The contribution of supply and demand factors to the reproduction of hierarchies online: The case of crowdfunding of scientific research

Abstract

We conceptualize mechanisms that explain how social uses of media technologies, especially online platforms and crowds, reproduce, or modify inequalities, and explore these in the context of the crowdfunding of science. We distinguish between “supply side” factors related to the ability of actors given their institutional standing to use this funding approach, and “demand side” factors related to the crowd’s sensitivity to the institutional standing of those actors. We collected data on scientists requesting funding for their studies on Experiment.com , arguably the most popular scientific crowdfunding platform, and investigated the factors contributing to initiation and success. Supply side factors were important: crowdfunding appeals tended to come from scientists affiliated with larger, wealthier, and more active and prestigious institutions. However, demand side factors were not as important at the institutional level. Crowdfunding projects’ success was not predicted by the institution’s status, but rather by the number of appeals from an institution.

Full text

Title: The contribution of supply and demand factors to the reproduction of
hierarchies online: The case of crowdfunding of scientific research
Authors: Roei Davidson1, Yariv Tsfati1*.
Affiliations:
1Department of Communication, University of Haifa, 199 Abba Khoushy Ave., Haifa,
3498838, Israel.
*Correspondence to: ytsfati@com.haifa.ac.il
The authors contributed equally to this study.
Abstract: We conceptualize possible mechanisms that explain how social uses of media
technologies reproduce, intensify or narrow socio-economic inequalities, and explore these in the
context of the crowdfunding of science. We distinguish between “supply side” factors related to
the ability of actors given their institutional standing to use such online crowd practices, and
“demand side” factors related to the crowd’s sensitivity to the institutional standing of those
actors. We collected data on scientists requesting funding for their studies on Experiment.com – a
popular scientific crowdfunding online platform and investigated the factors contributing to
initiation and success. Supply side factors were important: crowdfunding appeals tended to come
from scientists affiliated with larger, wealthier and more active and prestigious institutions.
However, demand side factors were not as important at the institutional level, as crowdfunding
projects’ success was not predicted by the institution's status, but rather by the number of appeals
from an institution.
Keywords: Science and economics, crowdfunding, knowledge industries, inequality, media and
science
1

Introduction
The production of knowledge, in science and technology, literature, and popular culture
could be conceptualized as a process of selection. Many knowledge creators have
embryonic ideas, large parts of them carry promise, but only a few come to fruition. In all
these fields, the selection favors socially, symbolically or economically stronger actors.
For decades, media scholars have been debating the role of communication technologies
in knowledge production, consumption and acquisition (e.g., Donohue et al., 1975; Pool,
1983), and more specifically the extent to which technology weakens or intensifies the
existence of social hierarchies in this domain. In this paper, we turn the spotlight on two
sides of the selection process: The “supply side” factors relate to the tendency of
economically or socially stronger actors to seek and successfully secure resources
necessary to fulfill their ideas and talents. The “demand side” factors relate to the
preferences of the public. Technology could potentially play a role on both sides of the
process, by lowering entry barriers on the supply side, and by allowing audience
preferences to independently express themselves on the demand side, with less
centralized gatekeeping. In this paper we examine the role of supply and demand factors
in the context of the crowdfunding of scientific research.
Crowdfunding is a novel distributed online approach for raising finance for diverse types
of endeavors and initiatives (Younkin and Kashkooli, 2016). Popular discourse regarding
crowdfunding (Grant, 2014) has echoed optimistic assessments of the egalitarian impact
of digital distributed technologies (Benkler, 2006) but some studies have found that
social and geographic barriers are also evident in crowdfunding (Mollick, 2014;
2

Davidson and Poor, 2015). In this study, we distinguish between supply and demand
barriers and their operations in the crowdfunding of scientific research. We focus on the
funding of science given that, unlike in realms such as artistic production, in which
reputation of contemporary producers is extremely fluid, and sometimes domain-specific
(Becker, 2008), in the academic realm ranking is well defined, both institutionally and
individually (Bourdieu, 1975).
The remainder of the paper is organized as follows: We first consider the role of status in
shaping supply and demand in general. We then turn to evidence of the operation of
status in the traditional funding of science, and in crowdfunding in general. Based on this
evidence we generate and test a hypothesis and pose as well as answer research questions
regarding the role of status in shaping the supply of scientific crowdfunding projects and
the crowd’s demand for them. We conclude the paper with a discussion of the findings
with a particular emphasis on the role of supply and demand-side factors in the effects of
technology on the allocation of resources in science and beyond.
Literature Review
Supply and demand factors in the selection of knowledge products
The tendency of economically or socially stronger actors to seek and successfully secure
resources necessary to fulfill their ideas and talents may stem from a variety of reasons.
On the supply-side, children of affluent parents receive material support that allows them
to persist in their chosen occupation even if at first it is not especially lucrative
(Friedman, O’Brien and Laurison, 2016). In addition, some middle-class children are also
exposed by their parents to diverse forms of culture and encouraged to engage in cultural
3

