An Exploratory Factor Analysis of Motivations for Participating in

2013 46th Hawaii International Conference on System Sciences
An Exploratory Factor Analysis of Motivations for Participating in
Zooniverse, a Collection of Virtual Citizen Science Projects
Jason Reed
Adler Planetarium
1300 South Lake Shore Drive
Chicago, IL 60605
[email protected]
M. Jordan Raddick
Department of Physics and Astronomy
Johns Hopkins University
Bloomberg 388, Baltimore, MD 21218
[email protected]
Andrea Lardner
Department of Physics and Astronomy
Johns Hopkins University
Baltimore, MD 21218
[email protected]
Karen Carney
Adler Planetarium
1300 South Lake Shore Drive
Chicago, IL 60605
[email protected]
As such, a long-standing concern for those who
create citizen science projects is how to attract and
maintain volunteers’ engagement with these projects.
One promising solution is virtual citizen science
(VCS), which uses forms of computer-mediated
communication (CMC) like email or websites for all
volunteer activities [24]. Use of CMC lowers barriers
to project participation; as long as volunteers can
access the CMC—usually a website—they can
contribute no matter where they are in the world. In
this way, VCS provides a wide and accessible
pathway for volunteer participation.
Despite a brief lifespan of about a decade, VCS
projects have already achieved some impressive
examples of success. Since its creation in 2007, over
600,000 volunteers have created accounts and
contributed to the Zooniverse, an ever-growing
collection of VCS projects [8]. The volunteers
directly contributed to the publication of dozens of
peer-reviewed papers in a variety of disciplines such
as astronomy, archeology, and climatology. The
Cornell Ornithology Lab created VCS projects like
eBird to enable birdwatchers to create and access one
of the largest depositories of biodiversity data in the
world [23]. Foldit, a VCS project that turned protein
folding into a game, discovered an ideal protein
structure before professional AIDS researchers
working on the same problem [12].
Examples like these encouraged research into
what motivates volunteers to participate in VCS.
This paper presents the results of an online survey
regarding volunteers’ self-reported motives for
participating in the collection of VCS projects of the
Zooniverse. We begin with a review of relevant work
that suggests that there are a variety of potential
Abstract
Virtual citizen science (VCS) uses technology like
the Internet to give volunteers the chance to
participate in real scientific research. The success of
VCS projects has prompted researchers to
understand what motivates volunteer participation.
199 registered users of Zooniverse, a successful
collection of VCS projects, completed a web survey
assessing various possible motives for their
participation. An exploratory factor analysis of their
responses suggests a three-factor solution
representing motives related to Social Engagement,
Interaction with the Website, and Helping. We
discuss the place these results have in helping to
understand the still emerging picture of the variety of
motives associated with various VCS projects and
activities.
1. Introduction
Citizen science gives volunteers who are not
trained experts in a field of study the chance to
participate in authentic research projects [4,25]. The
authenticity comes from volunteers taking part in
different aspects of the process of scientific inquiry
(e.g., data gathering, data cleaning, data analysis and
interpretation, and research question and hypothesis
formation) as part of these projects [10]. Importantly,
the volunteers are vital to the success of these
projects. It is their time and effort that allow these
projects to make meaningful contributions to the
body of scientific knowledge.
1530-1605/12 $26.00 © 2012 IEEE
DOI 10.1109/HICSS.2013.85
609
610
motives for participating in VCS. We then detail the
creation of a set of survey items designed to assess a
variety of motives for participation in VCS. We then
present the results of an exploratory factor analysis
(EFA) to identify any underlying dimensions to the
survey takers’ responses. We conclude by discussing
the results and detailing suggestions for future work
on this important topic.
a function of volunteers’ time and actions with VCS.
For example, awareness of a VCS project should be
more relevant to volunteers with relatively little
experience with a VCS project, whereas feelings of
obligation to a VCS project are more relevant to
volunteers with more experience with a project.
Although the propositions in [6] have yet to be
empirically evaluated, they are consistent with our
assumption that motivations to volunteer are sensitive
to the context of participating in VCS projects.
