
!152
workers often complain that they have limited ways to get in touch with requestors, I
created an email address (HashtagTheSnail@gmail.com) that served as an easy way to
communicate with me about any aspect of the project, whether before or after they
submitted their contributions. I also encouraged participants to let me know if they were
interested in seeing the entire book once it was finalized, since many workers had
expressed this desire in the feedback section of my experimental tasks.
After I gathered all the written and visual contributions needed, I edited and
designed the book in BookSmart, a free software by self-publishing company Blurb,
making sure to credit all the contributors at the end of the story. The final book, titled The
Adventures of Hashtag the Snail, is available to read at http://issuu.com/iliterat/docs/
theadventuresofhashtagthesnail/1 (full screen view recommended) and it is a true delight.
I could not be happier with the way it came out and I must admit that, although I was
optimistic to begin with, the quality, whimsy and artistry of the final book far exceeded
my initial expectations. For now, the online version of the book is private - i.e. only those
with the above link can access it - while I am looking into ways to distribute it in print
and raise funds for The Modern Story digital storytelling program.
!153
BIBLIOGRAPHY
Abbott, A. (2008). Co-opt this: Towards an appropriate language for the evaluation
and production of art under cultural capitalism. Unpublished Manuscript.
Retrieved from www.yvonnecarmichael.com/andyabbott.co.uk/files/
AACo-optThis.pdf
Adamowicz, E. (1998). Surrealist collage in text and image: Dissecting the exquisite
corpse. Cambridge, UK: Cambridge University Press.
Amabile, T.M. (1983). The social-psychology of creativity: A componential
conceptualization. Journal of Personality and Social Psychology, 45(2),
357-376.
Amabile, T.M. & Pillemer., J. (2012). Perspectives on the social psychology of
creativity. Journal of Creative Behavior, 46(1), 3-15.
Amazon.com (n.d.). Mechanical Turk FAQ. Retrieved from https://www.mturk.com/
mturk/help?helpPage=worker
Andrejevic, M. (2009). Exploiting YouTube: Contradictions of user-generated labor.
In P. Snickars & P. V onderau (eds.) The YouTube Reader. Stockholm:
National Library of Sweden.
Arnstein, S.R. (1969). A ladder of citizen participation. Journal of the American
Institution of Planners 35(4), 216-224.
!154
Banet-Weiser, S., Baym, N., Coppa, F., Gauntlett, D., Gray, J., Jenkins, H., & Shaw,
A. (2014). Participations| Part 1: CREATIVITY . International Journal Of
Communication, 8, 20. Retrieved from http://ijoc.org/index.php/ijoc/
article/view/2721/1117
Banks, J., & Deuze, M. (2009). Co-creative labour. International Journal of Cultural
Studies, 12(5), 419-431.
Barthes, R. (1979). From work to text. In J.V . Harari (Ed.), Textual strategies:
Perspectives in poststructuralist criticism (pp. 73-81). Ithaca, NY:
Cornell University Press.
Becker, H.S. (1974). Art as collective action. American Sociological Review, 39(6),
767-776.
Becker, H.S. (1984). Art worlds. Berkeley: University of California Press.
Beech, D. (2008). Include me out! Art Monthly 315, 1–4.
Beech, D. (2010). Don’t look now! Art after the viewer and beyond participation. In J.
Walwin (Ed.), Searching for Art’ s New Publics (pp 15-30). Bristol, UK:
Intellect Ltd.
Bishop, C. (2004). Antagonism and relational aesthetics. October 110, 51–79.
Bishop, C. (2006a). The social turn: Collaboration and its discontents. Artforum
(February issue): 179-185.
Bishop, C. (2006b). Introduction: Viewers as producers. In C. Bishop (Ed.),
Participation (pp. 10–17). Cambridge, MA: MIT Press.
!155
Bourriaud, N. (2002[1998]). Relational aesthetics. (S. Pleasance & F. Woods, Trans.).
Dijon, France: Les Presses du Reel.
Bourriaud, N. (2006). Relational aesthetics. In C. Bishop (Ed.). Participation (pp.
160-171). Cambridge, MA: MIT Press.
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.
Brabham, D. C. (2010). Moving the crowd at Threadless. Information,
Communication & Society, 13 (8), 1122-1145.
Brabham, D.C. (2011). The myth of amateur crowds. Flow, 13(6). Retrieved from
http://flowtv.org/2011/01/the-myth-of-amateur-crowds/
Brabham, D. C. (2013). Crowdsourcing. Cambridge, MA: MIT Press.
Browne, S. (2008). Crowd theory lite: 'The crowd' in participatory art and pop
economics. Circa, 126, 33-39.
Caramanica, J. (2014, January 17). ’HitRECord on TV ,’ based on the artistic
crowdsourcing site. The New York Times. Retrieved from http://
www.nytimes.com/2014/01/18/arts/television/hitrecord-on-tv-based-on-
the-artistic-crowdsourcing-site.html?_r=0
