The impact of switching research fields

When Karissa Sanbonmatsu joined the Los Alamos National Laboratory (LANL) as a postdoctoral researcher in 1997, she was eager to apply her plasma physics training to problems in space physics and laser fusion. Shortly after becoming a principal investigator, however, she became intrigued by the life sciences and decided to shift to studying biophysics. “I made it happen by leveraging transferable skill sets,” she says. Now she enjoys success as a structural biologist, combining wet lab biochemistry with computational biophysics and cryoelectron microscopy. “When you switch fields, you bring a fresh perspective that people haven’t heard about before,” she says.

For Sanbonmatsu, switching research fields led her to make meaningful contributions in new territory. And she’s not the only researcher who has increased her scientific impact after making a shift. Physics researchers who enter and stick with a new field are likely to produce papers that are more highly cited than their earlier works, according to a new analysis of the publication records of more than 14 000 authors of physics papers. The findings, published by Tao Jia of Southwest University in China and his colleagues in the August issue of the Journal of Informetrics, provide reassurance for early-career researchers who are considering establishing a new direction.

The paper “suggests an intrinsic link between research topic shift and scientific impact” that has not been identified before, says Junming Huang of Princeton University, who studies causal inference, science of science, statistics analysis, and social networks.

The study isn’t the first to use bibliometric data to analyze how topic changes and other factors influence scientific fields and the careers of the researchers who choose those fields. According to Jia, however, many studies tend to draw on comparisons between different groups of scientists to identify features that define a successful career. For example, they investigate traits shared by highly cited researchers or Nobel Prize winners. “Success stories can tell us a lot,” Jia says. “But what can be more informative is the comparison between current you and future you.”

To quantify the risks and rewards associated with changing research fields, Jia’s team used the 67 topic codes—such as for physical chemistry, quantum fluids and solids, or electron transport in condensed matter—that were assigned to papers published in the Physical Review journals from 1975 to 2010. For each author in the American Physical Society’s database of 300 000 papers published within that time frame, Jia and his colleagues collected two sets of papers corresponding to the author’s first n papers and last n papers for a series of different values of n up to half the author’s total publication count.

Within each set, they first composed a vector that represented how many times each topic occurred in the author’s publication sequence. Then, by calculating the degree of similarity between the two vectors, they quantified the extent of topic change between the earlier and the later set of papers. Finally, they measured whether an individual’s later papers received more citations than the earlier set and how quickly those citations were earned.

By grouping scientists who had undergone a similar extent of research topic change, Jia and colleagues compared how citation rates (which they used as their measure of impact) changed within each group. They found that for scientists who undertook a topic change, a larger difference between the earlier- and later-career fields of study was linked with the production of higher-impact work, even though there was no significant correlation with the number of papers produced each year. “The patterns we discovered could apply to a wide range of scientists, not just a few top elite ones,” says Jia.

Huang calls the study novel and insightful. However, it does not establish a causal link. Indeed, the quantitative approach purposefully excludes confounding factors such as whether the new field is hot or cold, whether a new research program is broad or narrow, and whether an individual tends to jump between topics. Those factors are important in real life. Jia and his colleagues also note that although their analysis highlights opportunity over risk in leaving one field to pursue another, it counts only the people who survived that change.

In addition, publication records and citation metrics do not reflect the entire picture of research impact. For example, Harry Miley, a senior scientist at Pacific Northwest National Laboratory who works on nuclear explosion monitoring, says that “a single dark matter paper that I contributed to more than 10 years ago drove my citations wild.” But he believes he’s doing his best work now with respect to societal impact.

A researcher’s institution could make a difference too, according to Andrew Sornborger, who leads the Quantum Science Center at LANL but spent his formative years in theoretical physics and computational neuroscience. “It was easiest to publish in high-impact journals when I was a postdoc at Cambridge University and with Fermilab. When I switched to computational neuroscience, my impact decreased and I struggled to publish,” he says. “Now, if you look at my citations since I moved to a major national laboratory, the difference is huge.” He cites a combination of rich research topics, work in teams, and a pure focus on research without teaching as helping to boost his impact at LANL.

An Zeng of Beijing Normal University, whose work has found that scientists with a narrow research agenda tend to secure a steady publication output and more citations, says that “the mechanism behind those correlations requires further investigation.” For example, do scientists have exciting new ideas in other fields before they change research direction and then accordingly produce high-impact work in the new field? Or do they decide to change direction first and then seek out new opportunities with high potential?

Researchers need to be careful when switching topics, cautions Zeng. Maria Molina, a project scientist at the National Center for Atmospheric Research, felt that need for caution when a firsthand experience with extreme weather events persuaded her to pursue a shift toward climate dynamics shortly after completing her PhD in convective storms. “As scientists we think changing direction might cause a slowdown in productivity because we have to learn new concepts,” she says. To her, Jia’s work provides encouraging evidence that a direction change can lead to greater success. So far, she’s had two papers accepted in areas that are new to her and has found satisfaction in expanding her knowledge base.

Jia says that future quantitative studies could analyze unsuccessful changes and the traits of individuals who choose to make a change and subsequently founder or flourish. Rebecca Holmes Sandoval, a LANL researcher whose work in single-photon detection has shifted her publication record from papers in quantum optics to ones in aerospace and space science, agrees that it’s important to pinpoint reasons for the career trends that can be teased out of big data sets. Doing so could help young scientists achieve their goals by guiding career development. But, she says, “don’t switch fields for the sake of publications. Switch because it’s fun!”