pursuits. These, in turn, aid them in securing employment in creative occupations in
which employers have similar diverse cultural tastes (Koppman, 2015). Further, access to
digital media technologies which is socio-economically stratified may be tied to the
“propensity to engage in entrepreneurial activity” and is ultimately linked to achieving
more financial success (Robinson et al., 2015, p. 575).
Demand-side reasons for the success of high-status actors – “superstars” (Rosen, 1983) –
in securing resources relate to the fact that familiarity and reputation are associated with
risk reductions in business, sports, entertainment and many other domains. So, for
example, when large Hollywood studios decide to fund projects offered by accomplished
screenwriters and directors, the risk is lower based on the assumption that audience
demand for the work of the acclaimed and well-known writer or directors will be high.
Similarly, art work valuations are pegged to an artist’s track record and the identity of their
associated galleries (Velthuis, 2005).
Science and institutional stratification
As in other domains, scientific resources such as research grants have traditionally tended
to accumulate in the hands of a small minority (Merton, 1968). Out of 562 American
research institutions that report any federal research funding (The Center for Measuring
University Performance, 2016), only 60 elite universities (10.7% of the 562 research
institutions described above) accounted for 59% of all federal funding of research and
development in universities (Association of American Universities, 2015).
4

The reasons for this concentration are linked both to supply and demand. On the supply
side, start-up funding from well-endowed universities enables researchers to carry out
pilot studies demonstrating the merit of their ideas. Such universities are also more likely
to adopt new media technologies (Stephan, 2012) which are especially relevant to
crowdfunding given its dependence on the efficient production and distribution of well-
crafted online messages.
On the demand side, which has attracted more research attention, peer-reviews of grant
proposals evaluate more positively proposals that include such preliminary evidence
(Stephan, 2012). In interactions within grant-review panels, faculty affiliated with elite
institutions might enjoy more influence and might take institutional affiliation into
account in their grant evaluations (Lamont, 2009, p. 147). There is some evidence that the
institutional affiliation of manuscript submitters or grant applicants impacts peer-review
decisions although the evidence is not conclusive (Bornmann, 2011). Referees tend to
judge scientific projects according to their own intellectual habitus and grant panelists
tend to prefer research familiar to them from their own professional networks. As a result,
the process “is perceived to be biased against daring and innovative research” (Lamont,
2009, p. 241). In other words, there is some evidence that peer-review as an expression of
professional scientific demand for research is guided by status.
We wish to examine whether, when appealing to the public through crowdfunding rather
than to dedicated institutional mechanisms controlled by the scientific establishment,
status shapes the supply of crowdfunding projects and the public demand projects enjoy
as reflected in the funding projects attract.
5

Hierarchies in crowdfunding
We turn now to research that has touched whether obliquely or more directly on the role
of status in shaping the supply of crowdfunding projects or the public’s demand for such
projects. A large portion of the multi-disciplinary literature focused on crowdfunding
since its emergence less than a decade ago has attempted to understand and identify the
project attributes that predict funding success, often as a means of enunciating best
practices for crowdfunding founders rather than as a means for building or testing theory
(e.g., Mollick 2014; Frydrych et al. 2014). On the supply side, the advantages enjoyed by
producers located in large cultural hubs, where social norms legitimize cultural work as a
worthwhile pursuit, and the talent and knowledge needed to develop creative projects is
abundant (Currid, 2007; Scott, 2014), could also advantage producers who wish to
crowdfund their project. Further, there is evidence that platforms enhance such
geographic advantage by selectively promoting centrally-located projects more than
projects located in the periphery (Davidson and Poor, 2018).
On the demand side, some studies have focused on the importance of social ties both
offline and online, and more generally on the existence of a community as a requirement
for crowdfunding success. Such studies have generally found that many founders rely on
existing friends and family for funding (Hui, Greenberg and Gerber, 2014; Davidson and
Poor, 2015; Muller et al., 2016; Wang, 2016). Hence, the role of social hierarchies in
crowdfunding has been identified indirectly as the reliance on friends and family signals
that those from more affluent backgrounds might be more likely to financially benefit
from links to affluent close others.
6

When taken as a whole, the results of more direct tests of how status might impact
crowdfunding decisions are mixed. Scholars have found that the racial identity of
entrepreneurs is related to their ability to raise funding from the crowd even when
controlling for pitch quality (Younkin and Kuppuswamy, 2017) suggesting that the
funding public is also sensitive to identity cues. In an analysis of an Australian equity
crowdfunding platform the proportion of directors with an MBA degree on a firm’s
board, an indicator of status, was positively related to success (Ahlers et al., 2015).
Turning specifically to the crowdfunding of science, a study of scientific projects on a
number of platforms found that academic ranking was not significantly related to
crowdfunding success (Schäfer et al., 2016). A recent extensive analysis of individual
level factors predicting crowdfunding success on the particular platform we also study
here found that junior scholars enjoy more crowdfunding success than senior scholars
(Sauermann, Franzoni and Shafi, 2018).
Turning to gender, women enjoy more crowdfunding success than men. This success
seems to be related to the support they receive from activist female funders who perceive
female entrepreneurs to be under-represented (Greenberg and Mollick, 2015). Similarly,
an analysis of scientific crowdfunding found that “women have higher odds of reaching
their funding goal than men” (Sauermann, Franzoni and Shafi, 2018, p. 13). A
comparison of the spatial dispersion of venture capital investment and successful
crowdfunding in the U.S. found that successful crowdfunding projects tend to be more
dispersed and located outside cultural hubs (as well as inside them) suggesting on the
supply side that the crowd does relax some status-related constraints (Sorenson et al.,
2016). Further, a comparison of expert judges to the crowd found that the crowd
7