Carpentier, N. (2009). Participation is not enough: The conditions of possibility of
mediated participatory practices. European Journal of Communication,
24(4).
!156
Carpentier, N. (2011a). The concept of participation: If they have access and interact,
do they really participate?. CM-časopis za upravljanje komuniciranjem
[Communication Management Quarterly], 6(21), 13-36.
Carpentier, N. (2011b). Contextualizing author-audience convergences: ‘New’
technologies' claims to increased participation, novelty and
uniqueness. Cultural Studies, 25(4-5), 517-533.
Carpentier, N. (2012). Bastard Culture! How User Participation Transforms Cultural
Production [Review]. European Journal of Communication, 27(2),
211-212.
Carpentier, N., & Dahlgren, P. (2011). Introduction: Interrogating audiences:
Theoretical horizons of participation. CM-časopis za upravljanje
komuniciranjem [Communication Management Quarterly], 6(21), 7-12.
Chandler, D., & Kapelner, A. (2010). Breaking monotony with meaning: Motivation
in crowdsourcing markets. Technical report. University of Chicago.
Retrieved from http://www.danachandler.com/files/
Chandler_Kapelner_BreakingMonotonyWithMeaning.pdf
Chandler, J., Mueller, P., & Paolacci, G. (2014). Nonnaïveté among Amazon
Mechanical Turk workers: Consequences and solutions for behavioral
researchers. Behavior Research Methods, 46(1), 112-130.
Couldry, N. (2011). More sociology, more culture, more politics: or, a modest
proposal for ‘convergence’ studies. Cultural Studies, 25(4-5), 487-501.
!157
Crawford, T.H. (2013, April 29). Rediscovering the material world. The Chronicle of
Higher Education. Retrieved from http://chronicle.com/article/
MOOCsthe-Material-World/138787/
Dahlgren, P. (2011). Parameters of online participation: Conceptualizing civic
contingencies. CM-časopis za upravljanje komuniciranjem
[Communication Management Quarterly], 6(21), 87-109.
De Certeau, M. (1984). The practice of everyday life. Berkeley, CA: University of
California Press.
Deci, E. (1975). Intrinsic motivation. New York, NY: Plenum Press.
Deci, E., Koestner, R. & Ryan, R. (1999). A meta-analytic review of experiments
examining the effects of extrinsic rewards on intrinsic motivation.
Psychological Bulletin, 125 (6): 627–668.
Delwiche, A., & Henderson, J.J. (2012). Introduction: What is participatory culture?
In A. Delwiche, J.J. Henderson. The participatory cultures handbook (pp.
3-9). New York: Routledge
Downey, A. (2009). An ethics of engagement: Collaborative art practices and the
return of the ethnographer. Third Text, 23(5), 593- 603.
Eco, U. (1989). The open work. Cambridge, MA: Harvard University Press.
Fish, A., & Srinivasan, R. (2012). Digital labor is the new killer app. New Media &
Society, 14(1), 137-152.
!158
Fish, A., Murillo, L. F. R., Nguyen, L., Panofsky, A. & Kelty, C. M. (2011). Birds of
the Internet: Towards a field guide to the organization and governance of
participation. Journal of Cultural Economy, 4(2), 157-187.
Foster, H. (2006). Chat rooms. In C. Bishop (Ed.), Participation (pp. 190–195).
London, UK: Whitechapel/MIT Press.
Foucault, M. (1979). What is an author? Screen, 20(1), 13-33.
Frey, B. (1997). Not just for the money: An economic theory of personal motivation.
Cheltenham, U.K.: Edward Elgar.
Friedman, T.L. (2013, January 26). Revolution hits the universities. The New York
Times. Retrieved from http://www.nytimes.com/2013/01/27/opinion/
sunday/friedman-revolution-hits-the-universities.html?_r=2&
Giddens, A. (1976). New rules of sociological method: A positive critique of
interpretive sociologies. London: Hutchinson.
Goodman, J. K., Cryder, C. E., & Cheema, A. (2013). Data collection in a flat world:
The strengths and weaknesses of Mechanical Turk samples. Journal of
Behavioral Decision Making, 26, 213–224.
Gupta, N., Martin, D., Hanrahan, B. V ., & O'Neill, J. (2014). Turk-life in India. In
Proceedings of the 18th International Conference on Supporting Group
Work. Association for Computing Machinery (ACM). Retrieved from
http://www.acm.org/conferences/group/conferences/group14/p1.pdf
Hall, S. (1993). Encoding, decoding. In S. During (Ed.), The cultural studies reader
(pp. 90-103). New York: Routledge.
!159
hitRECord (n.d.). Frequently Asked Questions (FAQ). Retrieved from http://
www.hitrecord.org/faq
Howe, J. (2006a). The rise of crowdsourcing. Wired, 14(6), Retrieved from http://
www.wired.com/wired/archive/14.06/crowds.html
Howe, J. (2006b). Crowdsourcing: A definition. In Crowdsourcing: Tracking the rise
of the amateur (weblog). Retrieved from http://
crowdsourcing.typepad.com/cs/2006/06/crowdsourcing_a.html
Howe, J. (2008). Crowdsourcing: Why the Power of the Crowd Is Driving the Future
