The Data-Driven Discovery Initiative was initially approved in 2012 by the Moore Foundation’s board of trustees to explore ways of accelerating research that uses data science. In its initial six years and $60 million, the initiative employed three strategies:
- Institutions (fostering academic environments that nurture data-driven research, researchers, and tools through data science environments),
- Practices (fostering the development and dissemination of tools and resources that enable analyses of large, complex datasets) and;
- People (increasing the visibility of individuals at the intellectual forefront of data-driven science).
An external independent evaluation of the initiative’s first phase was conducted by Abt Associates from September 2016 through August 2017. The evaluation assessed the results that the Data-Driven Discovery Initiative had achieved, the role it had played in the data science landscape, how positive outcomes of the initiative could be sustained and future opportunities. Concurrently, a panel of academic and corporate experts assessed the initiative’s scientific impact to date and suggested future directions for work in this field.
Lessons from the external evaluation and the expert panel report include:
- Moderate investments in software tools can yield high impact. The evaluators noted the high value-to-dollar ratio of tools funded by the initiative’s Practices strategy, which was echoed by the expert panel. Tools and software are important for the future of data-driven research, and many of these tools are useful in disciplines different from the one for which they were developed. Both the expert panel and evaluators noted that funding for these tools is scarce.
- Investing in people is critically important for advancing data-driven science. Evaluators reported that funding individual data-driven scientists was a valuable way to advance the field. Awards from the People strategy helped grantees to take more risk and build the infrastructure of labs that are conducting data-driven research. In addition, it allowed researchers to develop and refine tools and software, strengthening the ecosystem of tools available for data-driven science. The expert panel also noted that investing in individuals advanced data-driven research at the awardees’ universities.
- There is value in having an interconnected network of strategies. Both the evaluators and expert panel found that the initiative’s three intertwined strategies are mutually reinforcing. Strategies that intersect with one another can amplify the impact of each individual strategy.
- Influencing academic culture is challenging. Evaluators found that, while the initiative had begun to foster academic environments that nurture data-driven research, challenges remained. The data science environment collaborations have nurtured data-driven scientists, demonstrated the value of data-driven science, and have provided opportunities for training in data-driven skills and methods. In addition, the Data-Driven Discovery Initiative helped individuals practicing data-driven science to advance professionally. However, attempting to change academic culture is a large undertaking, and evaluators found that the initiative had not yet produced changes in the review criteria for promotion of data-driven scientists.
- Measuring culture change is difficult. Capturing and quantifying culture change in academia is challenging. It is a slow process influenced by multiple actors. Tracking the initiative’s contribution to culture change in an academic setting relies heavily on individuals’ opinions and requires that those individuals have full knowledge of the landscape.
The evaluation was useful to the Data-Driven Discovery Initiative team in identifying successes, challenges and potential new opportunities. We hope by sharing the evaluation that academic institutions, data science researchers, software engineers and other funding entities can learn from our first six years in funding this exciting and rapidly growing field.
Message sent
Thank you for sharing.