supported the funding of more diverse theatrical projects than the experts did (Mollick
and Nanda, 2015). Given the mixed findings we will ask to what extent are crowd
preferences related to the academic rank of project initiators and the status of the
institution with which they are affiliated.
Hypothesis and research questions
General evidence of the “intergenerational gifting of capital” (Friedman, O’Brien and
Laurison, 2016, p. 9) in the culture industries, and evidence of the economic and cultural
advantages researchers in elite institutions enjoy are manifested in academia in the start-
up funds and light course load that young faculty in well-funded research institutions
enjoy. In turn, these advantages allow them to devote time to research and grant writing.
Such a dynamic is consistent with initial findings suggesting status shapes the supply of
crowdfunding projects. Similarly to the way parents in creative occupations expose their
children to diverse cultural experiences (Koppman, 2015), the cultural emphasis in elite
research institutions on grant seeking as an indicator of academic excellence as well as
the intellectual infrastructure (e.g., students, mentors, training) provide an intellectual
advantage to scholars in elite institutions. Further, better access to new media
technologies and products (e.g., software) in elite universities, required to produce
effective crowdfunding campaigns, provides such scholars a technical advantage. For all
these reasons, supply factors could reflect status differentials in the crowdfunding of
science. We therefore hypothesize:
H1: The institutional standing of a research institution will be positively related to the
likelihood that researchers affiliated with that institution will attempt to crowdfund.
8

On the demand side, there is some evidence that the peer-review process advantages
researchers in elite institutions. In crowdfunding, the evidence is mixed: some studies
suggest supporters back crowdfunding projects affiliated with high status individuals,
while other studies point out that in comparison to traditional forms of funding backers
support a more diverse array of projects. Therefore, we ask:
RQ1: Is the standing of the institution with which a project is affiliated related to its
chances of success?
RQ2: Is the academic rank of the scientist initiating a project related to that project’s
chances of success?
Method
To study these questions, we collected data on scientists requesting funding for their
studies on Experiment.com, a platform dedicated to the crowdfunding of science. As of
April 2014, this site had by a considerable margin the largest number of completed
crowdfunded science projects of any crowdfunding platform whether general-purpose or
dedicated to science (Schäfer et al., 2016). We collected data from July 20, 2014 till
April 7, 2016. Data were manually collected from Experiment.com by research assistants.
An assistant collected initial data (including URL) after the project was first posted
because failed projects were usually less accessible after the funding period had
concluded. A research assistant then returned to the project to collect full data on it after
its funding period had concluded. In this period, 427 projects were posted on the website,
of which we removed 77 projects that were proposed by teams that did not include at
least one university-affiliated scholar (student, post-doc, researcher or faculty member).
9

The analysis focuses on 333 projects that were proposed by teams that included at least
one scholar affiliated with a US university, as data on university endowments and
research grants were available only for US universities.
To study the institutional underpinnings of the crowdfunding of science, we compared
US academic institutions from which project initiators originated, to US institutions from
which no projects were initiated at the time of study, on four indicators of academic
status and resources. Appropriate data were available from the Center for Measuring
University Performance (MUP) at Arizona State University. Institutional data includes
562 U.S. institutions that as of 2011 had reported any federal research funding in the five
preceding years according to MUP data. The full dataset included 782 institutions.
However, due to missing data on institutional endowments, the number of cases dropped
to 562 institutions for institutional analyses.
We used the most recent wave of data released by the center that preceded our data
collection beginning in 2014. Student enrollment served as an indicator of an institution's
size. Measures were for 2009-2011 in thousands of students. Endowment was taken as an
indicator of an institution's wealth, and transformed into deciles because of its skewed
distribution from MUP data on Endowment in 2011 in thousands of Constant 1998 US
Dollars for research institution (some institutions failed to report endowment data to
MUP contributing to a decreased sample size). A measure of federal research grants
received by the institution was included as an indicator of the volume and quality of
research activity, and in particular of research requiring resources (indicator was
measured in deciles transformed from MUP data in 2010 on Total Federal Research in
thousands of Constant 1983 US Dollars for research institution). Finally, the number of
10

