Crowdfunding money for research levels the playing field

Chiara Franzoni, Henry Sauermann, Kourosh Shafi 14 February 2019

a

A

Research has grown increasingly expensive. In the life sciences, a biotech lab at a public university in the US costs about $550,000 per year (Stephan 2012). The funding to do basic science comes predominantly from government, often through dedicated granting agencies like the European Research Council in Europe and the National Institutes of Health (NIH) in the US. 

At these agencies, funding is distributed by committees of scientists who assess the scientific merit of the proposals and the qualifications of the principal investigators. This system has worked well in selecting promising research (Li 2017), but has also been criticised for perceived bias against women, young scholars, and novel or risky research (Ceci and Williams 2011, Daniels 2015, Edwards et al. 2011, Lamont 2010). 

New systems of science funding

Following the rapid growth of crowdfunding in other domains, such as entrepreneurship and the arts (Mollick 2014), platforms dedicated to crowdfunding research are now providing scientists with the opportunity to raise money directly from citizens, without going through the peer review system (Colombo et al. 2015). The emergence of this new system raises many questions: Who are the scientists that try to raise money in this way? What kind of projects do citizens like to fund? Does the ‘crowd’ suffer from the same biases argued to exist in the traditional funding system? 

To answer these questions, we analysed data about fundraising campaigns conducted in Experiment.com (Sauermann et al. 2019). This crowdfunding platform has been operating since 2012 and is dedicated only to scientific research. With over 1,000 funded projects, it is by far the largest platform if this type. 

Experiment.com uses a standard crowdfunding model called ‘all or nothing’. In this system, a scientist who launches a campaign chooses a target amount of funding for a clearly identified project. If the project reaches its target within a pre-determined time (typically between 30 and 45 days), the campaign is successful and the scientist gets the money. If not, the campaign fails and the scientist gets nothing. 

We collected and coded data about all 725 campaigns from the platform launch in May 2012 until August 2015. Of these campaigns, 68% were started by a single creator, and 32% were posted by teams of between two and seven creators, which gave a sample of 1,148 creators.

Who raises money from the crowd?

Figure 1 shows that 44% of the people who created crowdfunding campaigns were students (20% PhD students and 24% Master’s or undergraduate students). About one in four were university professors and of these, 14% were associate or full professors, and 9% were assistant professors. Postdoctoral scholars accounted for 6% of all creators. 

Figure 1 Share of project creators on Experiment.com by position, May 2012 to August 2015 Source: Sauermann et al. (2019).

The creators were predominantly from the US (89%) – which partly reflects that Experiment.com was founded in the US and focuses its outreach on that market – and 57% were male.

Also 81% of creators were based at educational institutions, including universities, medical schools and colleges; another 5% were at companies; and approximately the same amount were independent researchers, or creators with no affiliation. The few remaining creators were at other organisations, such as foundations and museums. 

What the crowd funds, and what work scientists want to be funded

Collectively, $4.37 million in raised funds was distributed to the scientists in our sample. Although crowdfunding for science tends to be used mostly for small projects compared to traditional government grants, it can also reach the magnitude of traditional grants. The median project raised $3,105, but several raised tens of thousands of dollars, and one raised more than $2.6 million. 

Projects belonged to many research areas, including (in descending order from most frequently listed field) biology, ecology, medicine, engineering, education, psychology, social sciences, and chemistry.

To understand what type of work scientists were pitching, we manually coded project descriptions. The results show that 78% of the projects involved the scientific investigation of a topic (treatment of diseases or climate change studies, for example), and often the goal is to publish the results in a peer-reviewed publication. Another 12% aim to develop devices, tools, software, or methods. The other 10% had different aims, such as the restoration of dinosaur skeletons.

We explored what kind of expenses scientists planned to cover with the funding, especially given that average targets appeared too small to pay salaries or the costs typically included in government grants. Materials, equipment, and services (60%) were the most frequent expenses specified. Other items included travel (16%) and salaries for personnel such as student assistants (11%). Creator salaries accounted for only 3% of the average budget.

