Highlighting Hidden Statistical Gems

April is Mathematics and Statistics Awareness Month, a time to celebrate the contributions of mathematics and statistics in driving discoveries, solving real-world problems, and helping us understand the world. In the spirit of MSAM, I want to highlight contributions from our community that are sometimes overlooked but are vital in science, especially in multidisciplinary research teams. In my own work in cancer research, I see this every day—my colleagues express their appreciation, and I can clearly see the impact of statistical thinking and leadership on scientific progress.

Many believe artificial general intelligence will be a reality in the next few years. We know AGI is not possible without statistics, but our essential contributions are sometimes overshadowed by other disciplines. Public perception and funding trends often favor AI-driven solutions, while overlooking the statistical foundations that ensure rigor and reliability. We need to raise awareness outside our community about the essential role of statistics in AI technologies.

Many peers talk about the paradox of statistics: Our work is invaluable, yet often invisible. The joke about statistical power captures this irony well: “We calculate power, but we don’t have power.” John Tukey once described the pleasure of doing statistics as the opportunity to “play in everyone’s backyard”—an essential tool that belongs to all but is rarely given the spotlight.

As a biostatistician in oncology research, I know firsthand how essential our discipline is. Without the Kaplan-Meier curve (Kaplan & Meier, 1958) and log-rank test (Mantel, 1966; Peto et al 1972), the Cox proportional hazards model (Cox, 1972), generalized linear models (Nelder & Wedderburn, 1972; McCullagh & Nelder, 1989)—including logistic regression, Random Forest methods (Breiman, 2001), Simon’s two-stage design for Phase II clinical trials (Simon, 1989), and, more recently, BOIN design (Yuan, Hess, Hilsenbeck, Gilbert, 2016), my daily work could not move even one step forward. Study designs, sample size calculation, visualization approaches, regression models based on least-squares fitting, and multiple imputations are so fundamental that they are often no longer cited in the literature. They have become default tools, deeply embedded in the framework of medical and clinical research. They are hidden gems.

This month, I want to make several “hidden gems” visible by celebrating their profound impact on statistics. I’ll highlight major contributions—from internationally recognized statisticians to the most highly cited statistical papers in history to methods that have shaped entire fields, especially in the AI era.

A celebration of statistical gems must include the winners of the International Prize in Statistics, so I begin by sharing the citations associated with the award.

C. R. Rao is recognized for his remarkable 1945 paper published in the Bulletin of the Calcutta Mathematical Society. Calyampudi Radhakrishna (C. R.) Rao demonstrated three fundamental results that paved the way for the modern field of statistics and provided statistical tools heavily used in science today. The first, now known as the Cramér-Rao lower bound, provides a means for knowing when a method for estimating a quantity is as good as any method can be. The second result, named the Rao-Blackwell Theorem (because it was discovered independently by eminent statistician David Blackwell), provides a means for transforming an estimate into a better—in fact, an optimal—estimate. Together, these results form a foundation on which much of statistics is built. And the third result provided insights that pioneered a new interdisciplinary field that has flourished as “information geometry.” Combined, these results help scientists more efficiently extract information from data.

Sir David Cox is recognized for his 1972 Journal of the Royal Statistical Society, Series B, paper in which he developed the proportional hazards model that today bears his name. The Cox Model is widely used in the analysis of survival data and enables researchers to more easily identify the risks of specific factors for mortality or other survival outcomes among groups of patients with very different characteristics.

Bradley Efron, professor of statistics and biomedical data science at Stanford University, is recognized for the “bootstrap,” a method he developed in 1977 for assessing the uncertainty of scientific results that has had extraordinary impact across many scientific fields. With the bootstrap, scientists are able to learn from limited data in a simple way that enables them to assess the uncertainty of their findings.

Nan Laird, Harvey V. Fineberg Professor of Biostatistics (Emerita) at Harvard T. H. Chan School of Public Health, is recognized for her work on powerful methods that gave researchers the tools they needed to answer important questions in health, medicine, psychology, and more. This jump-started a field known as “random effects modeling for longitudinal data analysis,” and the methods Laird introduced in 1982 are still the most widely used techniques in both observational studies and clinical trials today.

Andrew Gelman and Aki Vehtari, in a 2021 Journal of the American Statistical Association paper, highlighted some of the most important statistical ideas of the past 50 years—“counterfactual causal inference, bootstrapping and simulation-based inference, overparameterized models and regularization, Bayesian multilevel models, generic computation algorithms, adaptive decision analysis, robust inference, and exploratory data analysis.” These methodological breakthroughs have not only advanced statistical theory but also transformed practical applications in fields ranging from medicine and social sciences to artificial intelligence and engineering. Emphasizing the field’s deep ties to real-world applications, they underscored the importance of remaining open to ideas from other disciplines.

In a separate effort, Michael Schell and I have been working on a classification system for statistical papers to track and update our understanding of key research areas. So far, we have identified 4,492 highly cited statistical papers from 147 peer-reviewed journals and built a publicly available database. Among the top, high-impact papers, several influential statistical ideas emerge, and here are a few examples: false discovery rate (79,257 citations); agreement of continuous data (40,939); AIC (37,973); support vector networks (36,069); neural net overfitting (26,328); nonlinear least squares (24,084); and elastic net (11,785).

What stands out is the profound contributions of statistics to modern science. Like many of you, I was struck by how deeply these statistical concepts are woven into modern machine learning, AI, and other scientific advancements. Recognizing these sometimes-hidden gems not only increases our appreciation for the field but also reinforces the vital role statistics continues to play in shaping discovery and innovation. These methods are not simply tools; they are foundational ideas that have propelled science forward.

But recognition shouldn’t stop with famous names and landmark methods. Another group also deserves attention: the statisticians embedded in team science. These everyday statisticians quietly ensure the validity of research, design rigorous studies, and safeguard the integrity of scientific conclusions. I often call us “Public Safety Officers in Science.” We navigate complex collaborations, bridge gaps between disciplines, and serve as the final guardrails against misinterpretation of data. Though we rarely receive public recognition, our impact is profound.

Without statisticians, many breakthroughs in medicine, public health, economics, and engineering would lack a solid foundation. We need to be better at advocacy to secure our place in funding discussions, decision-making circles, and publications. Addressing this gap also means expanding recognition—not just for those who develop widely known methods, but also for those who apply them effectively, ensuring science moves forward with accuracy, reliability, and integrity.

It’s time to make all statisticians—both the celebrated and the unseen—more visible.

Thank you for the important role you play in our ASA community.