The variety of motives for participation in VCS is
also apparent in empirical research about volunteers
in distributed computing projects. These projects
involve researchers dividing large, complex
computational problems into smaller bits and asking
volunteers to install software onto their own personal
computers to process these smaller bits. Although
volunteers usually don’t actively contribute to them,
distributed computing projects share the core
qualities of VCS; both involve volunteers making
active choices to contribute their resources to aid in
the progress of scientific investigations. For example,
a well-known VCS project named seti@home makes
use of the idle processing cycles of volunteers’ home
computers to see of radio signals from space that may
contain unusual signals that could indicate intelligent
life. A survey of participants from various distributed
computing projects found a wide variety of reported
reasons for participation, such as contributing to
scientific research, competition, effective use of
computer resources, acquisition of technical
knowledge, and interaction with other volunteers
[11].
Whereas some investigations like [11] examined
motivations for a category of projects similar to VCS,
other investigations focused on one or two actual
VCS projects. For example, intrinsic, norm-oriented,
identification, and collective motives were salient
reasons for participation in a passive (seti@home)
and an active (Stardust@home) astronomy VCS
project [16]. A survey of both the scientists who
created and volunteers participating in a data curation
VCS project suggested that egoism, collectivism,
altruism, and principalism were salient motives for
both parties [22].
One particularly interesting bit of research
involves two studies that examine the same VCS
project. [18] interviewed 20 registered users of
Galaxy Zoo, a well-known VCS project in which
volunteers classified the morphological features of
galaxies. Coding of the interviews yielded a list of
twelve different motives for participation including
contributing to scientific research, learning about
galaxies, making discoveries, interacting with other
people, teaching other people, looking at pleasing
1.1. Motivations for participation in VCS
To inform the development of items for our
survey of motivation to participate in VCS, we relied
primarily on research about motivation to participate
in VCS projects. This does not mean that we ignored
research about motivation to volunteer outside of this
particular context. To be sure, work such as [3] laid
important groundwork to understand that a variety of
personal and social motives drive volunteering in
general. Furthermore, some of these motives are
relevant to volunteering that occurs primarily in
online situations. For example, volunteers who create
open source software are driven mainly by personal
motivations such as creative fulfillment and skill
development [13]. Personal motives like having fun
and sharing ideological beliefs are important drivers
of volunteering in crowdsourcing applications like
Wikipedia [15]. Such work underscores the
importance of understanding motivation to volunteer
in online contexts.
However, motivation for volunteering in VCS
may reflect differences between VCS and other
online contexts. Indeed, volunteering in VCS occurs
in its own particular context of the research protocols
and procedures created by scientists conducting
actual research [10]. It does not require volunteers to
self-organize
like
crowdsourcing
nor
to
autonomously determine all aspects of their activities
as in open-source software development [24]. If
motivations for volunteering “depend on the
interaction of person-based dynamics and situational
opportunities” [2], it follows that we should pay
particular attention to research about motivation to
volunteer in VCS so as to best capture the nature of
motivation to volunteer in this particular context.
A theoretical model by [6] applies previous work
about motivation to volunteer to online contexts like
VCS marked by people collectively contributing
effort online. Volunteering in VCS is motivated by
some combination of volunteers’ awareness of a VCS
project and its needs, their particular interaction with
it, their perceived capacity to do what the VCS
project asks of them, their obligation to volunteer for
that VCS project, and their evaluation of their VCS
experience. The particular combination of motives is
611
610
images, fun, helping, amazement about the vastness
of the universe, interest in the Galaxy Zoo project,
interest in the field of astronomy, and interest in
science in general.
[14] analyzed the content of forum posts of
Galaxy Zoo and found a variety of motives for
participation complementary to those in [18]: social
interaction as part of the scientist-volunteer
collaboration process, interest in astronomy-related
topics like the space race, spiritual aspects related to
consideration of the topic material, and a strong
pleasing aesthetic appeal of the galaxy images. To
our knowledge, this is the only example of research
about what motivates participation in VCS in which
the same VCS project is examined by more than one
study with different methods of investigation.
Clearly, one of the primary challenges for this
burgeoning area of research is how to make sense of
the variety of methods and conclusions found in the
existing research. On the one hand, researchers can
and should continue to draw upon established
literature and methods to assess which are
appropriate to assess the motivations most relevant to
a particular VCS project’s goals and needs. Some
motives will be specific to a particular VCS project
because of that project’s unique design and features
necessary to successfully meet its research goals [17,
21]. However, it would also be beneficial for
subsequent research to explore if there is any general
underlying structure to the reported motivations for
participating in VCS. As the number and variety of
VCS projects increases, so does the desire for a sense
of whether there are general motives that apply to
more than one project.