Success rates and predictors of success in fundraising

A high number of campaigns (48%) reached their funding target, with many even exceeded their target. This is a high success rate compared with traditional granting agencies in the US (the NIH funds 16% of its applications) and general-purpose crowdfunding platforms that work in a similar way (on Kickstarter, for example, 36% of campaigns are successful). 

Some factors correlate with fundraising success. Students and postdocs were more likely to succeed in fundraising compared to professors, unlike the traditional funding system, in which juniors are often at a disadvantage (Daniels 2015). Women have been more successful than men, and receive larger donations. The success rate was no higher for projects that listed prior publications – in contrast to traditional funding mechanisms, for which prior scientific accomplishments are central to a scientist's ability to secure funding (Li 2017). Note these correlations are not causal.

We identified significant differences in the rates of success among scientists, but the scope or content of the proposals appeared to play a minor role (although our coded variables were also limited). Projects aimed at research were equally likely as those aimed at the development of devices, tools or software to get funded. To understand how funding success relates to the risk of a project, we analysed the frequency of words conveying risk and uncertainty in project descriptions. This did not predict funding success, suggesting that the crowd may not penalise scientists for describing work as risky. 

Campaigning for a project

It can take substantial time and effort for scientists to promote crowdfunding campaigns, but this effort appears to pay off. Campaigns that featured a video presentation of the project were more likely to succeed, as were campaigns that provided project background and campaign updates in 'lab notes' written by the creators. Endorsements from professional scientists or known personalities were also associated with funding success. 

Even though scientific research does not generate the kinds of products or services that can be sold, such as on other crowdfunding platforms, some campaigns offered creative rewards such as acknowledgements of donor names in future publications, photographs of animals that were being observed, or visits to the research lab. Campaigns with rewards like this were much more likely to succeed. Again this is not necessarily causal, since the success may partly reflect unobserved factors such as outreach effort, or the 'coolness' of the idea.

Our findings suggest that crowdfunding may open up new sources of funding for scientists. Funding decisions are fast, and the success rate is high. Young scholars and women, who have low success rates with traditional granting agencies, seem to have an advantage in crowdfunding. Risky projects, and projects with no prior publications, do not seem to be disadvantaged. 

But the amounts currently raised through crowdfunding are small, and so crowdfunding can only complement the funding of most traditional research projects. Success may also require a big investment of time and effort, as well as the ability to communicate effectively with the general public. Finally, crowdfunding is a relatively new phenomenon. As more scientists use it, the nature of the platform will change.

References

Ceci, S J and W M Williams (2011), “Understanding Current Causes of Women’s Underrepresentation in Science”, Proceedings of the National Academy of Sciences 108(8): 3157–3162.

Colombo, M G, C Franzoni, and C Rossi-Lamastra (2015), “Cash from the Crowd”, Science 348(6240): 1201–1202.

Daniels, R J (2015), “A Generation at Risk: Young Investigators and the Future of the Biomedical Workforce", Proceedings of the National Academy of Sciences 112(2): 313–18.

Edwards, A M, R Isserlin, G D Bader, S V Frye, T M Willson, F H Yu (2011), “Too Many Roads Not Taken”, Nature 470(7333): 163–65.

Lamont, M (2010), How Professors Think. Inside the Curious World of Academic Judgement, Harvard University Press.

Li, D (2017), “Expertise versus Bias in Evaluation: Evidence from the NIH”, American Economic Journal: Applied Economics 9(2): 60–92.

Mollick, E (2014), “The Dynamics of Crowdfunding: An Exploratory Study”, Journal of Business Venturing 29(1): 1–16.

Sauermann, H, C Franzoni, and K Shafi (2019), “Crowdfunding Scientific Research: Descriptive Insights and Correlates of Funding Success”, PLoS ONE 14(1): e0208384.

Stephan, P (2012), How Economics Shapes Science, Harvard University Press.

a

A

Topics:  Frontiers of economic research Productivity and Innovation

Tags:  crowdfunding, experiment.com, research, Grants, NIH

Associate Professor, Politecnico di Milano

Associate Professor of Strategy, ESMT European School of Management and Technology Berlin

Post-doctoral Research Fellow in Entrepreneurship, University of Florida

Events

CEPR Policy Research