of Business. New York, NY: Crown Business.
Ipeirotis, P. (2010). Demographics of Mechanical Turk. NYU Center for Digital
Economy Research Working Paper CeDER-10-01. Retrieved from http://
www.ipeirotis.com/wp-content/uploads/2012/02/CeDER-10-01.pdf
Janissary Collective, The (2012). Participatory culture and media life: Approaching
freedom. In A. Delwiche & J.J. Henderson, The participatory cultures
handbook (pp. 257-265). New York: Routledge.
Jenkins, H. (2006a). Convergence culture: Where old and new media collide. New
York: NYU Press.
Jenkins, H. (2006b). Fans, bloggers, and gamers: Exploring participatory culture.
New York: NYU Press.
Jenkins, H. (2013). Rethinking ‘rethinking convergence/culture’. Cultural Studies,
(ahead-of-print), 1-31.
!160
Jenkins, H., & Bertozzi, V . (2008). Artistic expression in the age of participatory
culture: How and why young people create. In S. J. Tepper & Bill Ivey
(eds.), Engaging art: The next great transformation in America’ s cultural
life (pp.171-195). New York: Routledge.
Jenkins, H., & Carpentier, N. (2013). Theorizing participatory intensities: A
conversation about participation and politics. Convergence: The
International Journal of Research into New Media Technologies (ahead-
of-print), 1-22.
Jenkins, H., Clinton, K., Purushatma, R., Robison, A., & Weigel, M. (2006).
Confronting the challenges of a participatory culture: Media education
for the 21st century. White paper. Retrieved from https://
mitpress.mit.edu/sites/default/files/titles/free_download/
9780262513623_Confronting_the_Challenges.pdf
Jenkins, H., Ford, S., and J. Green (2013). Spreadable Media: Creating Value and
Meaning in a Networked Culture. New York: New York University Press,
2013.
Jenkins, H., Purushotma, R., Clinton, K., Weigel, M., & Robinson, A.J. (2006).
Confronting the challenges of participatory culture: Media education for
the 21st century. Chicago: The John D. and Catherine T. MacArthur
Foundation.
Kelty, C. (2012). From participation to power. In A. Delwiche & J.J. Henderson, The
participatory cultures handbook (pp. 3-9). New York: Routledge
!161
Kester, G. (2004). Conversation pieces: Community and communication in modern
art. Berkeley, CA: University of California Press.
Kester, G. H. (2011). The one and the many: contemporary collaborative art in a
global context. Durham, NC: Duke University Press Books.
Koblin, A. (2006). The Sheep Market: Two cents’ worth (Masters Thesis, Design and
Media Arts Program, UCLA) Retrieved from www.aaronkoblin.com/
work/thesheepmarket/TheSheepMarket.doc
Koblin, A. (2010). TEDx Talk [untitled]. TEDx Amazonia, Brazil. Retrieved from
http://en.tedxamazonia.com.br/tedtalk/aaron-koblin
Kolowich, S. (2013, April 29). Why some colleges are saying no to MOOC deals, at
least for now. The Chronicle of Higher Education. Retrieved from http://
chronicle.com/article/Why-Some-Colleges-Are-Saying/138863/
Levy, P. (1997). Collective intelligence: Mankind’s emerging world in cyberspace.
Cambridge, MA: Perseus Books.
Literat, I. (2012). The work of art in the age of mediated participation: Crowdsourced
art and collective creativity. International Journal of Communication
6, 2962-2984.
Literat, I. (2015, forthcoming). Implications of massive open online courses
(MOOCs) on higher education: Mitigating or reifying educational
inequities? Higher Education Research & Development.
!162
Losh, E. (2014). The war on learning: Gaining ground in the digital university.
Cambridge, MA: MIT Press.
MacDonald, D. (2011 [1952]). Masscult and midcult: Essays against the American
grain. New York, NY: New York Review of Books.
Malone, T.W., Laubacher, R., & Dellarocas, C. (2010). The collective intelligence
genome. Sloan Management Review 51 (3), 21-31.
Marvit, M.Z. (2014, February 4). How crowd workers became the ghosts in the
digital machine. The Nation. Retrieved from http://www.thenation.com/
article/178241/how-crowdworkers-became-ghosts-digital-machine#
Mesch, C. (2007). Institutionalizing social sculpture: Beuys’ Office for direct
democracy through referendum installation, 1972. In C. Mesch & V .
Michely (Eds). Joseph Beuys: The Reader (pp. 198-227). Cambridge
MA: MIT Press.
Microsoft Research (n.d.), Crowded: Digital piecework and the politics of platform
responsibility in precarious times. Retrieved from http://
research.microsoft.com/en-us/projects/crowdwork/
National Geographic (n.d.) Life in a Day: About the production. Nat Geo Movies.
Retrieved from http://movies.nationalgeographic.com/movies/life-in-a-
day/about-the-production/
Onah, D. F.O., Sinclair, J., & Boyatt, R. (2014). Dropout rates of massive open
online courses : Behavioural patterns. In Proceedings of the 6
th