academy members in an institution, an indicator of institutional achievement and
prestige, was taken from MUP data on National Academy Memberships of faculty
members by institution in 2011.
In the institutional analysis, the number of initiated projects per institution served as one
dependent variable. This measure related to the number of projects with at least one
crowdfunding founder (member of the crowdfunding team initiating the project) affiliated
with a given research institution. The affiliation data were collected manually from
project pages on the crowdfunding platform where initiators were able to note current
affiliation. As an alternative dependent measure we transformed the above measure into a
dichotomous initiation measure coded as a dummy variable: 1 = at least one project by a
project founder affiliated with institution; 0 = no researchers affiliated with institution
had initiated a project.
We coded the rank of project founders – a key independent variable at the individual
level – on an ordinal scale: 1 = no rank, unknown, student, postdoc, adjunct; 2 = assistant
professor; 3 = associate professor; 4 = full professor. All ranks were coded from the
project page as indicated by project founders. Maximal rank is the rank of the highest
ranked member in the team. Minimal rank is the rank of the lowest ranked member in the
team.
To examine crowdfunding outcomes, we collected data on pledged funds raised by a
project. These were collected manually from project pages after the funding period had
concluded where they appeared in a dedicated field at the top right corner of the page.
Further we collected data on the project goal – the amount of money the project founders
11

indicated as their goal on the project page in the dedicated field at the top right corner of
the page. We should note that in 21 of 350 cases we detected that the goal was lowered
during the funding period. In those cases we utilized the updated goal. We then derived
from these data a measure of success. We coded a project as successful when funds raised
were equal or greater than the goal. We used this measure because the platform operates
according to an “all or nothing model” whereby scientists received funding only when the
funds raised surpassed the goal.
In our regression analysis, we controlled for geographic location for the institution. We
created a dichotomous measure of institutional location in a hub or outside of it
aggregated from U.S. State as indicated in 2015 MUP data. Missing data were added
from university name if the state was included, or else from the Google search profile for
the institution in March 2017. Universities in Midwest (IN, MI, OH, WI, IL, IA),
Northeast (CT, DC, DE, MA, MD, ME, NH, NJ, NY, RI, VT, PA) and Pacific (CA, OR,
WA) states coded as located in research hubs (1), all other institutions coded as located
outside research hubs (0). We also controlled for the existence of a project video for the
number of team members given prior evidence of their positive impact on funding levels
(Mollick, 2014; Muller et al., 2016).
Given the differences between scientific disciplines in the propensity to engage the public
(Jensen, 2011), we also controlled for the disciplinary affiliation of the projects. We
coded the research domain of a project based on disciplinary tags that appeared in
structured fields on project pages: Life sciences (Paleontology, Medicine, Neuroscience,
Biology, Ecology) = 1; Exact sciences (Mathematics, Physics, Materials Science, Earth
Science, Engineering, Data Science, Computer Science, Chemistry) = 1; Human Sciences
12

(including both social science and humanities: Education, Psychology, Anthropology, Art
and Design, Political Science, Economics, Social Science) = Reference category.
Data analytic strategy
To study the supply side, at the institutional level we regressed initiation of at least one
project by an institution, the number of initiated projects per institution and the success of
at least one project affiliated with an institution on the aforementioned indicators of
institutional standing as well as the geographic control variable. In the case of project
success, we also included a measure of the number of projects affiliated with the
institution. Further we examined the indirect effect of the independent variables on
success of at least one Crowdfunding project affiliated with a given US institution
(mediated by the number of projects affiliated with an institution), using the PROCESS
macro test for the significance of the indirect effect (Hayes, 2013).
We also conducted analysis at the level of the individual project to examine demand side
factors. Here we regressed two measures of success – pledged funds and a ratio of
pledged funds on initial goal – on a number of measures of the standing of individual
members of the team: the highest academic rank of a member of a project team, and the
lowest rank as well as the decile for the institution that raised the most federal funding
among the institutions with which team members were affiliated among American
academic research institutions in general. We also controlled for disciplinary affiliation,
for the existence of a video on the project page, and for the number of project founders
for a given project.
Results
13

In the period of study (roughly 22 months), on the most popular online platform for
crowdfunding scientific research, USD 979,905 were donated by 12,359 individuals to
support research at academic institutions worldwide. On average, the projects we
analyzed on the prominent science-dedicated crowdfunding platform, Experiment.com,
set funding goals of USD 5094.85 (SD = 7167.96), and were able to solicit USD 2871.18
(SD = 4330.92) per project. A majority of the projects (54.79%) reached their pre-
determined funding goal and got funded. This funding rate is higher than the overall
success rate for projects on Kickstarter (Kickstarter.com, 2018) – the most popular
general-purpose crowdfunding platform. Of those that succeed, most cluster just above
the preset goal, similarly to the funding patterns for Kickstarter (Mollick, 2014). A
majority of the projects belonged to the life sciences (69.74%), 11.01% to exact sciences,
and 19.35% to the human sciences (social sciences and humanities). Sample project titles
include: "How does our blood absorb light?", "Eating disorders in college women – A
new treatment approach", "Next generation non-seeing eye prosthesis", "Editing
photographs in three dimensions", "Can bacteria turn light into fuel?" and "Can cognitive
tests of shelter dogs improve their chances of adoption?"
14