Materials. Participants’ responses to the items were
then analyzed using an EFA. Because this is first
research using these items, EFA is desirable because
it will provide information about which of the items
from our item pool best represent any underlying
structure of the data.
2. Method
2.1. Participants
Following the rationale of [6] that there are
different types of volunteers in VCS projects, we
used a stratified sampling strategy to create a sample
of three distinct types of registered Zooniverse user.
Zooniverse volunteers were distinguished by either
the amount of time they spent contributing to the
main scientific research tasks of the project or by
communicating with other volunteers or scientists
about their actions via project features like the project
forums. Our analysis of Zooniverse activity logs
suggested that 38% of volunteers spent less than 20
minutes contributing to the main research task of at
least one Zooniverse project, 51% spent at least 20
minutes contributing to the main research task of at
least one Zooniverse project, and 11% were in the top
half of volunteers who started or replied to discussion
threads. All email addresses available as part of
registered user accounts were assigned to one of
these three groups.
We then sent 3,883 emails to randomly selected
user accounts from each of the three groups. All
selected accounts received an initial invitation to take
the survey, followed by a reminder email sent once a
week for the next two weeks. If a selected email
address did not take the survey, it was removed from
subsequent waves of survey sampling.
Response rates from each of the three groups
were monitored to keep the sample proportions
similar to those in the population of registered
Zooniverse accounts. 220 participants responded to
the invitation and completed at least part of the
survey (response rate = 5.6%). Twenty-one
participants of the 220 participants who responded to
the invitation completed less than 20% of the survey
and were removed from the sample used in analyses.
Thus, this research used a final sample of 199
participants (5.1% of the invited sampling frame). Of
these 199, 83 were from the first group (42% of the
sample), 75 were from the second group (38% of the
sample, and 41 were from the third group (20% of the
sample).
Of the 199 participants who completed the
survey, the majority were male (67.3% vs. 30.2%
1.2. The current study
The current study attempts to address both of
these issues by investigating motivation to participate
in Zooniverse. As previously mentioned, Zooniverse
is a group of VCS projects that share a common
brand and investigate a variety of topics that evolved
from and because of the success of the Galaxy Zoo
project [8]. The current study builds upon the
previous work of [14] and [18] by extending the
investigation of motives for participating in a group
of VCS projects that are the progeny of Galaxy Zoo.
In exploring whether and how the motives for
participating in a single project like Galaxy Zoo
relate to the set of projects in Zooniverse, the current
study will also detail motivations unique to the
Zooniverse. To do this, we created a survey of selfreport items to assess the multiple motivations
proposed in [6]. We detail in the creation of an item
pool and the selection of items in section 2.2.
612
611
female vs. 2.5% no response) with a mean age of
40.7 years (SD = 15.69). Although participants
reported residing in countries on 6 different
continents, the majority resided in either the United
States (51.3%) or the United Kingdom (20.6%). Of
the 119 participants who provided information about
the highest level of education they completed, the
majority had achieved some form of college degree
(72.9% of participants using the U.S. education
system designation and 87.5% of participants using
the U.K. education system designation). Of the 197
participants who provided information about their
occupation, 90% reported being employed in a
variety of fields such as “Computer and
Mathematical” (19.8%) and “Education, Training,
and Liberty” (10.2%).
featured the official Zooniverse logo and motto
centered at the top of the page to clearly
communicate to the participants that this was indeed
an official part of the Zooniverse.
The survey website consisted of 10 individual
webpages with some design features common to all
webpages. Each of these webpages feature the phrase
“Zooniverse Volunteer Survey” below the
Zooniverse logo and motto to clearly communicate
the activity of taking a survey related to Zooniverse
volunteers. All survey webpages featured buttons
labeled “Previous section” and “Next section” to
allow participants to move through the survey
backwards and forwards, respectively. Finally, all
survey webpages contained a progress indicator in
the form of a phrase that communicated participants
were on “Page xx of 10.”