International Conference on Education and New Learning Technologies
!163
(EDULEARN14), Retrieved from http://www2.warwick.ac.uk/fac/sci/
dcs/people/research/csrmaj/daniel_onah_edulearn14.pdf
Pateman, C. (1970). Participation and democratic theory. Cambridge, UK:
Cambridge University Press.
PBS Frontline (2010). Roundtable: The crowd. PBS Digital Nation. Retrieved from
http://www.pbs.org/wgbh/pages/frontline/digitalnation/forum/the-crowd/
Peer 2 Peer University (n.d.) About P2PU. Retrieved from https://p2pu.org/en/pages/
about/
Peer 2 Peer University (n.d.) What we’re all about. Retrieved from http://
info.p2pu.org/about/
Rancière, J (2006). Problems and transformations in critical art. In C. Bishop (Ed).
Participation (pp. 83-93). London: Whitechapel/MIT Press.
Rivard, R. (2013, March 8) Measuring the dropout rate in MOOCs. Inside Higher Ed.
Retrieved from http://www.insidehighered.com/news/2013/03/08/
researchers-explore-who-taking-moocs-and-why-so-many-drop-out
Rogstadius, J., Kostakos, V ., Kittur, A., Smus, B., Laredo, J., & Vukovic, M. (2011).
In Proceedings of the Fifth International AAAI Conference on Weblogs
and Social Media. Retrieved from https://www.aaai.org/ocs/index.php/
ICWSM/ICWSM11/paper/view/2778
Rosenberg, A. (2014, January 18). Can Joseph Gordon-Levitt reinvent television with
his new show, ‘HITRECORD ON TV’? ThinkProgress. Retrieved from
!164
http://thinkprogress.org/alyssa/2014/01/18/3181191/joseph-gordon-levitt-
hitrecord/
Schaefer, M. T. (2011). Bastard culture! How user participation transforms cultural
production. Amsterdam: Amsterdam University Press.
Schenk, E., & Guittard ,C. (2011). Towards a characterization of crowdsourcing
practices. Journal of Innovation Economics, 1 (7), 93-107.
Schneider, N. (2015, February 16). Intellectual piecework. Chronicle of Higher
Education. Retrieved from http://chronicle.com/article/Intellectual-
Piecework/190039/?cid=wc&utm_source=wc&utm_medium=en
Sontag, S. (1966). Against interpretation, and other essays. London: Macmillan.
Strickland, C. (2011, January 14). Crowdsourcing: The art of a crowd. Christian
Science Monitor. Retrieved from http://www.csmonitor.com/The-Culture/
Arts/2011/0114/Crowdsourcing-The-art-of-a-crowd
Surowiecki, J. (2004) The wisdom of crowds: Why the many are smarter than the
few and how collective wisdom shapes business, economies, societies,
and nations. New York: Doubleday.
Tabarrok, A. (2012). Why online education works. CATO Unbound. November issue.
Retrieved from http://www.cato-unbound.org/2012/11/12/alex-tabarrok/
why-online-education-works Terranova, T. (2000). Free labor: Producing
culture for the digital economy. Social Text, 18(2), 33-58.
!165
Vaidhyanathan, S. (2012). A new era of unfounded hyperbole. CATO Unbound.
November issue. Retrieved from http://www.cato-unbound.org/
2012/11/16/siva-vaidhyanathan/new-era-unfounded-hyperbole
Wiegand, D. (2014, January 14). 'HitRecord on TV' review: A work in progress - and
it rocks. SFGate. Retrieved from http://www.sfgate.com/tv/article/
HitRecord-on-TV-review-A-work-in-progress-5143171.php
Williamson, V . (2014). On the ethics of crowdsourced research. Manuscript under
review. Retrieved from http://scholar.harvard.edu/files/williamson/files/
mturk_ps_081014.pdf
Wolff, J. (1981). The social production of art. New York: St. Martin's.
Zheng, H., Li, D., & Hou, W. (2011). Task design, motivation, and participation in
crowdsourcing contests. International Journal of Electronic Commerce
15 (4), 57– 88.
!166
APPENDIX A: MECHANICAL TURK TASK DESCRIPTIONS
1. Standard instructions for writing tasks
Help me write a collaborative children’s book about Hashtag the Snail and his Internet
adventures!
Here's the beginning of the story:
“On his way home, Hashtag the Snail stumbled upon another snail's shell. He looked
around, wondering who this mysterious shell belonged to, but there was no other snail in
sight.”
•
What happens next? How does Hashtag use the Internet to find the shell's true owner
and return it to them? What kind of surprises and/or obstacles does he encounter on
the way? Write the next sentence in the story.
•
The only requirement is that Hashtag must somehow use the Internet (or an online
service or community) to accomplish his goal.
2. Standard instructions for illustration tasks
Help me write a collaborative children’s book about Hashtag the Snail and his Internet
adventures!
Please draw and upload an image that represents the following situation:
"On his way home, Hashtag the Snail stumbled upon another snail's shell. He looked
around, wondering who this mysterious shell belonged to, but there was no other snail in
sight."
•
If you don't have Paint or Photoshop on your computer, Pixlr is a free online
alternative: http://pixlr.com/editor/
•
You can also draw it by hand and scan the drawing or take a picture of it.
•
I'm also accepting collages, photos, ASCII art, or any other creative way you can
think of representing this!
•
Thanks! And have fun!
!167
Table A1: Task titles and descriptions for each incentive condition
Condition
Type of
task
Task title Task description
