Scientific Progress: More Evolutionary than Revolutionary

“Mr. Watson. Come here. I want to see you,” Alexander Bell said, and with those short sentences, Eureka! The telephone was born.

For anyone who holds dear the idea of such thunder-claps of invention, the stories in which science kicks over its traces to shout a revolutionary Aha!, the results of a study published Oct. 30 in the journal Nature may feel a bit disappointing. The study, led by the director of Stanford’s Meta-Research Innovation Center John P.A. Ioannidis and co-authored by Aaron A. Sorensen, Director of Informatics, Temple University School of Medicine, among others, suggests that most important scientific advances are evolutionary, not revolutionary. That is, the thunderclaps are small and occur only after a long, uphill accumulation of other discoveries.

To reach this conclusion, the authors created a list of 400 biomedical scientists whose work was most often cited in other publications from 1996 to 2011. That is, they identified 400 scientists whose articles were mentioned most often by other scientists in support of their own work. Once identified, these most-cited authors were asked to rate their 10 most-cited papers on a scale of 0 to 100 on a series of measures. Three of the measures -- continuous progress, broader interest, and greater synthesis – suggested the work in question was more evolutionary than revolutionary. The other three measures -- disruptive inventiveness, surprise, and publication difficulty -- correlate with paradigm -- shifting research. Considering the prestige of the scientists being surveyed – authors of the most influential papers over 15 years – one might expect lots of eureka moments.

In fact, it was just the opposite. Of the 123 researchers who responded to the survey and scored a total of 1,214 of their own papers, authors, on average, interpreted their papers as evolutionary, giving lower scores for things like surprise and innovation.

“To our knowledge, no one has ever systematically surveyed a large group of highly cited biomedical investigators and asked them what they thought about some of their most cited work,” Sorensen said. “That by itself is a treasure trove of information.”

But it may upset some dearly held notions about science. “People have this romantic idea of science as a series of breakthrough moments, like the discovery of penicillin or the structure of DNA. But most of science is incremental,” Sorensen said. “Even for highly successful scientists, 95 percent of the projects they spend their time on will end up being incremental, and our study confirmed that.”

While some may be surprised by the results, Ioannidis said that when it comes to science, patience is a necessity. “Careful synthesis of evidence, looking at the broader picture, and step-wise progress are probably more important and influential in science than disruption and surprise,” he said.

Twenty of the authors, 16 percent of the total surveyed, actually thought that their most important paper was not among their most cited papers. But in a sense, these outliers helped demonstrate how important citations are in determining influence.

“Since only 16 percent said their most important paper fell out of the top 10 most cited, that’s a fairly good confirmation that citations are a good indication of what’s important and what’s not,” Sorensen said.

Before anyone jumps to the conclusion that the inventor of the telephone was some kind of outlier in science’s evolutionary advance, it may be worth noting what Bell said on the matter: “The most successful men in the end are those whose success is the result of steady accretion.”