The first survey webpage asked participants to
provide their informed consent to participate in the
survey. Informed consent was achieved by providing
a link to an electronic version of a consent form for
participants to read and a box to check if they agreed
with the statement, “I understand the nature of this
research and wish to participate.” If participants
checked the box, they then activated a button in the
lower right corner of the screen allowing them to
proceed.
After providing their informed consent,
participants were directed to the second webpage of
the survey website that repeated the message of the
invitation email and stated that participants could also
elect to receive compensation of US$10 if they
wished to provide a mailing address once they
completed the survey. This allowed participants to
choose whether they wished to receive compensation
and if they wished to provide personal information.
Those participants who did provide a mailing address
had that information stored on a separate secure
database that is only accessible by the researchers
and programmers associated with this project.
Page 3 of the survey website asked participants to
provide self-reports of information about their
activities on the Zooniverse and science in general.
Because these items did not specifically ask people
about their motives for doing or not doing these
behaviors, we do not discuss them in the remainder
of this paper.
Pages 4-9 of the survey website presented the 54
items used to assess self-reported motives for using
the Zooniverse. All participants received a randomly
ordered presentation of the 54 items across these 6
webpages. Each webpage presented no more than 10
items. All items were individually numbered with
respect to their overall place in the survey (e.g. the
first motivation item on page 4 of the survey was
2.2. Materials
We created a pool of 114 items designed to assess
the motives proposed in [6]. The authors reviewed
the literature for measures related to the various
motives in [6] and reworded or created items so that
they applied to activity on the Zooniverse and its
projects. Two groups of judges (one of 10 educators
and another of 20 volunteers naïve of this research)
evaluated the items by sorting each individual item
based on which of the motives proposed in [6] they
thought it best represented. A total of 54 items were
kept because all of the judges in one group or at least
75% of both groups of judges correctly assigned it to
its motive category. All items were written as Likerttype scale items using a 5-point scale. The scale used
the following values and labels: 1 = “Strongly
Disagree,” 2 = “Disagree,” 3 = “Neither Agree Nor
Disagree,” 4 = “Agree,” 5 = “Strongly Agree.” An
additional response option of “Not Applicable” was
also provided on a line below the 5 scale agreement
scale points.
The survey instrument also contained two other
sets of items in addition to those used to assess
motivation for participation in VCS. One set of items
assessed demographic characteristics of the survey
takers, specifically their self-reported gender, age,
country of residence, highest level of education
achieved, and profession. A different set of 11 items
assessed various behaviors associated with the
participants’ activity on the Zooniverse and science;
these items were not analyzed in the current
investigation.
2.3. Procedure
Participants who visited the survey website were
asked to login to their Zooniverse user account. After
successfully logging in, participants were directed to
the survey website. All pages of the survey website
613
612
numbered 12). Each item began with its statement
directly to the right of its number, with the response
scale directly below the statement. Participants were
free to move between pages and answer questions at
will.
Page 10 of the survey website presented
participants with demographics questions asking
them to report their age, gender, country of residence,
highest level of education achieved and profession.
This page also contained a button labeled “Click here
to finish” instead of “Next section” to clearly indicate
that the survey was complete.
After finishing the 10 pages of the survey
website, participants were directed to another
webpage that provided fields for participant names
and mailing addresses and a message at the top center
of the webpage that read, “Thank you for completing
the survey. If you would like compensation please
send us your address using the form below.”
Participants could click a “Submit” button in the
lower left corner of the webpage once finished. If so,
they would receive a message that their responses
had been recorded. Payment of US$10 was sent to all
mailing addresses in the form of a pre-paid gift
checks.
likelihood for factor extraction and the geomin
rotation to allow extracted factors to correlate.
Because factors could correlate, we retained survey
items that had both high factor loadings and
correlations with only a single factor in order to aid
interpretation of the factors. Practically, we only
considered keeping items with factor loadings and
correlations of .40 or greater and cross-loadings of
.35 or less.
We interpreted the results to best be described by
a three-factor structure. Because the eigenvalues for
the first 15 factors were greater than one and its
general unreliability, we did not rely on the Kaiser
rule to determine the number of factors to keep [7].
