The Strong Researcher Hypothesis

There’s a tension between two basic metascientific instincts. On the one hand, metascience promises that, although scientific and technological progress is wild, powerful, and mysterious, it can be understood deeply enough that we can craft interventions to effectively influence it. Through reforms to scientific publication, changes to granting, and other procedural tweaks, we may be able to accelerate and shape progress. These reforms are built on the idea that by shaping the incentives of researchers, we can shift their decision making calculus in ways that will produce faster progress. 

On the other hand, metascience has also lionized the lone, self-motivated researcher, toiling away to produce a scientific breakthrough in spite of the skepticism of others. Katalin Karikó has been (rightfully!) celebrated as a dogged figure in this mold. One gets the feeling that Karikó would have pursued her explorations regardless of the broader research landscape. In other words, the very researchers metascience romanticizes the most are weird characters who are apparently deeply immune to the lure of external incentives. 

Metascience therefore proposes interventions that are least likely to influence the behavior of the kinds of researchers it loves best. At the end of the day, there’s just not much you can do to direct how a Karikó will want to invest her time and mental energy. 

But the problem might go further. What if it is not just the occasional standout genius that marches to the beat of their own drum? What if researchers, by and large, have strong intrinsic motivations guiding them that are hard to shape through external interventions? 

This is the Strong Researcher Hypothesis. Simply stated, it is the presumption that researchers are mostly driven by strong intrinsic motivations. We should assume that they have high agency and that it will be very hard to change their behavior significantly through the simple modification of mechanisms like grants.

There’s a corollary to the Strong Researcher Hypothesis. If the hypothesis is true, then frequently the apparent influence of funders over researcher behavior is typically more a demonstration of researchers finding ways of manipulating funders to meet their own interests, rather than the other way around. Whether the buzzword is “artificial intelligence” or “nanotechnology,” researchers will find clever ways of clothing their work in the topic of the moment without fundamentally shifting what they were going to do anyways.

There’s some evidence to support this view. Barham, Foltz, and Melo (2021) conduct a large survey of faculty at American land grant universities in 2005 and 2015, finding that intrinsic motivations like “enjoy[ing] doing this kind of work” and “scientific curiosity” were consistently the most highly ranked factors in defining choice of research, over extrinsic motivations. Stern (1999) estimates that biologists are willing to accept a 25% discount on their wages in order to take positions where they are permitted to publish and engage with the broader scientific community. Science itself, and being able to participate in it, is clearly a good that researchers value highly. As Stern puts it, scientists pay to be scientists. 

Second, money – the external motivator par excellence – seems curiously ineffective as a way of shaping researcher decision making. Myers and Tham (2023) report a survey where theoretical increases in the size of a grant do not result in appreciable changes in research direction. In fact, “larger grants lead researchers to increase the size of ongoing projects at the expense of moving in new directions.” In reviewing NIH data, Myers (2019) similarly finds that it is very expensive to attract new applicants to a particular research program with a grant, and that the stronger factor in whether they apply is the similarity of a grant opportunity to a researcher’s existing work. 

Where does a world populated by Strong Researchers leave the metascience faithful? It does not mean that activating extrinsic motivations through mechanisms like cash are not useful at all. But it may mean that these tools are the least efficient ways of producing desired outcomes in researcher behavior and scientific progress. 

Cash and external incentives are attractive since they are relatively easy to experiment with, measure, and establish as programs. But if researchers – especially the most talented and desirable researchers – are largely motivated by intrinsic factors, then the levers for genuinely modifying their behavior are much less clear. Given “strong” enough researchers, metascience may be left with only really being able to accelerate researchers in doing the work they were going to do anyways, an approach taken by programs like FastGrants. 

Ultimately, the Strong Researcher Hypothesis demands further investigation. How do intrinsic motivations manifest into a specific research agenda? What types of different intrinsic motivations are the most powerful, and how do they vary across fields and institutions? How do these intrinsic motivations around science come to be in the first place? These are questions in the realm of the psychological, sociological, and anthropological. But metascience may need to journey there to really figure out what makes progress tick. 

---

Thanks to Matt Clancy and Kyle Myers, who directed me to a number of the papers cited. And, thanks as always to Santi Ruiz and Caleb Watney, who provided review and edits on earlier drafts of this article.