When Science Was Still Disruptive

Why is science becoming less disruptive? Michael Park, Erin Leahey, and Russell J. Funk’s recent publication sparked a lively debate in the scientific community. Many believe it is an inherent feature of the field that more disruptive findings are made at the moment of conception of new areas of study: “low-hanging fruit” breakthroughs. But the study authors argue that such hypotheses do not adequately explain their observations. Instead, they suggest several systemic problems may explain the decline in disruptiveness, like a focus on publication quantity instead of quality.

The main problems leading to a decline in “disruptive science,” in my view, are structural. Chief among those is the increasingly competitive, metrics-driven nature of academia. Although this system purports to offer objective criteria of scientific merit, it actually takes away the freedom that is necessary for disruptive science and incentivizes researchers to boost their “success scores” instead of focusing on innovative science.

Nowadays, a career like the one Smithies describes is largely unthinkable. Scientists do not switch their research focus. Rather, they tend to become narrower and narrower in their research, something Park et al. quantify. It is also almost impossible to have a scientific career without publishing important papers every step of the way.

Publish or Perish

Why do scientists today shy away from taking the liberty Smithies found so crucial for his own career? The reason it is so rare for scientists to take a sabbatical or switch fields is simple: they are ensnared in a system of brutal competition. If you take a break or don’t publish for a while, you’re out.

In an elegant article, French sociologist Christine Musselin shows how competition came to structure academic science. Competition between universities for status spirals into a rivalry fueled by the state as “competition organizer.”

Initially, the National Institute of Health (NIH) awarded funding mostly to centers or common projects (“P01 grants”). In the 1970s, this funding scheme was rapidly replaced by grants for individual researchers handed out in ever more standardized competitions (“R01 grants”). Through the mechanism of an “indirect cost rate,” part of the money individual researchers receive from these grants flow into their universities. So federal funding for universities came to depend on how well their employed researchers perform in competitions for federal grants.

In theory, contests between scientists don’t have to be a bad thing. As Musselin says, competition existed in science even when it was more disruptive. What changed was the nature of this competition between scientists. In the search for measurements universities and the state can use to rank competitors, these institutions look for objective metrics of researcher quality. It is this attempt to “objectify genius” that ultimately erodes disruptive science.

These metrics are based on researchers’ publications. Some measurements, like the “H-Index,” measure how often a scientist’s publications get cited by other scientists. Others, such as the “impact factor,” use the citation record of journals the scientist publishes in as a proxy. The “objectified” value of researchers has not only served for university rankings but has also come to determine the distribution of federal grants and faculty positions.

At first glance, the system seems like an elegant way to address a problem that was likely even worse in the past: if we attribute objective quality scores to scientists and use these, for example, to distribute faculty positions, we are less reliant on subjective decisions, which can allow for nepotism and individual prejudice to determine who advances. But the measured decline in disruptive science suggests that the system does not really work as intended. Instead, it creates incentives that are poison for innovative research.

Incentives for Senior Researchers: the “Productive Lab”

Once a career depends on a score system, researchers will seek to optimize their scores. Instead of a competition to do the best science, scientists hunt for “impact points.”

How does one become the highest scorer? First, you get a better score when you increase your output of papers. The easiest way to increase that output is to hire people whose work and brainpower allow you to produce more papers that you will get credit for.

The incentive for professors is clear: get as many subordinate workers as possible, and you will have more publications. A certain feature in the publication system ensures hiring more trainees is never detrimental: the splitting of “first” and “last” author. Professors get their currency by being last authors (the last name in the list of people publishing the paper), while the workers receive first-author credits. For researchers, “last author” means “this person is the brain of the study,” and “first author” means “this person did the hands-on work.”

The example of Smithies shows that disruptive scientists need freedom to pursue questions out of curiosity. Smithies had this freedom because his professors, at all career stages, saw him as a peer and not as an employee. In modern laboratories with professors who fully embrace the competition model in academia, young researchers are hands for hire, not peers.

Like a recent commentary in the debate around disruptive science suggests, young scientists these days focus on an “executive and results-based approach” rather than engaging in creative curiosity-driven research. In my opinion, this change in training for young researchers is not due to flawed teaching styles. Instead, it is the logical consequence of the transformation of the professor-trainee relationship, fueled by the current scheme of competition in science.

Incentives for Junior Researchers: Productivity and Specialization

The focus on “research productivity” not only shapes how senior scientists operate, but also fundamentally restricts junior scientists. These restrictions are most obvious at the transition point between trainee and professor.

To become a professor, you need to acquire “starting grants.” In life sciences in the United States, the main starting grant is the K99 from the NIH. To receive a K99 grant, you have to demonstrate your productivity. And your productivity is demonstrated by publications over time.

