Beyond open big data: addressing unreliable research

Abstract

The National Institute of Health invests US $30.9 billion annually in medical research. However, the subsequent impact of this research output on society and the economy is amplified dramatically as a result of the actual medical treatments, biomedical innovations, and various commercial enterprises that emanate from and depend on these findings. It is therefore a great concern to discover that much of published research is unreliable. We propose extending the open data concept to the culture of the scientific research community. By dialing down unproductive features of secrecy and competition, while ramping up cooperation and transparency, we make a case that what is published would then be less susceptible to the sometimes corrupting and confounding pressures to be first or journalistically attractive, which can compromise the more fundamental need to be robustly correct.

Keywords: collaborative learning; knowledge discovery; open data; peer review; research culture; unreliable research.

Figure 1

Visualization of NYC Open Data. Figure courtesy of Yuan Lai.

Figure 2

Graphic illustrating how to address unreliable research. Figure courtesy of Kai-ou Tang.