Figure 1 shows that the scree plot of eigenvalues
begins to flatten out between 3 and 4 factors. The
final decision to keep a 3-factor solution was based
on a theoretical reason. The results of an exploratory
structural equation analysis of a different survey of
Zooniverse users using different items to assess
motivation to participate suggested a three-factor
solution similar in concept to the one here [20]. The
three factors in the current research are moderately
correlated but not redundant with one another:
rFactor1Factor2 = .43; rFactor1Factor3 = .21; rFactor2Factor3 = .25.
As such, we decided to retain 3 factors because of
their consistency with this previous work.
3. Results
The responses of the final sample of 199
participants were subjected to an EFA using LISREL
5.1. Although our survey used items derived form
motives for participating in Zooniverse based on the
theoretical model in [6], this is the first empirical data
gathered about these items. As such, we are
essentially in the beginning stages of developing a
new measure of motivation for participating in VCS,
and EFA is an appropriate analysis technique to
examine any latent motivation constructs underlying
the survey response dataset from the current study
[5].
We also conducted the EFA using the structural
equation modeling analysis program LISREL 5.1
because of its ability to handling missing data. Every
participant who completed the survey did not respond
to at least one of the survey items related to
motivation for participating in Zooniverse (M = 3.96
missing items, SD = 5.68). LISREL 5.1 can easily
deal with missing data to this degree via its use of the
Full Information Likelihood Method (FIML) to
estimate missing portions of a variable for latent
variable models like EFA [1, 9].
We followed recommendations in the literature
on EFA to make the most appropriate choices for our
analysis [5, 7]. Specifically, we used maximum
Figure 1. Scree plot
To interpret the factors, we needed to consider
both the factor loadings that show the linear
combination of items for each factor (see Table 1)
and the structure matrix that shows the correlation
between each item and its factor (see Table 2). For
ease of presentation, only the factor loadings and
correlations that were kept are displayed in Tables 1
and 2; crossloadings less than .35 are not shown in
the tables for ease of presentation. The eigenvalues
and factor determinacies for each factor are located at
the bottom of each table; both of these suggest that
each of the factors is reasonably represented by its
constituent items.
We labeled each of the three retained factors to
reflect the content of their constituent items. We
614
613
Table 1. Factor loadings with geomin rotation
Item label
Awareness3
Awareness4
Capacity7
Capacity8
Evaluation13
Evaluation15
Evaluation17
Evaluation18
Obligation5
Obligation6
Obligation8
Awareness6
Awareness7
Capacity4
Capacity6
Item wording
People like me are generally aware of X Zoo.
I have seen or heard of other people using X Zoo.
I know what other X Zoo users expect of me as a forum
contributor.
I know the values of other X Zoo users.
I get to use a number of complex skills in X Zoo.
My work in X Zoo involves doing a number of different
tasks.
I enjoy being a part of the community of X Zoo
volunteers.
I am good at working with other X Zoo volunteers.
I trust my fellow X Zoo volunteers.
Other X Zoo volunteers can be respected as co-workers.
I fulfill the commitments I make to other volunteers in X
Zoo.
X Zoo's needs are clearly stated.
I know what X Zoo needs from me.
I feel confident that I can use the available technology to
get things done in X Zoo.
I feel confident that I can log on and navigate on the X
Zoo Web site.
When I first saw the X Zoo website, I thought that the X
Zoo Web site was attractive.
Visitor
Website
Interaction1
Visitor
When I first saw the X Zoo website, I thought that I
Website
liked the graphics and images used on the X Zoo Web
Interaction2
site.
Visitor
When I first saw the X Zoo website, I thought that the
Website
screen layout of the X Zoo Web site was attractive.
Interaction3
Visitor
I find X Zoo to be easy to use.
Website
Interaction4
Visitor
I find it easy to get X Zoo to do what I want it to do.
Website
Interaction6
Visitor
My interaction with X Zoo is clear and understandable.
Website
Interaction7
Visitor
It was easy for me to become skillful at using X Zoo.
Website
Interaction8
Evaluation10
Helping science makes me feel good about myself.
Obligation1
Volunteering makes me feel important.
Obligation2
Volunteering makes me feel good about myself.
Obligation4
I want to make the world a better place.
