In the world of science, it’s known as the “replication crisis.”
For at least the last two decades, scientists have become increasingly aware that the results of many studies can’t be reproduced, which means they’re likely flawed.
In a 2016 survey of 1,500 scientists in the journal Nature, 70 per cent — mostly in medicine and psychology — reported that they had been unable to replicate at least one of their peers’ studies. A study in 2021 showed that studies with unreproducible results were cited more than those that could be replicated, with the authors noting that standout papers tend to attract more funding and media coverage.
Another report found that 90 per cent of cancer biology studies could not be reproduced.
To combat the diminishing confidence in scientific findings, evolutionary biologist Shinichi Nakagawa of the University of New South Wales in Australia has been keeping a keen eye on the big picture in his field. He applies what’s called meta-science — or the science of science — to quickly growing mountains of scientific data, using artificial intelligence and statistical analysis to weed out flawed research.
As a global leader in the practice, he now brings his expertise to the University of Alberta’s Faculty of Science as the newly named Canada Excellence Research Chair in Open Science and Synthesis in Ecology and Evolution. The chairs are considered one of the most prestigious and generous programs in the world, awarding leading researchers and their teams up to $8 million over eight years to establish internationally recognized knowledge centres.
Nakagawa will head the first systematic, large-scale meta-science project in ecology and evolution by building diverse and global collaboration networks. The project will aim to continuously synthesize and update large bodies of evidence while improving reporting guidelines and best practices in science research.
The overall goal of his chair, according to the project proposal, is to “make methods, data and inferences in ecology and evolution more accessible, transparent, reliable and rigorous.”
Crucial to his approach is the promotion of “open science” — communities of researchers sharing methods and results to make their collective evidence more robust and credible.
In addition to using artificial intelligence to analyze the millions of research articles published every year, Nakagawa and his team plan to generate online resources — as well as hold workshops and conduct outreach activities such as school visits — to make the science of ecology and evolution more accessible to the public.
His program will also aim to train more than 250 early-career researchers in highly sought-after quantitative data and synthesis skills.
“The U of A has one of the best computing science departments in the country, but is also very strong in ecology and evolutionary biology,” says Nakagawa. “That’s quite rare, and it makes this a unique opportunity.”
The replication crisis has long been recognized in medicine and psychology, he says, but has taken longer for evolutionary biologists to accept. As a doctoral student studying the parental behaviour of house sparrows in the U.K., he couldn’t help but notice the poor quality of reporting: “I saw opportunities to write something about it.”
That was the beginning of his mission in meta-science.
“The whole eco-evolutionary dynamic is changing, and we need to understand the forces that are changing the selection pressure,” he says. To do that properly takes a lot of researchers sharing and analyzing quantities of data that are growing exponentially.
Nakagawa’s purview extends far beyond the boundaries of his discipline, however. At his current Interdisciplinary Ecology and Evolution Laboratory in New South Wales, he and his team have been collaborating with researchers in statistical and computational sciences as well as medical and social sciences to exchange expertise: “They can teach us how to use those most advanced techniques.”
His work as a Canada Excellence Research Chair is expected to transform the fields of ecology, evolution and beyond, enabling researchers to identify trustworthy evidence and rebuild research where evidence has been flawed or weak.
“Open science is changing how scientists conduct and share research, making research methodology and outputs more open and transparent,” says computing science professor Eleni Stroulia, vice-dean of the Faculty of Science and an expert in applying AI to improve or automate industrial systems.
“Dr. Nakagawa will amplify our current strengths in this area, leading the establishment of an international knowledge hub for evolutionary ecology and open science at the U of A.”
Nakagawa joins two other new recipients of the prestigious national chairs at the U of A: Kalyan Das, Canada Excellence Research Chair in Antiviral Drug Design, and Rebecca Hull-Meichle, Canada Excellence Research Chair in the Islet Microenvironment.
The U of A now has seven current or past Canada Excellence Research Chairs, including virologist Michael Houghton, who shared the Nobel Prize in Physiology or Medicine in 2020 after coming to the U of A as one of the inaugural chairs in 2010.