 !
"#$%&'()
(*%
"*)(
'*#(
*%
)*&
)$*(%
$)*&#
"*"
Figure 1: Comparing instuons from which crowdfunding projects were
iniated to other instuons +,-
./
,-
./
Note: All differences are statistically significant at the p < .001 level, based on
independent sample t-tests
H1 concerned the supply side of crowdfunding, and predicted that the standing of the
institution will be related to the likelihood that researchers affiliated with the institution
will crowdfund. To test for this hypothesis, we first compared institutions from which at
least one crowdfunding appeal originated, to all other institutions, on a set of markers of
academic standing. Results, presented in Figure 1, demonstrate that crowdfunding
appeals are more likely to come from larger institutions (in terms of student enrollment,
19.61K, compared to 8.5K on average; t (199.74)= 9.81, p < .001), from institutions
with larger endowments (948.5M compared to 209.8M, t (171.56) = 4.10, p < .001) and
larger federal research budgets (49.6M compared to 7.38M, t (177.3) = 7.31, p < .001),
15

and from institutions with more members of the National Academy of Sciences among
their ranks (20.2 compared to 1.5; t (170.69) = 4.92, p < .001).
16

Table 1. Regression models predicting initiation, number of initiated projects, and
success of at least one Crowdfunding project in experiment.com (n=562 institutions)
Model 1:
Initiation
(logistic regression)
Model 2:
# of initiated projects
(poisson regression)
Model 3:
Success
(logistic regression)
B (SE) EXP (B) B (SE) EXP (B) B (SE) EXP (B)
Federal research
funding (by decile)
.13 (.06)* 1.14 .24
(.04)***
1.27 .07 (.08) 1.08
Endowment (by
decile)
.13
(.05)**
1.14 .07
(.03)*
1.07 .02 (.06) 1.02
Student enrollment (in
thousands)
.06
(.01)***
1.06 .03
(.00)***
1.03 .02 (.01) 1.02
Academy members .02 (.01)* 1.02 .00
(.00)**
1.00 -.01 (.00) .99
Geographic location
(research hubs = 1)
.16 (.23) 1.17 .21 (.13) 1.23 .22 (.31) 1.25
Number of projects 1.03
(.16)***
2.81
Pseudo R2.23 .42 .19
Note. * p <.05, **p <.01, *** p <.001
As a test of whether these differences are accounted for by other factors, we conducted a
logistic regression analysis predicting initiation of projects at the institution level using
federal research funding, endowment, student enrollment, and the number of Academy
17

members, controlling for geographic location, known in the research literature to be
correlated with the success of crowdfunding projects (Mollick, 2014). Results (reported
in Table 1, Model 1) demonstrate that these differences remained statistically significant
even when controlling for one another, and for geographic location. This analysis
explained 23.2% of the variance in initiation. The same was true when predicting the
number of projects initiated in each of the institutions (the Poisson regression model,
explaining 42% of the variance in the number of projects, is reported in Table 1, Model
2). That is, the more students enrolled at an institution, the more well-funded and well-
endowed the institution, and the more the institution has reputable and accomplished
faculty members among its ranks, the higher the likelihood that more crowdfunding
projects were initiated by scholars at that institution.
In sum, initiation of crowdfunding projects at an institution is positively associated with
its size, funds, current research grants and prestige, as expected by H1. Do these factors
also contribute to the success of scholars from these institutions in securing funding for
their projects from the crowd? This was the focus of RQ1.To investigate this question, we
ran a logistic regression model predicting success of at least one of the projects from the
institution, that is, that at least one of the crowdfunding projects initiated at an institution
has succeeded in reaching its goal. Of course, more projects increase the odds of an
institution to succeed, and this is why the number of projects is controlled for in the
model. Results (Table 1, Model 3) demonstrate that, all else being equal, neither grants,
endowments, Academy of Sciences memberships nor student enrollment were
significantly associated with success. The only factor significantly associated with
success was the number of projects. Each additional project initiated at an institution was
18

associated with an increase of 281% in the odds that one of them will be successfully
crowdfunded after controlling for all other factors.
Given that, as demonstrated above, the number of projects initiated at an institution was
itself associated with endowments, federal grants, student enrollment and Academy of
Sciences memberships, we tested the possibility of an indirect impact of these factors on
success at the institutional level, mediated by the number of projects initiated at an
institution. This test found that the number of projects was a significant mediator in the
association between these factors and success (see Table 2). That is, prestige, size and
financial resources contributed positively to the initiation of more crowdfunding projects,
which in turn increased the odds of success. In other words richer, larger and more
prestigious institutions do succeed more in crowdfunding because more projects are
initiated in such institutions.
19

