Question: How should researchers be rewarded for data sharing and reproducible research?
The current reward system in science is based on publishing a good amount of papers in high impact factor journals, which will ultimately lead to acquisition of further research funding. Here, speed and novelty are greatly rewarded. The public availability and thorough documentation of generated data should be the basis of high-quality and reproducible research. According to recent polls, 70% of conducted studies cannot be reproduced by scientists, a term coined as the “reproducibility crisis”. It is becoming increasingly clear that the scientific community needs to come up with solutions to address these arising problems. Researchers that invest time and effort in generating reproducible high-quality science should be met with certain rewards and incentives. As a starting point, generating publicly available datasets that meet the TOP guidelines (1) should be a requirement for the publication of research findings in scientific journals. In order to decrease the amount of labor invested by scientists, some publishing houses like Springer Nature have launched a Research Data Support Service, where data and metadata is organized as to meet data sharing requirements and standards. This would ultimately lead to a significant increase in the reusability of published data and in the number of citations of a particular paper. This time-saving incentive could be easily taken over by other journals, making it a standard publishing practice. Furthermore, funding agencies, research institutes or universities could designate one part of their research funds for these specific purposes. In fields such as machine learning, datasets are published as research papers with some of them achieving as many as 3300 citations. Bearing this is mind, creating a journal dedicated to the publication of high-quality datasets would surely draw the interest of many scientists, who could eventually end up publishing several extensively cited papers from one research project. The journal would also benefit from the numerous citations by earning a high-impact factor. Another measure to increase data sharing could be granting financial rewards to researchers that consistently produce high-quality publicly available datasets or that re-utilize them to answer further research questions. This is being carried out by institutions such as the Berlin Institute of Health (BIH) or the whimsically termed “Symbiont” and “Parasite” awards. Regarding the reproducibility crisis, very much like the implementation of a female quota in academia, a “Research Replicability Quota” could be adopted by many scientific research journals, where 5-10% of the journal would be dedicated to the publication of research validations. This way, scientists that invest time and funds in the reproduction of published studies would be rewarded by high impact publications. In conclusion, the scientific community needs to find solutions that improve research transparency, data sharing and reproducibility. In the past years, awareness of the current problematic has been raised and some action has been taken by research institutions and scientific journals. We can only hope that these efforts continue and that, eventually, investigators will make the generation of high-quality research a top priority.