Eigenvalues
Factor Determinacies
615
614
Social
engagement
.521
.429
.472
Interaction
with
website
Helping
.533
.507
.525
.597
.662
.562
.506
.574
.475
.530
.609
.586
.564
.462
.537
.774
.692
.700
.658
10.166
.946
3.549
.950
.625
.614
.769
.488
3.083
.916
Table 2. Structure matrix
Item label
Awareness3
Awareness4
Capacity7
Capacity8
Evaluation13
Evaluation15
Evaluation17
Evaluation18
Obligation5
Obligation6
Obligation8
Awareness6
Awareness7
Capacity4
Capacity6
Item wording
People like me are generally aware of X Zoo.
I have seen or heard of other people using X Zoo.
I know what other X Zoo users expect of me as a forum
contributor.
I know the values of other X Zoo users.
I get to use a number of complex skills in X Zoo.
My work in X Zoo involves doing a number of different
tasks.
I enjoy being a part of the community of X Zoo
volunteers.
I am good at working with other X Zoo volunteers.
I trust my fellow X Zoo volunteers.
Other X Zoo volunteers can be respected as co-workers.
I fulfill the commitments I make to other volunteers in X
Zoo.
X Zoo's needs are clearly stated.
I know what X Zoo needs from me.
I feel confident that I can use the available technology to
get things done in X Zoo.
I feel confident that I can log on and navigate on the X
Zoo Web site.
When I first saw the X Zoo website, I thought that the X
Zoo Web site was attractive.
Visitor
Website
Interaction1
Visitor
When I first saw the X Zoo website, I thought that I
Website
liked the graphics and images used on the X Zoo Web
Interaction2
site.
Visitor
When I first saw the X Zoo website, I thought that the
Website
screen layout of the X Zoo Web site was attractive.
Interaction3
Visitor
I find X Zoo to be easy to use.
Website
Interaction4
Visitor
I find it easy to get X Zoo to do what I want it to do.
Website
Interaction6
Visitor
My interaction with X Zoo is clear and understandable.
Website
Interaction7
Visitor
It was easy for me to become skillful at using X Zoo.
Website
Interaction8
Evaluation10
Helping science makes me feel good about myself.
Obligation1
Volunteering makes me feel important.
Obligation2
Volunteering makes me feel good about myself.
Obligation4
I want to make the world a better place.
Eigenvalues
Factor Determinacies
616
615
Social
engagement
.473
.409
.471
Interaction
with
website
Helping
.555
.568
.557
.642
.665
.563
.539
.564
.509
.507
.576
.554
.533
.471
.495
.706
.670
.698
.637
10.166
.946
3.549
.950
.675
.642
.798
.524
3.083
.916
labeled Factor 1 Social Engagement to reflect that 11
of the items communicate awareness of and
interaction with other members of the Zooniverse
community. This factor also featured two items—
Evaluation 13 and Evaluation 15—that speak mainly
about the tasks and skills involved in Zooniverse
activity. The content of these items is consistent with
the idea of interaction with the Zooniverse
community for a possible subgroup of volunteers
who extensively participate in Zooniverse.
Volunteers who would engage in activities that put
them in contact with the Zooniverse community
would be especially likely to endorse these
statements. We labeled Factor 2 Interaction with
Website to reflect that all 11 items communicate a
sense of awareness, facility, and enjoyment from
using the various features of Zooniverse projects. We
labeled Factor 3 Helping to reflect that all 4 items
communicate how participants experience positive
feelings from helping or volunteering to participate in
Zooniverse.
of contact or if they decline the invitation to
participate, researchers such as ourselves must
graciously accept the volunteers’ choice. This also
precluded us from comparing people who did or did
not participate in the survey with any of the
Zooniverse registered accounts because no personal
information is gathered about Zooniverse users when
they create their accounts.
As such, the current project used a stratified
sampling strategy as one way to attempt to preserve
the qualities of our target population of registered
Zooniverse accounts. Although the proportions of
subgroups in the sample are not a perfect match with
the subgroups in the population of Zooniverse users,
we did maintain comparable proportions of users who
primarily spent time contributing to the main research
task of the project (89% in the population vs. 80% of
the sample) and users who primarily used the
communication tools of the project (11% of the
population vs. 20% of the sample). We have taken
the steps that we could to gather responses indicative
of the Zooniverse volunteer base while respecting the
primary purpose of the Zooniverse to treat its
volunteers as valued collaborators.
Besides the response rate, the absolute value of
the sample size is another possible limitation.
