UK MPs are encouraged to pay attention to the role of software when preparing a report on reproducibility in the science and technology industry, which adds around £ 36 billion to the economy.
According to the joint Software Sustainability Institute academic group, about 69% of research is done with specialized software, which can range from short scripts to solve a specific problem, to complex spreadsheets analyzing the collected data, to complex spreadsheets analyzing the collected data. millions of underlying lines of code. the large hadron collider and the square kilometer array.
“With many studies, research published without the underlying software used to produce the results is unverifiable,” the institute said in its submission to the parliamentary committee’s research reproducibility and integrity inquiry. science and technology.
The institute said a 2014 workshop found that research software “is rarely subject to the same scientific rigor as applied to more conventional experimental devices.”
The committee comes from published 86 similar evidence tackle this large and difficult problem.
The reproducibility crisis arose in 2005 when Professor John Ioannidis of the Stanford School of Medicine published an article titled: “Why Most Published Research Results Are Wrong”.
Since then, the problem has arisen in a number of studies demonstrating the prevalence of non-reproducible data, the committee said. UK Research and Innovation (UKRI), the non-ministerial public body funded by the Department of Business, Energy and Industrial Strategy, is setting up a national research integrity committee.
However, the specific issue of reproducible research has so far been overlooked, the Science and Technology Committee said.
Others wonder if “crisis” is the right word. King’s College London said in its submission: “We believe that all stakeholders have a responsibility to use sober and precise language to discuss these challenges and consider whether the ‘crisis’ threatens to generalize a complex set of problems and may allow misrepresentation in the media which has the potential to unnecessarily damage public confidence.
“Referring to nonreproducible research results as a ‘crisis’ can also imply that it is an acute problem that can be quickly resolved, rather than a feature embedded in our current research culture. “
Crisis or not, the problem has many different facets, from medicine to neuropsychology, and even battery performance research, according to the submissions.
None of this should prevent the underlying role of software from being ignored, argued the Software Sustainability Institute.
“To enable systemic change to improve the reproducibility and integrity of research, the quality and transparency of software must be improved,” he said in his submission.
“Research software engineers have a key role to play in making the software used in research more robust and reusable, and in helping to train researchers in the fundamentals of publishing code so that others can revise it and read it. ‘inspect.”
According to its website, the Software Sustainability Institute has facilitated the advancement of software in research by cultivating better and more sustainable research software to enable world-class research since 2010. It has already received funding from the Seven Research Councils. and its mission is to become the global hub for the practice of research software.
The Science and Technology Committee will be holding oral sessions to collect testimony starting in December and a report will follow. Then we will see if the software is duly taken into account. ®