The ‘nostalgia effect’: Scientists produce less disruptive work as they age

You probably know that Einstein changed the face of physics with his theory of relativity in his twenties. What you may not know is that he spent his later career on a crusade against quantum mechanics, the model that would go on to drive the next century of advances in the field.  

New research from the University of Pittsburgh and University of Chicago shows that Einstein was far from alone in his shift from a revolutionary to a gatekeeper. Older scientists produce fewer disruptive new ideas than younger ones, the study shows, creating a “nostalgia effect” that influences scientific advancement on the scale of whole fields and even nations. 

“There's a linkage between memory and innovation — how you tie yourself to the past leads to how you position yourself in the current day,” said Lingfei Wu, an assistant professor in Pitt’s School of Computing and Information and a senior author of the study. “It’s not just about science, it's about humanity.”

The idea that science advances “one funeral at a time” dates back to 1950. Aging researchers, the adage implies, cling to old ideas and prevent new ones from taking root. Yet, even three quarters of a century after physicist Max Planck introduced the idea, the evidence remains mixed. One limitation of this body of research, according to Wu, is that it failed to recognize the different types of creativity that are necessary to do strong science. 

Wu's lab has spent years studying the factors that affect innovation, from the size and structure of teams to the ways in which they collaborate. Those efforts have given the team methods to quantify the different types of impact new research can have. In their new study, they analyzed the careers of 12.5 million scientists and looked at patterns of how they cited research and were cited by others. The team considered “disruptive” papers to be ones that subsequent studies cite without citing older research in the field, and “novel” papers to be ones that cite older pieces of research in new combinations.

As scientists age, the team showed, the papers they produce tend more toward novelty than disruptiveness — that is, they grow less likely to brew up wholly new ideas but more likely to create new links between existing ones. Not only that, but the research they cite grows older as their career progresses. For each year a researcher ages, the average paper they cite ages by around one month.

The team published their research today in the journal Science. 

Even late in one’s career, the researchers show, the ideas encountered as a young scientist loom large: The paper they choose to cite more than any other was typically published as they were just starting out, two years before the publication of their own first paper.

“You stick to a certain kind of idea or taste, and as time goes by you keep sticking to that. We see this happen again and again,” Wu said. 

And the influence of these researchers is passed down, both through their leadership of labs and as they review the papers of younger scientists. By looking at patterns within research groups and comparing citations in preprints with those in the final published papers, Wu and colleagues showed that older researchers have the effect of pushing their younger colleagues to cite older research, transmitting the nostalgia effect through academic hierarchies and the scientific review process.

Wu himself isn’t immune: Having been steeped in network theory in grad school, he now finds himself recommending two-decade-old papers to his students. This phenomenon is an important one — a way that fields maintain institutional memory.

“Aging people are not less creative, they are just creative differently,” Wu said. “They tend to recombine things, because as you age, you know more things.”

But there needs to be balance, Wu says. The research has particular importance now, the authors write, as changes to mandatory retirement policies and long training periods have led to what they call an "aging core" of scientific research in the U.S.

And these effects hold true when looking at a much broader scale. Countries with more young researchers like China and India produce more disruptive research than those, like the U.S., with an older research community, the team found. This makes the nostalgia effect a matter not just of scientific importance, but of national competitiveness, according to Wu. 

“You need to welcome global young talent. A society needs to be open to exchange students, open to international students, open to scientific talent and immigration,” Wu said. “They don't have certain kinds of knowledge, they don't have certain kinds of attachment. But this is their advantage.”

There are other ways to maintain a balance between research that breaks new ground and connects old ideas. The paper, for instance, suggests encouraging “intergenerational, flat collaborations” and valuing the skills of younger researchers who bring with them perspectives that can push science forward.

“Science needs both continuity and renewal,” Wu said. “We need to keep the continuity of science and these classic papers, and also welcome new ideas that challenge them.”
 
Other authors include Yiling Lin of the University of Pittsburgh, James Evans of the University of Chicago and Haochuan Cui, now at Nanjing Normal University in China.