iStock.com/milos-kreckovic

To predict what research will be transformative, a crystal ball might be your best option

Last year, marine ecology postdoc Sarah Gravem was having a routine discussion with her adviser and some graduate students about the experiments they were planning when the conversation took an unexpected turn. Gravem, who studies the ecological consequences of sea star wasting disease for intertidal ecosystems at Oregon State University in Corvallis, told them that she “didn't think the experiment was ‘transformative’ enough,” she writes in an email to Science Careers. Her adviser, Bruce Menge, “bristled and said that he didn't like that word, because he thinks scientists have no way of knowing if something will be transformative before performing an experiment,” she recalls.

The conversation prompted Gravem to investigate the scientific process behind several transformative advances in ecology. The resulting paper, which Gravem co-authored with Menge as well as a number of Ph.D. students, provides quantitative evidence supporting Menge’s point. Even prominent scientists typically did not anticipate that their research would be transformative—defined as “radically changing our understanding of a concept, causing a paradigm shift, or opening new frontiers”—at the time they conceived it.

Gravem and her co-authors identified six ecologists whose studies they considered transformative and asked them whether they thought it was possible for researchers to deliberately conduct transformative work. They generally believed it wasn’t. Most of the researchers also said that the studies they had initially set out to pursue were conceived of as incremental, and that they only recognized the groundbreaking potential of their research later in the process. Five of the six hadn’t received dedicated grant money for the work, instead relying on peripheral sources of funding, such as postdoctoral money or university travel funds, and sometimes having no funding at all.

These findings were echoed in a survey of 72 other ecologists with highly cited work. Only 6% of the respondents said that they realized the transformative nature of what would later become their most highly cited paper while preparing the proposal. The numbers gradually increased over the course of conducting the research and publishing: 10% recognized the impact of the work while collecting data, 18% while analyzing the data, and 28% during the writing and publication process. Another 19% didn’t recognize the impact of the work until after the paper was published. (The final 19% didn’t identify any of these options as their lightbulb moment, with some respondents saying that it occurred while talking to a colleague or brainstorming and a few others declining to see their paper as transformative.) Overall, approximately 47% of the respondents believed that, at the proposal stage, researchers cannot predict whether their work will be transformative, versus 36% who believed it may be possible. Transformative research “is essentially impossible to plan for,” Gravem concludes in her email.

These findings are in line with previous studies in this area, notes Roberta Sinatra, an assistant professor at the Central European University in Budapest who has conducted some of the work. For example, she and her co-authors have found that each paper a researcher publishes is equally likely to become their most-cited work. Others have shown that, although most papers accrue citations in a standard pattern, many major scientific advances deviate from the norm.

The ability—or not—to predict what is going to be transformative may have important implications for early-career scientists looking to establish themselves while competing for scarce funding. These days, the authors write in the paper, “[p]rioritization of [transformative research] has become pervasive among granting agencies.” The U.S. National Science Foundation, in particular, has made transformative statements a ubiquitous requirement, Gravem argues, with the U.S. National Institutes of Health, the Natural Sciences and Engineering Research Council of Canada, and the European Research Council also asking for transformative statements in targeted calls, she adds. But their definition is so lofty that “this is a problem because it sets up a straw man for grant reviewers to strike down essentially any project proposal.” To juggle the funding pressure with the realities of research, Gravem recommends that early-career researchers pursue multiple projects so that they can balance a range of likely risks and rewards. “The lower risk, incremental projects may end up producing high rewards, because that’s how most [transformative] researchers ended up getting there anyway,” she writes.

Sinatra, however, disagrees that there is a funding push for transformative research. “There are other works … that actually show the opposite, that transformative research is somehow less likely to be funded,” says Sinatra, warning of “a disconnect between the narrative [in] a call and what is funded among the proposals.” She and others suggest that, given the current system, a safer choice for young scientists is to start with more incremental research proposals because they are more likely to be funded. Then, once a researcher is established and tenured, their existing funding and reputation will make it easier and less risky for them to explore new ground, she adds.

Daniel Nettle, a behavioral scientist and professor at Newcastle University in the United Kingdom, recently wrote about how he stayed in the research game by steadily pushing his research forward rather than planning big and risky leaps ahead. As he sees it, if there is a funding trend in the United Kingdom, it is toward research with societal impact, which “tends to militate against transformative basic thinking,” he writes in an email to Science Careers. “But I certainly agree that funding is very competitive, and of course that means people have to make transformative claims,” he adds. Eventually, though, “given that it is hard to predict what will turn out to be important, the main thing you can do is to do what interests you, do it as well as you possibly can, be honest about it, don’t believe your own hype, and communicate what you have done clearly,” Nettle writes. “Then, it’s up to history what happens next.” 

Follow Science Careers

Search Jobs

Enter keywords, locations or job types to start searching for your new science career.

Top articles in Careers