Although there is no set standard, research into the
practices associated with EFA suggests that having a
participant-to-item ratio of 5:1 is a desirable goal [5].
This does not mean that our results are invalid for not
reaching this ratio, but we take this recommendation
to heart in our plan to gather a larger sample to retest
the latent content structure of motivation for
participation in Zooinverse.
Our future plans for this line of research center
around establishing the most plausible latent
construct structure of motivation for participating in
Zooniverse. The immediate steps to reach this goal
include gathering a larger sample to take our survey
so that we can conduct more analyses on different
possible models of underlying latent constructs. We
also seek to develop a set of items that will reliably
measure the latent constructs in this model. Our
current work used an EFA analysis strategy as a
starting point towards accomplishing these goals.
After this phase of work is completed, we plan to
refine our measurement instruments and conduct
more detailed tests of these models. For example, we
plan on examining whether the same model of
motivation constructs applies to the different types of
users proposed by [6]. That is, do the motives that
best describe the actions of initial users apply to
participants who exhibit different types and patterns
of behaviors on Zooniverse? In this way, we hope to
continue to make contributions to this fledgling and
4. Discussion
The current study provides evidence that
participation in the VCS projects of Zooniverse may
be motivated by reasons related to social
engagement, interaction with the website, or helping.
The current research both connects and extends what
is known about motives for participating in VCS.
Like previous work, the current results suggest a
variety of motives for participating in the Zooniverse
collection of VCS projects. However, it brings new
theoretical and empirical details into the mix as well.
Despite these contributions, the current study is
not without limitations. One limitation of the current
research is the response rate final sample analyzed in
the EFA. The 199 participants in the final sample
made up about 5% of the invitations sent to registered
Zooniverse users. Even though this is a very low
response rate, it is similar to a 6% response rate
found in a different survey of registered Zooniverse
users [19].
Both of these low response rates are potentially
due to the Zooniverse’s approach to the process of
VCS. From its beginnings, the Zooniverse strove to
offer its volunteers an authentic experience in
contributing to scientific research and to treat them as
valued collaborators in the process. So as to not
violate the Zooniverse’s mission, research projects
like this one must reach out to the Zooniverse user
base as equals and ask them to do the researchers a
favor by accepting the invitation to participate. If
volunteers cannot be reached by the email address
they provide as the only mutually agreed upon form
617
616
Mediated Communication, vol. 10, no. 4, article 16, 2005,
http://jcmc.indiana.edu/vol10/issue4/holohan.html.
important line of research of what motivates
volunteering in VCS.
[12] F. Khatib, F. DiMaio, Foldit Contenders Group, Foldit
Void Crushers Group, S. Cooper, M. Kazmierczyk, M.
Gilski, S. Krzywda, H. Zabranska, I. Pichova, J.
Thompson, Z. Popović, M. Jaskolski, and D. Baker,
“Crystal structure of a monomeric retroviral protease
solved by protein folding game players,” Nature Structural
& Molecular Biology, published online September 18,
2011, doi:10.1038/nsmb.2119.
Acknowledgement: This material is based upon work
supported by the National Science Foundation under
Grant No. 0917608.
5. References
[1] R. L. Carter, “Solutions for missing data in structural
equation modeling,” Research & Practice in Assessment,
vol. 1, no. 1, March 2006, pp. 1-6.
[13] K. R., Lakhini and R. G. Wolf, “Why hackers do what
they do: Understanding motivation and effort in free/open
source software projects,” MIT Sloan working paper no.
4425-03, 2003.
[2] E. G. Clary and M. Snyder, “The motivations to
volunteer: Theoretical and practical considerations,”
Current Directions in Psychological Science, vol. 8, no. 5,
October 1999, pp. 156-159.
[14] T. A. Mankowski, S. J. Slater, and T. F. Slater, “An
interpretive study of meanings of citizen scientists make
when participating in Galaxy Zoo,” Contemporary Issues in
Education Research, vol. 4, no. 4., April 2011, pp. 25-42.
[3] E. G., Clary, M. Snyder, R. D., Ridge, J. Copeland, A.
A., Stukas, J. Haugen, and P. Miene, “Understanding and
assessing the motivations of volunteers: A functional
approach,” Journal of Personality and Social Psychology,
vol. 74, 1998, pp. 1516-1530.