Table 2: Indirect effect of independent variables on success of at least one Crowdfunding
project in experiment.com through the number of projects initiated in US institutions
of higher education (n = 562 institutions)
B
(bootstrap SE)
LLCI ULCI
Federal research funding
(by decile)
.04 (.02)* .00 .09
Endowment (by decile) .04 (.02)* .01 .09
Student enrollment (in
thousands)
.03 (.01)* .01 .05
Academy members .01 (.01)* .01 .05
Geographic location
(research hubs = 1)
.10 (.09) -.08 .29
Notes: * p < .05, Indirect effects, standard errors and confidence intervals were
calculated using PROCESS macro (Hayes, 2013)
We also examined success at the individual project level as a means of analyzing the
demand for scientific crowdfunding projects, in addition to examining the supply of
projects at the institutional level. To do so, we investigated the factors contributing to the
sum pledged by each of the projects (in US Dollars) and the funding success rate (the
ratio of funding raised to the original funding goal). We regressed these two outcomes on
several project characteristics: the highest and lowest academic rank of project initiators
(on a scale varying between 1=student or junior scholar to 4= full professor), the field the
project came from (both exact and life sciences were compared to human sciences),
20

whether the initiators used a video to promote their projects, and the number of project
initiators. As an indicator of the funding resources available at the institutions in which
the teams operated, we ranked (in deciles) the universities in research grant terms. In
cases where several scholars cooperated on a project, we used the highest ranking
institution among team members. Results are presented in Table 3.
Table 3: Regression models predicting pledged funds and pledged to goal ratio,
crowdfunding projects with at least one creator originating from a US university (n=
335)
Model 1:
Pledged funds
Model 2:
Pledged to goal
ratio
Life sciences (=1) 274.37 (587.65) .05 (.09)
Exact sciences (=1) -1513.18 (858.42)#-.28 (.13)*
Video included (=1) 1436.77
(489.63)**
.21 (.08)**
Maximal rank -482.47 (300.38)#-.12 (.04)*
Minimal rank 433.00 (440.59) .03 (.07)
Number of project founders 896.58
(206.04)***
.14 (.03)***
Maximal federal research grants (decile) 85.18 (76.63) -.02 (.01)
R2.11 .12
Notes: # p < .10, * p <.05, **p <.01, *** p <.001
21

RQ2 concerned the association between the academic rank of a scientist and her odds of
securing funding in crowdfunding. We found that each one unit-increase on the 1-4
indicator of the maximal academic rank was associated with a decrease of USD 482 in
the amount pledged by the team. That is, ceteris paribus, teams with a full professor
raised on average USD 1929.6 less than teams including only students or junior scholars.
Similarly to the institutional-level models, the university-level indicator of available
research resources did not significantly impact success at the individual level. This result,
when coupled with the institutional level analysis above provides a consistent negative
answer to RQ1.
As in past research, each additional team member was associated with an increase of
USD 896.6 in the pledged amount. Echoing past research on crowdfunding, promoting
the funding request with a video was associated with a USD 1,436.8 increase in funds
raised. Life sciences did not significantly differ from human sciences on both indicators
of success. However, the exact sciences raised significantly less funds than the human
sciences.
Discussion
Is crowdfunding equalizing the funding of science? While a previous analysis of
individual level factors related to the crowdfunding of science has argued “that
crowdfunding of scientific research broadens access to resources for groups that have
been excluded or disadvantaged in traditional funding systems” (Sauermann, Franzoni
and Shafi, 2018, p. 16), we draw a more nuanced conclusion. Aided by our application of
the distinction between supply and demand factors to the online funding of science, a
22

particular form of knowledge production, we find that the answer to this question is
complex. Our data show that, by far, initiators of crowdfunded projects come from large
elite universities, well-resourced in terms of endowments and grants. Hence, the supply
of scientific crowdfunding projects is shaped by institutional status, with scientists
affiliated with higher ranked institutions more likely to consider a crowdfunding project.
However, while scholars from such elite institutions tend to participate more in
crowdfunding initiatives, their chances of success are not significantly higher than their
counterparts from less prestigious, less well-funded universities. Thus, lay crowd
members, who constitute the demand for scientific crowdfunding projects, do not seem to
look more favorably upon projects initiated by scientists enjoying higher status.
We conclude that it is the supply side, but not the demand side, of crowdfunding that
drives inequalities in recruiting funds from the public. This implies that, currently,
graduate students and junior scholars from leading institutions benefit most from
crowdfunding when they set out to crowdfund without a senior collaborator. Given that
our findings show that institutional status is not associated with greater success, the
scientific community should encourage junior scholars to independently crowdfund,
especially if they operate in the academic periphery. This could allow gifted young
researchers working in sparsely financed institutions to raise funds for their studies, and
thus fulfill the utopian vision of the internet as a more egalitarian and meritocratic public
sphere. Further, our findings suggest that “the august array of insignia adorning persons
of “capacity” and “competence” - the red robes and ermine, gowns and mortar-boards of
magistrates and scholars in the past, the academic distinctions and scientific
qualifications of modern researchers” (Bourdieu, 1975, p. 20) are of muted importance
23

when science funding decisions are handed over from members of the scientific
community to the crowd.
Our study is theoretically situated within a fundamental question in the study of new
media: do digital media technologies narrow, reflect or broaden gaps in access to
resources between the more and less affluent, and adds an important distinction between
the supply and demand sides of resource allocation in the crowdfunding of science and
cultural products more generally.
Our findings paradoxically support both sides of the debate on the role of the internet in
shaping social hierarchy. Early expectations (now called by some utopian) that online
technology would work to narrow gaps between haves and have-nots by lowering entry
barriers to the public sphere received support from the finding that status did not
significantly contribute to the success of individual projects, above and beyond controls.
On the other hand, skeptics’ argument that stronger actors enjoy the benefits of online
technologies more than relatively weaker ones also enjoyed the support of our data, in
that, when not controlling for the number of projects initiated by institutions, larger,
stronger and more prestigious institutions received on average more funding.
It is possible that scholars from stronger universities simply invest more effort in
crowdfunding. In turn, these differential efforts could themselves result from, or
otherwise reflect, difference in the academic hierarchy. For example, it could be that it is
not only that the academic culture in stronger institutions stresses the importance of more
expensive research or of being funded. Rather than, or in addition to these pressures,
junior scholars operating in a stronger environment enjoy more resources that are relevant
24