To measure this productivity, you need a set time frame. Junior scientists can only apply for a K99 grant during the first three and a half years of their postdoc. During this time, scientists have to demonstrate their productivity with first-author papers.

But different kinds of research are not rationally comparable in this way. Let’s say there are two researchers: one is a computational biologist who uses preexisting data for their research and the other researcher studies the effect of an aging immune system and must perform their own experiments. The computational biologist has no problem publishing in three and a half years. But for the researcher focusing on aging, each experiment takes a year. Unless they are extremely lucky, there is no way they can publish in time.

It should be obvious that time restraints like those imposed by the need to win starting grants select for a certain type of research. The researcher interested in aging will likely have to choose between pursuing their curiosity-driven research and risking their career, or pursuing a project that is “feasible” for publishing more papers quickly. Unfortunately, the most easily publishable science is likely the least disruptive. The chance of publication is best if you follow the research of your supervisor and study questions that yield predictable results.

The restrictions imposed on researchers by “feasibility” and “productivity” are not limited to starting grants: the NIH explicitly lists “feasibility” as one of the key criteria in the evaluation of all grants. Behind this decision lies a valuing of “productivity” over “creativity” in the competitive structure of academia.

The Corset of Neoliberalism Does Not Fit Academia

The incentives that come from the competition model of modern academia limit the freedom of researchers in a way that suppresses disruptive science. But how can we undo it?

A first step is to understand why academia was transformed like this in the first place. And at the core of this transformation is the neoliberalization of science. The reigning viewpoint of neoliberal capitalism says that a (supposedly) meritocratic competition is the best way to structure society and maximize economic growth. The objectification of research value is a form of the broader phenomenon of ever-expanding commodification under capitalism; the transformation of trainees into hands for hire is an instance of the alienation described by Karl Marx, in which workers are separated from the fruits of their own labor and their control over the productive process. And behind the current methods of assessing the “feasibility” of scientific research, we can find the same practices financial institutions deploy for “risk analysis” of investments.

Facing climate catastrophe and a crisis in wealth distribution should make us rethink this approach to organizing social life. But for science, the problem is obvious: the structure of a competitive marketplace is not conducive to good research in the first place.

First, objectification of scientific exploration and innovation in the way that capitalism demands is not conducive to scientific breakthroughs, because most breakthrough discoveries, by their nature, are unpredictable. For instance, when Francis Mojica began studying repetitive patterns in the DNA of bacteria, no one cared. Big journals refused to publish his findings. Today we know that this work was in fact the basis for maybe the biggest discovery in modern biology: the gene scissors CRISPR/Cas9, which is revolutionizing molecular biology and life science.

Second, the transformation of the mentor-trainee relationship from peer-to-peer to boss-and-wage-laborer also makes little sense for academia at a large scale: today’s trainees are tomorrow’s professors. Suppressing the autonomy and creativity of the trainees by turning them into wage laborers is detrimental for the future generation of professors, who then have lost their ability to think creatively and have been trained to take less risky options.

Last, if we accept that breakthroughs are unpredictable, we should understand that good science can never be “quantified” like a product. The most disruptive science likely requires much more time than other research. It also requires taking great risks — for example, scientists choosing to switch fields or study something entirely new. If we continue to measure the quality of research as “predictable productivity” and distribute resources and positions accordingly, we will miss out on much disruptive science.

Bring Back Disruption by Limiting Competition

To bring back disruptive science, we need to limit the competition scheme that has ultimately impaired our ability to conduct curiosity-driven research. A first step could be to strengthen guaranteed funding from institutions and reduce the resources that have to be acquired in grant competitions, especially for young researchers.

Moreover, attempts to “score” the value of researchers via their publication record should be dramatically scaled back. Instead, we need to embrace the fact that scientific value cannot be quantified. Decisions about faculty positions therefore must be based largely on qualitative judgments. To make sure this does not lead to nepotism or unfair discrimination, we should radically increase democratic input into institutional decision-making. Faculty hires, for instance, could be voted on by the whole faculty, and even postdocs.

Last, we need to reverse the recent transformation of the mentor-trainee relationship. Limits on the composition of research groups could help here, since most “exploitative” structures are characterized by a large number of highly qualified postdocs that stay for an extended time under the control of a single professor. And graduate student and postdoc unions are essential to empower trainees and make their concerns heard in a way the current system does not allow.

Kati Kariko’s work on mRNA vaccines was not predicted to be of any value. As a result, she was almost forced to leave academia because she couldn’t get funding or a senior faculty position. Kariko, a New York Times profile said, “needed grants to pursue ideas that seemed wild and fanciful. She did not get them, even as more mundane research was rewarded.”

Her work, of course, would turn out to be the basis for lifesaving COVID-19 vaccines. By reforming science to put curiosity-driven research back at the center, we can make sure we don’t miss out on more important discoveries like hers.