Recently, I was contacted by a Program Director from the National Academies of Sciences, Engineering, and Medicine (NASEM) and asked the following question:

If you could make one change that would shift incentives in research in favor of open science practices in research, what would it be?

The purpose was to provide input for an upcoming roundtable on open science NASEM plans to host. By replying to the email, I could hopefully make a small impact towards accelerating science through improved openness.

While I generally provide very specific recommendations to make science more open — such as open licensing of data and code and performing research publicly in realtime — I decided to go with a more general recommendation that is relevant for a diverse range of scholarly disciplines.

Here's what I replied:

When evaluating scientists, the primary factor should be how their past work has contributed to the research of others. Specifically, scientists should be asked "how have others used the outputs of your research?" When scientists are incentived to provide useful outputs to the community, they will gravitate towards openness because open science maximizes one's contribution.

For example, the biologist might say, "we created a gene editing technology, which we declined to patent, such that hundreds of labs and companies now freely use it in their research." A geneticist might say, "we created an openly licensed dataset which is now incorporated into 10 integrative resources and genome browsers created by third parties." A psychologist might say, "we published our research on healthy eating in an open access journal and the article has been downloaded over 1 million times and translated into Spanish, Mandarin, and Arabic." A data scientist might say, "we created a software package that has 200 stars on GitHub and is used by over 50 studies a year."

Publishing in a high impact journal is not, in and of itself, a contribution to science at all. Studies should not be judged by the journal which published them, but rather a verifiable record (beyond just citations) that the outputs of that study are driving the next generation of research. When the incentives to provide open outputs become larger than the incentives to keep outputs proprietary, science will shift towards openness. Funders & institutions must reward and promote those who have a history of providing open & useful outputs versus those who have a history of not sharing their research to the maximum extent possible.

Let me know in the comments if you agree, disagree, or have other ideas on what we can do to make science more open!