for crowdfunding: more time given lower teaching loads, better studios to create more
successful pitches and so forth.
We cannot tell whether funding goes to the more scientifically-solid and deserving
projects, but it is very likely (given the positive effects of posting a video and having
more team members) that putting more effort into promoting the crowdfunding campaign
increases the chances of success. It is unclear why teams without senior scholars are on
average more successful in crowdfunding their research. It is possible that teams of junior
scholars were more active in the promotion of their projects. In addition, it is possible
that members of the public are reluctant to support researchers that they perceive as
already well-funded.
Limitations and directions for future research
An important implication of our findings is that future research should pay closer
attention to deciphering the mechanisms behind empirical findings demonstrating
inequalities in the digital pooling of resources in crowdfunding, crowdsourcing and other
forms of collective action online. Such research should consider whether involving the
public more directly in funding decisions will open up opportunities for additional
scientists, and what impact would such involvement have on the quality of the research
produced. Such careful attention could potentially broaden our understanding of the
complex association between technology and resource allocation in other domains and
therefore offer more nuance to our understanding of the social implications of emerging
media and online platforms.
25

We argued above that the context of science is suitable for exploring supply and demand
factors’ role and their interaction with status in shaping resource allocation online given
the fact that in science status is more explicit and perhaps salient compared to other
domains. However, this choice has its limitations.
One should note that our study is limited to one particular mechanism - crowdfunding, in
a particular knowledge industry – science, on a dominant though particular funding
platform. On the one hand, in other cultural domains such as fine arts quality is
“inscrutable” (Gambetta, 1994) and especially difficult to ascertain, and therefore status
is a convenient shortcut shaping both supply and demand (Aspers, 2009). On the other
hand, in knowledge industries beyond science, barriers to participation are now
considerably lower (Doyle, 2013; Waldfogel, 2017) and therefore status could be of less
importance in shaping the supply of knowledge producers appealing to the public directly
for funding. Future studies should therefore test similar hypotheses and research
questions in additional domains, contexts and platforms.
26

References
Ahlers, G. K. C. et al. (2015) ‘Signaling in Equity Crowdfunding’, Entrepreneurship
Theory and Practice, 39(4), pp. 955–980. doi: 10.1111/etap.12157.
Aspers, P. (2009) ‘Knowledge and valuation in markets’, Theory and Society. Springer
Netherlands, 38(2), pp. 111–131. doi: 10.1007/s11186-008-9078-9.
Association of American Universities (2015) By the numbers. Available at:
http://www.aau.edu/WorkArea/DownloadAsset.aspx?id=13460 (Accessed: 18 May
2016).
Becker, H. S. (2008) Art Worlds. 25th anniv. Berkeley: University of California Press.
Benkler, Y. (2006) The Wealth of Networks: How Social Production Transforms Markets
and Freedom. New Haven: Yale University Press.
Bornmann, L. (2011) ‘Scientific peer review’, Annual review of information science and
technology, 45(1), pp. 197–245.
Bourdieu, P. (1975) ‘The specificity of the scientific field and the social conditions of the
progress of reason’, Information (International Social Science Council), 14(6), pp. 19–
47.
Currid, E. (2007) The Warhol economy: How fashion, art, and music drive New York
City. Princeton, NJ: Princeton University Press.
Davidson, R. and Poor, N. D. (2015) ‘The barriers facing artists’ use of crowdfunding
platforms: Personality, emotional labor, and going to the well one too many times’, New
27

Media & Society, 17(2), pp. 289–307.
Davidson, R. and Poor, N. D. (2018) ‘Location, location, location: How digital platforms
reinforce the importance of spatial proximity’, Information, Communication & Society.
Donohue, G. A., Tichenor, P. J. and Olien, C. N. (1975) ‘Mass Media and the Knowledge
Gap A Hypothesis Reconsidered’, Communication Research, 2(1), pp. 3–23.
Doyle, G. (2013) Understanding media economics. London: Sage.
Friedman, S., O’Brien, D. and Laurison, D. (2016) ‘“Like Skydiving without a
Parachute”: How Class Origin Shapes Occupational Trajectories in British Acting’,
Sociology. doi: 10.1177/0038038516629917.
Frydrych, D. et al. (2014) ‘Exploring entrepreneurial legitimacy in reward-based
crowdfunding’, Venture Capital, 16(3), pp. 247–269. doi:
10.1080/13691066.2014.916512.
Gambetta, D. (1994) ‘Inscrutable markets’, Rationality and Society, 6(3), pp. 353–368.
Grant, T. (2014) To find seed money, Canadian startups follow the crowd, The Globe and
Mail. Available at: http://www.theglobeandmail.com/report-on-business/to-find-seed-
money-canadian-startups-follow-the-crowd/article17631251/.
Greenberg, J. and Mollick, E. (2015) ‘“Leaning In or Leaning On? Gender, Homophily,
and Activism in Crowdfunding”’, Academy of Management Proceedings , 2015(1). doi:
10.5465/AMBPP.2015.18365abstract.
Hayes, A. F. (2013) Introduction to mediation, moderation, and conditional process
28