control
writing
Write the next sentence
in this children's story!
Help me write a children's book
about the internet, starring Hashtag
the Snail.
illustration
Draw an image to
illustrate a children's
story!
Help me illustrate a children's book
about the internet, starring Hashtag
the Snail.
$1 bonus
writing
Write the next sentence
in this children's story!
$1 BONUS for best
submission!
Help me write a children's book
about the internet, starring Hashtag
the Snail. If your sentence is chosen
as the best, you get an extra $1
bonus!
illustration
Draw an image to
illustrate a children's
story! $1 BONUS for
best submission!
Help me illustrate a children's book
about the internet, starring Hashtag
the Snail. If your illustration is
chosen as the best, you get an extra
$1 bonus!
$5 bonus
writing
Write the next sentence
in this children's story!
$5 BONUS for best
submission!
Help me write a children's book
about the internet, starring Hashtag
the Snail. If your sentence is chosen
as the best, you get an extra $5
bonus!
illustration
Draw an image to
illustrate a children's
story! $5 BONUS for
best submission!
Help me illustrate a children's book
about the internet, starring Hashtag
the Snail. If your illustration is
chosen as the best, you get an extra
$5 bonus!
charity
writing
Contribute to a charity
project by writing the
next sentence in this
children's story!
Help me fundraise for schools in
India! Write the next sentence in
this children's book about the
internet, starring Hashtag the Snail.
illustration
Contribute to a charity
project by drawing an
image to illustrate a
children's story!
Help me fundraise for schools in
India! Illustrate a children's book
about the internet, starring Hashtag
the Snail.
!168
commercial
writing
Help write a children's
story! Book will be
sold commercially!
Write the next sentence in a
children's book about the internet,
starring Hashtag the Snail.
illustration
Draw an image to
illustrate a children's
story! Book will be
sold commercially!
Help me illustrate a children's book
about the internet, starring Hashtag
the Snail.
credit
writing
Co-author a children's
story! Authorship will
be credited!
Write the next sentence in a
children's book about the internet,
starring Hashtag the Snail. If you
include your name (or a nickname
or pseudonym), I will mention you
as an author in the final book.
illustration
Draw an image to co-
author a children's
story! Authorship will
be credited!
Help me illustrate a children's book
about the internet, starring Hashtag
the Snail. If you include your name
(or a nickname or pseudonym), I
will mention you as an author in the
final book.
no credit
writing
Contribute to a
children's story!
(Authorship will NOT
be credited)
Write the next sentence in a
children's book about the internet,
starring Hashtag the Snail.
illustration
Draw an image to
illustrate a children's
story! (Authorship will
NOT be credited)
Help me illustrate a children's book
about the internet, starring Hashtag
the Snail.
!169
APPENDIX B: CODING SCHEME FOR MEASURING QUALITY
Both written and visual submissions were assessed for usefulness (0-3 points) and novelty
(0-3 points), as per the rubric below.
** NOTE: Rebellious contributions (i.e. submissions that have nothing to do with the snail
or this story) get 0 for usefulness, 0 for novelty, and must be marked as rebellious in the
appropriate column!
A. WRITING TASKS
• usefulness
+1 point (for a max. of 3 pts), if the contribution:
♣ 1. fits narratively, by building on the first sentence
• it must continue the story without finishing it. If it
completely finishes the story (thus rendering any future
contributions unfeasible), it should not get a point.
♣ 2. mentions Hashtag using the Internet or an Internet technology/
service/community
• including mobile media (smartphones, tablets)
♣ 3. is grammatically correct
• beyond ONE minor error, such as a spelling or grammar
mistake
• punctuation errors don't count
• novelty
o +1 point (for a max. of 3 pts), if the contribution:
♣ 1. exceeds the requirements of a basic sentence:
• i.e. it goes beyond basic narration (“who did what”) and
contains deeper descriptions of actions, characters, places,
etc (beyond a mere adjective or adverb)
• this includes the decision to name places, characters or
things (e.g. Old Slug Road; Herman the Hermit Crab; The
Happy Forest Daily Newspaper)
♣ 2. contains an innovative and less common solution:
!170
• i.e. if it is different from the following scenario:
o Hashtag wants to return the shell to its rightful
owner so he posts about it on social media (any
social media or online community) OR posts it on
Craigslist OR looks up information on Google or
Wikipedia in order to find the rightful owner
• examples of ideas that would get a point:
o he uses an unexpected online resource (NOT the
sites/networks mentioned above)
o he does not want to return the shell to its rightful
owner and decides to keep it or sell it
o he goes inside the shell and finds something
o he encounters another character who will help/
hinder his search
• if Hashtag does nothing (e.g. he is scared; he goes home;
