Great scientific discoveries are increasingly rare, study finds

Miracle vaccines. Videophones in our pockets. Reusable rockets. Our technological bounty and the scientific progress related to it seem undeniable and unparalleled. But an article published by the journal Nature points out that the general pace of major advances has slowed dramatically over the last nearly three quarters of a century.

The researchers responsible for the article analyzed millions of patents and scientific articles and showed that researchers and inventors made relatively few advances and important innovations, considering the growing mountain of scientific and technological research carried out around the world.

The three analysts found between 1945 and 2010 a steady drop in disruptive discoveries as part of the increase in research, suggesting that scientists today tend to move forward incrementally rather than making intellectual leaps.

“We should be living in a golden age of new discoveries and innovations”, says Michael Park, one of the authors of the article and a doctoral student in entrepreneurship and strategic management at the University of Minnesota, in the USA.

Park and his colleagues’ findings suggest that investments in science are spiraling into ever-diminishing returns and that, in some ways, quantity is trumping quality.

Although this is not addressed in the study, the findings also raise questions about the extent to which science is in a position to open new frontiers and sustain the kind of daring that unlocked the atom and the universe and what can be done to combat the growing distance. between science and pioneering discoveries. Previous studies have pointed to a slowdown in scientific progress, but in most cases they have done so with less rigor.

Park, Russel J. Funk, also at the University of Minnesota, and Erin Leahey, a sociologist at the University of Arizona (USA), based their study on the kind of augmented citation analysis that Funk helped to create. Citation analysis generally tracks how researchers cite each other’s published work as a way to separate bright ideas from less than exceptional ones in a flooded system of scientific papers. The improved method they used broadens the analytical scope.

“It’s a very smart metric”, evaluates Pierre Azoulay, professor of technological innovation, entrepreneurship and strategic management at MIT (Massachusetts Institute of Technology), in the USA. “I was blown away when I saw it. It’s like a new toy.”

It is not new that researchers are looking for objective ways to assess the state of science, seen as crucial for economic growth, national pride and military strength. It became more difficult to do so as scientific articles published increased in number, reaching over 1 million each year. There are more than 3,000 articles each day — by any measure, an indecipherable blur.

Contesting this increase, experts have debated the value of incremental advances versus the “aha” moments that change everything known about any given field.

The new study may deepen this debate. One surprise is that findings popularly hailed as major advances are seen by the authors of the new study as, in many cases, representing little more than routine science. The authors consider that true jumps are sometimes completely absent from the discussion.

For example, the biggest advance cited in the list of examples drawn up by the study is in the recombination of genes, still little known by popular science. This discovery allowed foreign DNA to be inserted into human and animal cells, not just bacterial cells.

The New York Times mentioned this discovery in 1983, in a note of just four paragraphs. But the feat yielded a series of awards to its authors and to their institution, Columbia University, in the USA, in addition to almost US$ 1 billion (about R$ 5 billion, in current quotations) in licensing fees, since it has enabled a breakthrough in biotechnology operations around the world.

In contrast, analysts would view two of the most celebrated discoveries of this century as triumphs of ordinary science rather than bold leaps. The mRNA (messenger RNA) vaccines that successfully fight the coronavirus are the result of decades of low-key work, the authors point out.

And the 2015 observation of gravitational waves — subtle ripples in the fabric of space-time — wasn’t exactly a breakthrough: it was confirmation of a centuries-old theory that required decades of hard work, testing and sensor development.

“Disruptive advances are good”, says scientist Dashun Wang, from Northwestern University (USA), who used the new analytical technique in a study in 2019. “We want novelty. But we also want everyday science.”

The three analysts found the trend of incremental advances when they were using the enhanced form of citation analysis to examine nearly 50 million scientific articles and patents published between 1945 and 2010. They examined four categories—life sciences and biomedicine, physical sciences, technology, and social sciences—and found a steady drop in discoveries they described as “disruptive.” “Our results suggest that the slower pace of disruption may reflect a fundamental change in the nature of science and technology,” they wrote.

Its innovative method—and citation analysis more generally—gains analytical power because of the requirement that scientists cite studies that helped shape their published findings. From the 1950s, analysts began to count these citations as a way to identify important research. It was a kind of measurement of scientific applause.

But counting could be misleading. Some authors cited their own works with some frequency. And the big names in science could receive large numbers of citations for unremarkable discoveries. Worst of all, some of the most cited papers involved only minuscule improvements to techniques widely used by the scientific community.

The new method examines citations in greater depth, more effectively separating routine work from real breakthroughs. It counts not just citations that appear in research analyzed, but in the previous studies that it cites.

What was found was that prior work is cited much more frequently if the discovery is routine rather than highly innovative. The analytical method converts this difference into a new prism through which to look at the scientific endeavor.

The metric was named the CD index due to its scale, which ranges from consolidation to disruption of the existing knowledge set.

Funk, who helped devise the CD index, says the new study was so computationally intensive that the team had to use supercomputers at times to process the millions of data sets. “It took about a month,” he says. “This kind of thing wouldn’t have been possible ten years ago. It’s just now becoming within reach.”