analysis: A regression-based approach. New York: Guilford Press.
Hui, J. S., Greenberg, M. D. and Gerber, E. M. (2014) ‘Understanding the role of
community in crowdfunding work’, Proceedings of the 17th ACM conference on
Computer supported cooperative work & social computing. Baltimore, Maryland, USA:
ACM. doi: 10.1145/2531602.2531715.
Jensen, P. (2011) ‘A statistical picture of popularization activities and their evolutions in
France’, Public Understanding of Science , 20(1), pp. 26–36. doi:
10.1177/0963662510383632.
Kickstarter.com (2018) Stats. Available at: https://www.kickstarter.com/help/stats.
Koppman, S. (2015) ‘Different Like Me: Why Cultural Omnivores Get Creative Jobs’,
Administrative Science Quarterly, 61(2), pp. 291–331. doi: 10.1177/0001839215616840.
Lamont, M. (2009) How professors think: Inside the curious world of academic
judgment. Harvard University Press.
Merton, R. K. (1968) ‘The Matthew Effect in Science’, Science, 159(3810), pp. 56–63.
Available at: http://www.jstor.org/stable/1723414.
Mollick, E. (2014) ‘The dynamics of crowdfunding: An exploratory study’, Journal of
Business Venturing, 29(1), pp. 1–16. doi:
http://dx.doi.org/10.1016/j.jbusvent.2013.06.005.
Mollick, E. and Nanda, R. (2015) ‘Wisdom or Madness? Comparing Crowds with Expert
Evaluation in Funding the Arts’, Management Science, 62(6), pp. 1533–1553. doi:
10.1287/mnsc.2015.2207.
29

Muller, M. et al. (2016) ‘Social Ties in Organizational Crowdfunding: Benefits of Team-
Authored Proposals’, in Proceedings of the 19th ACM Conference on Computer-
Supported Cooperative Work & Social Computing. New York, NY, USA: ACM (CSCW
’16), pp. 1246–1259. doi: 10.1145/2818048.2819955.
Pool, I. de S. (1983) Technologies of freedom. Cambridge, MA: Belknap.
Robinson, L. et al. (2015) ‘Digital inequalities and why they matter’, Information,
Communication & Society, 18(5), pp. 569–582. doi: 10.1080/1369118X.2015.1012532.
Rosen, S. (1983) ‘The Economics of Superstars’, The American Scholar, 52(4), pp. 449–
460. Available at: http://www.jstor.org/stable/41210977.
Sauermann, H., Franzoni, C. and Shafi, K. (2018) Crowdfunding Scientific Research.
National Bureau of Economic Research.
Schäfer, M. S. et al. (2016) ‘Selling science 2.0: What scientific projects receive
crowdfunding online?’, Public Understanding of Science. doi:
10.1177/0963662516668771.
Scott, A. J. (2014) ‘Beyond the Creative City: Cognitive–Cultural Capitalism and the
New Urbanism’, Regional Studies, 48(4), pp. 565–578. doi:
10.1080/00343404.2014.891010.
Sorenson, O. et al. (2016) ‘Expand innovation finance via crowdfunding’, Science,
354(6319), p. 1526 LP-1528. Available at:
http://science.sciencemag.org/content/354/6319/1526.abstract.
Stephan, P. E. (2012) How economics shapes science. Cambridge (MA): Harvard
30

University Press.
The Center for Measuring University Performance (2016) American Research University
Data.
Velthuis, O. (2005) Talking prices: Symbolic meanings of prices on the market for
contemporary art. Princeton University Press.
Waldfogel, J. (2017) ‘How Digitization Has Created a Golden Age of Music, Movies,
Books, and Television’, Journal of Economic Perspectives, 31(3), pp. 195–214.
Available at: http://www.aeaweb.org/articles?id=10.1257/jep.31.3.195.
Wang, C. (2016) ‘The Promise of Kickstarter: Extents to Which Social Networks Enable
Alternate Avenues of Economic Viability for Independent Musicians Through
Crowdfunding’, Social Media + Society, 2(3), p. 2056305116662394. doi:
10.1177/2056305116662394.
Younkin, P. and Kashkooli, K. (2016) ‘What Problems Does Crowdfunding Solve?’,
California Management Review, 58(2), p. 20 LP-43. Available at:
http://cmr.ucpress.edu/content/58/2/20.abstract.
Younkin, P. and Kuppuswamy, V. (2017) ‘The Colorblind Crowd? Founder Race and
Performance in Crowdfunding’, Management Science. doi: 10.1287/mnsc.2017.2774.
31