he just ponders how strange it is and continues his trek),
do not award a point.
♣ 3. has a whimsical or humorous quality
• including:
o puns (e.g. shellular phone, Slimebook [instead of
Facebook], Snailslist [instead of Craigslist],
Underscore the Worm, “his computer booted up at
a snails pace”)
o actual hashtags (#) incorporated into the text
o rhyming
B. ILLUSTRATION TASKS
o usefulness:
♣ +1 point (for a max. of 3 pts), if the contribution:
• feels like an accurate representation of the sentence
• depicts the snail
• depicts the empty shell
!171
o novelty
♣ +1 point (for a max. of 3 pts), if the contribution:
• exceeds the requirements of a basic illustration and
denotes a higher level of complexity
o i.e. presence of “extra” details that are not
compulsorily needed for the effective illustration
of the sentence (e.g. drawn foreground or
background, flowers, trees, clouds, birds, houses,
intricate detailing on the snail or shell)
• uses colors and/or patterns
• has a whimsical or humorous quality
o snail’s depiction, facial expression or posture is
humorous (e.g. eyes atop antennae, expressive
face, hands on hips, etc)
o funny speech/thought bubbles
o other humorous elements (e.g. smoking chimney
on the snail’s house/shell, snail carrying a laptop
briefcase, etc)
o ** 0 pts if it is NOT an original drawing, created specifically for this task
(i.e. it is taken from the Internet). Please use Google reverse image search
to verify authenticity, for each snail image that is suspicious (doesn't
include the empty shell or specific story details, looks too professional or
“clean”…)
!172
APPENDIX C: CODING SCHEME FOR
CATEGORIZING PARTICIPANT MOTIV ATIONS
1. General categories (to be used for ALL conditions)
• For fun
-
includes words like: fun, interesting, cute, funny, and most other positive
attributes except different/new/unusual (code as “task seemed unique / unusual) and
creative (code as “I enjoy being creative”)
-
EXAMPLES:
-
“Because I wanted to have fun.”
-
“It sounded interesting.”
-
“I thought that the question was interesting and amusing.”
• For the money
-
includes assessments of time-reward ratios and statements as to how quick and
easy it was to complete the task
-
if it mentions a monetary bonus, code as “for the bonus”
-
EXAMPLES:
-
“Im broke and finding a job in this economy sucks for a philosophy major.”
-
“It looked like the right pay for a minute's work.”
-
“It’s fast.”
• I enjoy being creative
-
includes more specific mentions of enjoying writing and/or illustrations
-
EXAMPLES:
-
“I like to write. I am a journalist and I enjoy writing stories and such.”
-
“More creative tasks are always nice.”
-
“I love to be creative. If you give me an idea, I can come up with
something.”
• The task seemed unique / unusual
-
includes all instances where there is a comparison with other Mechanical Turk
or crowdsourcing tasks
-
EXAMPLES:
-
“It's different in a land of surveys.”
-
“Most hits are surveys, this one was a nice change of pace.”
-
“This was very different that any other HIT I've ever seen so I had to do it.”
• I enjoy children’s books
-
includes: enjoying reading and/or writing children's books, mentions of the
benefits of children’s books
!173
-
EXAMPLES:
-
“Because children's books are fun and make me feel happy.”
-
“I love reading to my daughter!”
-
“What drew me in is I thought how my 3 year old would love this book
about a snail who uses the internet like she does sometimes.”
• I wanted to contribute to something bigger
-
includes: mentions of collaborative creativity, exquisite-corpse-style
storytelling, working with others, etc
-
EXAMPLES:
-
“I wanted to be a part of the story.”
-
“Communal Storytelling = Genius. I cannot wait to discover how Hashtag's
story ends.”
-
“It's fun to be a part of something bigger.”
2. Condition-specific categories
• For the potential bonus (to be used in the $1 bonus, $5 bonus conditions)
-
anytime there is a mention of the bonus
-
EXAMPLES:
-
“Extra 5 bucks would be awesome”
-
“Saw bonus chance”
-
“The bonus was worth a shot”
• For charity (to be used in the charity condition)
-
anytime there is a mention of the charitable purpose of the book
-
EXAMPLES:
-
“I mostly accepted the hit, because I think you're doing a good thing, and
I'm glad to help if possible.”
-
“The proceeds being donated to education is definitely a positive thing.”
-
“I love kids and I think the motivation for this idea is really cool.”
• For authorship credit (to be used in the credit condition)
-
anytime there is a mention of being recognized as a co-author in the final
book
-
EXAMPLES:
-
“wanted to be an author in a book”
-
“I would love to be an author, even if it's only a small role.”
-
“It would be neat to be credited in a children's book.”
!174
NOTES:
*** If more two or more motivations are mentioned within the same response, code them in
all applicable categories.
*** If a specific motivation does not fit in any of the above categories, code it as “Other”
and make a note of what it contained.
!175
APPENDIX D: DISTRIBUTION OF V ALUES
ACCORDING TO REWARD AMOUNTS
1. Quantity

