The Invisible Hand of Citations: How Metrics Reshaped Scientific Careers
In the early 1960s, a quiet revolution reshaped scientific recognition. Eugene Garfield, the man behind the Science Citation Index (SCI), gave researchers a powerful new tool to track academic influence: the citation count. Before the publication of the SCI, it was virtually impossible to calculate citation counts as it required to manually review the reference list of countless papers published in scientific journals each year. With the SCI, a new metric was born—objective, comparable, and seductive in its simplicity. Suddenly, a scientist’s worth could be measured not only by what they published, but also by who paid attention.
The promise of clarity came with risks: what happens when a single number begins to define the arc of an entire career? In their paper Measuring Science: Performance Metrics and the Allocation of Talent (American Economic Review, December 2024) Carlo Schwarz of Bocconi’s Department of Economics, Sebastian Hager of Oxford University, and Fabian Waldinger of Ludwig-Maximilians-Universität München offer the first rigorous evidence on how citation metrics have reshaped the scientific labor market in sometimes surprising ways.
A quasi-accidental natural experiment
Carlo Schwarz and his co-authors noted a peculiar quirk: the SCI’s early editions neither covered citations in all years, nor in all journals. For technical reasons, it was only possible to analyze a subset of citations. As a result, some citations were recorded and made “visible” to the scientific community, while others remained “invisible”. This patchwork of coverage created a rare opportunity: a natural experiment where otherwise similar citations had different career consequences simply because of their visibility.
Imagine two scientists: both publish groundbreaking work in 1957. One was cited in Nature in 1961—a journal and year covered by the SCI. The other is cited in Nature in 1962—a year the SCI skipped. Only the first citation was visible to contemporaries. As the authors show, this arbitrary difference in visibility had a measurable and important impacts on careers, department placements, promotions, and funding.
When citations became career currency
The introduction of citation metrics immediately started to influence the matching of scientists to departments. Those with visible citations were significantly more likely to land at top-ranked universities. The correlation between citation counts and department rank rose by 61% between 1956 and 1969. Meanwhile, the predictive power of raw publication counts fell by nearly half.
Notably, the authors show this was not merely correlation. By comparing visible and invisible citations—randomly distributed by time and journal—they demonstrate that visibility itself drove outcomes. A scientist with a 10-percentile higher visible citation rank was, on average, placed in a department 2.5 percentiles higher; for invisible citations, the effect was only 0.6.
The SCI, in effect, made talent legible. But legibility came with amplification: visible stars were rewarded more than their equally brilliant but unseen peers.
Rising from obscurity
One of the most intriguing aspects of the study is who benefited the most. The authors identify a class of “hidden stars”: highly cited scientists stuck in lower-ranked departments. Once citations became visible, these researchers were finally noticed and often moved to more prestigious institutions. This visibility effect was particularly strong for women and those with PhDs from less prestigious universities.
Citation metrics also had a modest but positive effect on minority scientists—particularly those who were already star performers. In an academic world often marred by informal networks and implicit bias, the SCI acted as a kind of equalizer, helping exceptional scholars from overlooked backgrounds rise through the ranks.
A new geography of recognition
Citation metrics also reduced information frictions—especially across distance. Before SCI, it was harder for hiring committees to evaluate someone they hadn’t met or whose work hadn’t circulated locally. Citation data made talent portable. The effect of visible citations was strongest for moves to geographically distant, higher-ranked departments where prior personal connections were unlikely. In this way, the SCI acted as a passport for ambitious scientists, a way to transcend the constraints of geography.
The “Matthew effect”
Visibility also impacted promotions and research funding. Scientists with more visible citations were 5.8% more likely to be promoted and 19% more likely to win a National Science Foundation (NSF) grant. The authors consider this as an example of “Matthew Effect”: to those who have (visible citations), more is given. As Carlo Schwarz puts it, “These results indicate that citation metrics had direct impacts on the careers of scientists and changed the allocation of resources. Scientists with many visible citations accrued additional rewards and recognition, suggesting the presence of Matthew effects.” This is not merely academic bookkeeping. It speaks to how early visibility can compound advantages, creating trajectories of success—or obscurity—that persist for decades.
From SCI to Spotify
The implications go beyond academia. Today, we live in a metrics-saturated world. From social media likes to Spotify play counts, performance data shape visibility—and visibility shapes opportunity. Just as citations became the academic equivalent of "streams," our careers, tastes, and choices are increasingly filtered through what can be counted, ranked, and tracked. But, as Carlo Schwarz and his colleagues remind us, metrics are not neutral. They’re powerful. They shape the stories we tell about talent—and the future we allow to unfold.
