The Few & the Many: Why Big Science can be Detrimental
A Love Letter to Single-/Double-Author Publications
More, and more, and more. Not only did the number of scientific articles skyrocket in the last few years, the number of authors per paper also steadily increased over time, making single- or double-author papers as rare as the proverbial hen’s teeth. Instead, we’ve got 15-author papers, with three shared-first authors, two shared-seventh authors (I bet you there’s always a juicy story behind those), and authors that are simultaneously shared first and last. Often, this is seen as progress. Interdisciplinarity par excellence, demonstrating team spirit and the ineluctable journey of science into tackling tough questions requiring different approaches. Others critique the rising tide of authorship as padding, politics, or a product of the publish-or-perish mentality. Discussing the origin(s) of this phenomenon is a story for another time though. Here, I want to dissect the perils of Big Science; of large, multi-institute collaborations resulting in academic publications inundated with authors. Perils that could endanger both regular as well as groundbreaking research.
Let me start by saying that I decidedly do not think long author lists are always deplorable. There is a time and place for everything, even for a project that truly requires the collaboration of more than a dozen people. Some disciplines, such as particle physics, simply have such an exorbitant appetite for funds, equipment, and gigantic experiments that certain questions cannot be conclusively answered by the few. My point is just that author lists, on average, are currently inflated for various reasons, with potentially detrimental consequences for science. While there are many implications stemming from this change, I’ll focus on aspects from project development, science communication, and the intellectual enrichment of the academic community.
Project development is relevant for two reasons: (i) choosing an interesting topic/question and (ii) combining insights from multiple disciplines. Similar to the law of large numbers, project ideas developed by more and more people will regress to the mean, in the sense that only reasonable-sounding ideas will be pursued. Imagine if one of the involved researchers proposes an unlikely-sounding, ambitious approach. Even if only a fraction of their colleagues dismisses the idea, its chances of survival are dim. Ideas are fragile. Contrast that with the single- or double-author study, in which the suggestion of an individual carries far more weight. What’s more, the contributions of each author are also considerably clearer. Thus, similar to small sample sizes, small author pools enable more variance and therefore could have a higher chance of deviating from the average, of producing groundbreaking research rather than “normal science” in the sense of Thomas Kuhn. (As a side note, this increase in variance presumably also could lead to erroneous or subpar articles.) This is especially important for research areas shaped by dogma, such as research on Alzheimer’s disease plaques, strongly entrenched as the centerpiece of this disease by dogma, in which contrarian teams have to start small because nearly everyone’s against them; so the larger the gathering of researchers in that area, the higher the chances of including actively policing proponents of the current dogma. So, next to inefficiencies and the administrative costs of coordinating multiple institutions, Big Science is mostly limited to obvious, next-step kind of projects. All the worse then that ever more funding institutions require collaborative efforts via specialized funding structures, sometimes even of the multi-institution kind, regardless of their necessity.
While the mix of academic specialties in large collaborations seems enticing on paper — constructing a unique and synergistic path toward a solution — communication barriers often impede fruitful collaboration. Different disciplines speak different languages. This frequently presents an insurmountable obstacle, to the extent that individuals familiar with multiple disciplines (i.e., generalists) are more creative than groups of specialists. Combining disciplines — one of the key supposed advantages of large collaborations — thus seems to be easier for individuals than for groups and we might, paradoxically, impair this creativity with a focus on large collaborations. Many trailblazers switched scientific fields in their careers and made their most important and creative contributions shortly thereafter, in what has been termed the novice effect. By taking into consideration that these researchers now combined knowledge of multiple specialties in one person, this phenomenon is no longer that surprising. Contrast this with the conventional case, in which researchers from both disciplines are not only without connection and a connecting language but also have to be convinced of the merits of each other’s discipline prior to a collaboration.
The act of writing a scientific paper is, fundamentally, an act of science communication. And it suffers as well from too many cooks spoiling the broth. The first thing a manuscript with too many researchers wielding the editing pen loses is whimsy. Colorful language, innuendos, wit. A small sacrifice you might think. Maybe even a boon, endowing the resulting piece with rigor and objectivity. Yet also making it less approachable, less relatable, and ultimately less representative of what the scientific process encompasses, fostering unrealistic expectations and evaluations. The injection of personality into the writing, on the other hand, reflects the subjectivity inherent in research. With the fake-objectivity-inducing passive voice slowly but finally vanishing from scientific writing, I sincerely hope this artificial removal of personality in writing will soon follow.
Finally, large consortia not only affect how something is written but also what is being written in the first place. Analogous to the dismissal of daring ideas or approaches during project development and execution, speculations and concessions of method limitations are culled one by one as the manuscript is being “smoothened.” Smoothened until it doesn’t contain any rough edges that might elicit contention, smoothened until it is fundamentally bland. I posit that publications with fewer authors thus have the potential to incorporate an intellectually richer discussion and to inspire other researchers by sharing their interpretations and speculations, without having to navigate the consensus. It is fascinating to me that we cheer for the few in the form of academic talks but root for the many when describing the very same work in the written form.
Small author pools are not necessarily better but they are certainly different. So, if we truly care about diversity, we should view the ever-increasing trend toward large author pools, with all their perils, with care, especially as there is no naturally ordained progression in this direction. While some experimental questions might require unprecedentedly large teams, the plethora of publicly available data has facilitated work on important questions by much smaller teams. I am concerned that nowadays we merely tolerate single- or double-author publications, with the firm expectation that they will all but vanish in due course. Yet I’m convinced we should do more than just tolerate them; we should cherish them and all the whimsical variance they bring to science. Because small can be beautiful.