Figure A1. Hours until completion for writing tasks, per reward amount
Figure A12. Hours until completion for illustration tasks, per reward amount
0
275
550
825
1100
CONTROL $1 BONUS $5 BONUS CHARITY COMMERCIAL CREDIT NO CREDIT
5 CENTS
10 CENTS
20 CENTS
0
150
300
450
600
CONTROL $1 BONUS $5 BONUS CHARITY COMMERCIAL CREDIT NO CREDIT
5 CENTS
10 CENTS
20 CENTS
!176
2. Quality
Usefulness
Figure A3. Usefulness scores for writing tasks, per reward amount

Figure A4. Usefulness scores for illustration tasks, per reward amount
0
0.425
0.85
1.275
1.7
CONTROL $1 BONUS $5 BONUS CHARITY COMMERCIAL CREDIT NO CREDIT
5 CENTS
10 CENTS
20 CENTS
2.3
2.425
2.55
2.675
2.8
CONTROL $1 BONUS $5 BONUS CHARITY COMMERCIAL CREDIT NO CREDIT
5 CENTS
10 CENTS
20 CENTS
!177
Novelty
Figure A5. Novelty scores for writing tasks, per reward amount

Figure A6. Novelty scores for illustration tasks, per reward amount
0
0.35
0.7
1.05
1.4
CONTROL $1 BONUS $5 BONUS CHARITY COMMERCIAL CREDIT NO CREDIT
5 CENTS
10 CENTS
20 CENTS
0
0.225
0.45
0.675
0.9
CONTROL $1 BONUS $5 BONUS CHARITY COMMERCIAL CREDIT NO CREDIT
5 CENTS
10 CENTS
20 CENTS
!178
3. Time spent
Figure A7. Time spent on writing tasks (in seconds), per reward amount