[15] O. Nov, “What motivates Wikipedians?”
Communications of the ACM, vol. 50, no. 11, November
2007, pp. 60-64.
[4] J. P. Cohn, “Citizen science: Can volunteers do real
research?” BioScience, vol. 58, no. 3, March 2008, p. 192.
[16] O. Nov, O. Arazy, and D. Anderson, “Technologymediated citizen science participation: A model,” in
Proceedings of the AAAI International Conference on
Weblogs and Social Media (ICWSM 2011), Barcelona,
Spain, 2011.
[5] J. M. Conway and A. I. Huffuct, “A review and
evaluation of exploratory factor analysis practices in
organizational research,” Organizational Research
Methods, Vol. 6 No. 2, April 2003, pp. 147-168.
[17] N. Prestopnik and K. Crowston, “Citizen science
system assemblages: Understanding the technologies that
support crowdsourced science,” in Proceedings of the 2012
iConference, 2012, pp. 168-176.
[6] K. Crowston and I. Fagnot, “The motivational arc of
massive online collaboration,” 2012, submitted for
publication.
[18] J. Raddick, A. Szalay, J. Vandenberg, G. Bracey, P.
Gay, C. Lintott, P. Murray, and K. Schawinski, “Galaxy
Zoo: Exploring the motivations of citizen science
volunteers,” Astronomy Education Review, vol.9, no. 1,
2010, doi:10.3847/AER2009036.
[7] L. R. Fabrigar, D. T. Wegener, R. C. MacCallum, and
E. J. Strahan, “Evaluating the use of exploratory factor
analysis in psychological research.” Psychological
Methods, vol. 4, 1999, pp. 272-299.
[8] L. Fortson, K. Masters, R. Nichol, K. Borne, E.
Edmondson, C. Lintott, J. Raddick, K. Schwaniski, & J.
Wallin, “Galaxy Zoo: Morphological classification in
citizen science,” In Way, M.J., Scargle, J.D., Ali, K., and
Srivastava, A.N. (Eds.), “Advances in Machine Learning
and Data Mining for Astronomy,” Boca Raton: CRC Press,
2012.
[19] M. J. Raddick, G. Bracey, P. L Gay, C. J. Lintott, C.
Cardamone, P. Murray, K. Schawinski, A. S. Szalazy, and
J. Vandenberg, “Galaxy Zoo: Motivations of citizen
scientists,” Astronomy Education Review, in press.
[20] J. T. Reed and M. J. Raddick, “Factor structure of
motivations for participating in online citizen science,” talk
presented at the annual meeting of the Midwestern
Psychological Association, Chicago, IL, May 2012.
[9] J. W. Graham, “Missing data analysis: Making it work
in the real world,” Annual Review of Psychology, vol.6,
2009, pp. 549-576.
[21] J. Reed, W. Rodriguez, and A. Rickhoff, “A
framework for defining and describing key design features
of virtual citizen science projects,” in Proceedings of the
2012 iConference, 2012, pp. 623-625.
[10] M. Haklay, “Levels of participation in citizen science and
scientific knowledge production,” Retrieved April 15, 2012,
from https://povesham.wordpress.com/tag/citizen-science.
[22] D. Rotman, J. Preece, J. Hammock, K. Procita, D.
Hansen, C. Parr, D. Lewis, D. Jacobs, and A. Biswas,
“Dynamic changes in motivation in collaborative citizen
[11] A. Holohan, and A. Garg, “Collaboration online: The
example of distributed computing,” Journal of Computer-
618
617
science projects,” in Proceeding of the 2012 ACM
Conference on Computer Supported Cooperative Work,
CSCW’2012. Seattle, WA, 2012.
[23] A. Wiggins, “eBirding: Technology Adoption and
the Transformation of Leisure into Science,” poster
presented at the 2011 iConference, Seattle, WA, February
8–11, 2011.
[24] A. Wiggins and K. Crowston, “From conservation to
crowdsourcing: A typology of citizen science,” in
Proceedings of the Forty-fourth Hawai’i International
Conference on System Sciences (HICSS-44), Koloa, HI,
January, 2011.
[25] J. Silvertown, “A new dawn for citizen science,” Trends in
Ecology & Evolution, vol. 24, 2009, pp. 467–471.
619
618