Figure A8. Time spent on illustration tasks (in seconds), per reward amount
0
250
500
750
1,000
CONTROL $1 BONUS $5 BONUS CHARITY COMMERCIAL CREDIT NO CREDIT
5 CENTS
10 CENTS
20 CENTS
0
75
150
225
300
CONTROL $1 BONUS $5 BONUS CHARITY COMMERCIAL CREDIT NO CREDIT
5 CENTS
10 CENTS
20 CENTS
!179
4. Rebellious contributions
Figure A9. Number of rebellious contributions for writing tasks, per reward amount
Figure A10. Number of rebellious contributions for illustration tasks, per reward amount
0
6
12
18
CONTROL $1 BONUS $5 BONUS CHARITY COMMERCIAL CREDIT NO CREDIT
5 CENTS
10 CENTS
20 CENTS
0
1
2
3
4
CONTROL $1 BONUS $5 BONUS CHARITY COMMERCIAL CREDIT NO CREDIT
5 CENTS
10 CENTS
20 CENTS
!180
APPENDIX E: PARTICIPANT MOTIV ATIONS FOR EACH
REWARD AMOUNT AND INCENTIVE CONDITION

[cont’d on next page…]
!181

!182
APPENDIX F: EXAMPLES OF REBELLIOUS CONTRIBUTIONS
SUBMITTED FOR ILLUSTRATION TASKS
1) Animals
!183
2) Babies
!184
3) Miscellaneous
!185
APPENDIX G: EXAMPLES OF ILLUSTRATION SUBMISSIONS
! !186
! !187
APPENDIX H: hitRECord TASK INSTRUCTIONS
TITLE: Add a sentence or an illustration to a children's book about Hashtag the Snail!
DESCRIPTION:
Help me write and illustrate a collaborative children’s book about Hashtag the Snail and his
Internet adventures! A children's book about the Internet, by the Internet!
I love children's books, and I love the Internet, but all children's books about Internet culture
are so negative and restrictive: don’t talk to strangers on Facebook, don’t give out personal
information, don’t download pirated material, don’t… don’t… don’t. Why not tell kids
about all the amazing things that the Internet has to offer, and how you can find community
and meet people from all over the world?…
HitRECord is an awesome example of this positive potential. I appreciate this community so
much, and I'd love for you all to participate in the production of this book.
The finalized book will be available in two forms: for free online as an e-book (where
anybody can download it), AND as a printed book for those that want to purchase it, with all
profits going to an amazing NGO that implements digital storytelling programs in India.
WRITERS: This is going to work like a Round-robin story... Please copy/paste the story so
far and add your contribution (a sentence or a series of sentences - add as much as you'd
like!), marking it in bold. Next person, copy/paste everything that has been written so far,
and add your contribution in bold, and so on.
ILLUSTRATORS: Choose one sentence to illustrate and upload it as an image record.
Please make sure to specify which sentence you're illustrating. Each sentence will have one
illustration, so that the final book - like most children's books - will have one sentence and
one illustration per page. Thanks, and have fun! :-)
! !188
APPENDIX I: P2PU COURSE DESCRIPTION
(“ABOUT” PAGE)
Hi there, budding Internet artist! Thanks for joining this course, and welcome to the
community!
Course Overview
This initiative was born out of my belief that creativity should not be - and was never meant
to be - the prerogative of a few chosen individuals. Nor should creativity be limited to the
institutional space of museums, galleries, and art fairs. The Internet provides a unique
platform to engage a potentially global community in the creation of collaborative art, and I
think there’s something pretty special about that. This course is an introduction to
crowdsourced art, and an experiment in collective creativity. We will learn about
crowdsourced art, discuss key examples, and - to see how creative collaboration works in
practice - we will work together to develop a children's book about a snail called
Hashtag and his adventures on the Internet. Participation in this project will not require
too much time on your part (and you can be as involved as you'd like), but hopefully it'll be
fun to work on and it will allow us to build a strong and supportive creative community!
Structure
The course will be organized as follows:
• Module 1: intro to crowdsourced art and the children's book project
• Module 2: collaborative plot development and writing (20-30 sentences)
• Module 3: collaborative illustration (20-30 illustrations)
• Module 4: wrap-up and final details (title, cover, possibilities for online publishing, etc).
If you haven't officially signed up yet, please take a moment to do so: being registered for
the course ensures that we can all communicate and collaborate effectively.
Course Facilitator
Hi, I’m Ioana Literat. I’m a PhD Candidate at the USC Annenberg School for
Communication, where I research - guess what - crowdsourced art! You can find out more
about my academic work on my website. And when I’m not writing journal articles about
crowdsourced art, Romanian revolutions, or R. Kelly’s Trapped in the Closet, you’ll
probably find me in my garden, at the dog park, or maybe just on my couch watching
Trapped in the Closet one more time. I look forward to connecting over email
(ioanaliterat@gmail.com), Twitter (@ioanaliterat), or Facebook..
! !189
Our Community
You know the rules. They’re the same things your mom has been telling you since the first
day of kindergarten. Be nice. Assume good intentions. Criticize without being mean. Don’t
leave before dinner is finished (which in our case translates to: please don’t leave the course
before it’s over, because we’re counting on you for the collaborative children's book project!
And we would miss you, really.)
Our course is designed to reflect the P2PU values: openness, community and peer learning. I
am here to guide you along and to provide you with support and feedback if you need it, but
the most valuable insights and experiences will come from your interactions, your
collaboration, and your collective development as a creative community. Read more about
the P2PU pedagogical model here.

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

PDF

Participatory public culture and youth citizenship in the digital age: the Medellín model

PDF

Building a unicorn: management of innovation, collaboration, and change in a Silicon Valley start-up

PDF

The impact of engaged communication scholarship on the re-imagination of space and place in South L.A.

PDF

Alternative citizenship models: from online participatory cultures to participatory politics

PDF

Ecology and network evolution in online innovation contest crowdsourcing

PDF

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

PDF

Modes of cultural resistance post-Arab Spring: humor, music and the making of a new digital identity

PDF

From victim to subject: dramatization as research in Thailand's anti-trafficking movement

PDF

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

PDF

Corporate reputation crisis in the digital age: a comparative study on Abercrombie & Fitch’s reputation crisis in the U.S., China and Taiwan

PDF

Temporal dynamics in the lives of health practitioners

PDF

The changing nature of museology in the digital age: Case studies of situated technology praxis in U.S. art museums

PDF

Civic expression in Little Tokyo: how art and culture empowers communities and transforms public participation

PDF

Social motivation and credibility in crowdfunding

PDF

A seat at the table: navigating identity, power, and safety in work meetings

PDF

Infrastructures of the imagination: building new worlds in media, art, & design

PDF

Digital storytelling as participatory media practice for empowerment: the case of the Chinese immigrants in the San Gabriel Valley

PDF

The effects of poor practice in public relations: the Pennsylvania State football scandal

PDF

Immigrant health in context: a communication ecology approach to understanding health behavior

Asset Metadata

Creator Literat, Ioana (author)

Core Title Crowdsourced art: activating creative participation in online spaces

School Annenberg School for Communication

Degree Doctor of Philosophy

Degree Program Communication

Publication Date 07/16/2015

Defense Date 05/13/2015

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

Tag art,collaboration,creativity,crowdsourcing,Internet,OAI-PMH Harvest,participatory culture

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Jenkins, Henry (committee chair), Brabham, Daren C. (committee member), Gross, Larry P. (committee member), Hollingshead, Andrea B. (committee member)

Creator Email iliterat@usc.edu,ioanaliterat@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-594750

Unique identifier UC11301005

Identifier etd-LiteratIoa-3607.pdf (filename),usctheses-c3-594750 (legacy record id)

Legacy Identifier etd-LiteratIoa-3607.pdf

Dmrecord 594750

Document Type Dissertation

Format application/pdf (imt)

Rights Literat, Ioana

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