The 18th century theologian and mathematician Thomas Bayes, who first established the mathematical basis for probability inference, did not live to see the publication and application of his famous theorem.

Jeff Gill, SPA’s Distinguished Professor of Government and Math & Statistics, and who joined SPA in Fall 2017, manages this legacy with care and enthusiasm. He applies Bayesian modeling and data analysis to questions in general social science quantitative methodology, political behavior and institutions, and medical data analysis (especially physiology, circulation/blood, pediatric traumatic brain injury, and epidemiological measurement), using computationally intensive tools.

“When I introduce myself at talks, I say I study humans, biomedically, socially, and politically,” said Gill. “Humans are the hardest thing you can possibly study. [The physical sciences] agree on about 90% of knowledge and argue about the remaining 10%. In sociology, political science, or public policy, it's the other way around.”

Though his first-year statistics class left him cold, the one-time UCLA music major considered graduate school in math. However, he could not picture himself sitting alone proving thermos in a windowless office. “I knew that wasn't me.”

Gill’s next stop was an MBA from Georgetown University. He worked in the private sector in London and Glasgow before realizing that an academic path suited him better.

“In the business world, people can come into your office and tell you what to do,” he explained. “[No one] can come into my office and tell me what to do. And I love that. I'm also a child of academics.”

Gill headed here to American University, where he earned a double PhD in statistics and government. He then completed a post-doc year at Harvard University, working closely with world-class statisticians Don Rubin and Gary King. After successful tenure-track positions at other prestigious schools, he spent most of his career at Washington University in St. Louis in the Department of Political Science, with supplementary appointments in the School of Medicine and the Department of Mathematics, before returning to American University in 2017.

At AU, Gill was named inaugural director of SPA’s Center for Data Science, where he coordinates and supports empirical research across the campus by developing links with federal agencies, providing research support to faculty and graduate students, and building infrastructure to handle large and complex datasets. In 2019, he founded the Masters in Data Science program, now the largest in-person master's program on campus. Gill has published in a wide range of journals spanning social science and medicine, from the Journal of Public Administration Research and Theory to Lancet Neurology.

“The Data Century”

“We live in the data century,” said Gill, “whether we like it (or know it) or not.”

The Digital Revolution, like the Paleolithic, Neolithic, and Industrial Revolutions before it, is transforming global society. While the 20^th century brought gigantic advancements in engineering, chemistry, and physics, which surprisingly relied less on computing, the 21^st century, Gill maintains, will record gigantic leaps in the social and biomedical sciences, largely because of computational and methodological tools developed by scholars in recent years.

“Over 95% of big data machine learning is focused on understanding humans,” he said. “If you go to an astronomy or physics convention, they don't sit around and talk fixatedly about AI. But at a social science conference, it seems like it’s the most discussed topic.”

For example, while physicists still debate and explore the foundational elements of the universe, computational tools like large-scale statistical analysis and CRISPR have helped geneticists make huge advances in gene identification and manipulation.

“The gigantic leaps in genetics were facilitated by computation on a scale that was impossible in the 20^th century,” said Gill. “We're seeing this effect in the social sciences as well, particularly with large language models and other emerging tools.”

As the Digital Revolution evolves, scholars across fields of study are collaborating to tackle the trickiest human problems. Computer scientists often need societally substantive problems to address, Gill said, making them hungry to work with social scientists. Meanwhile, he fears that social scientists sometimes receive too little statistical and computational training.

“As a statistician who likes to study humans, I’ve had the privilege of working with lots of amazing people in medicine, political science, and the social sciences in general,” said Gill. “Though my personally most valued publications are in statistics journals, about a third are in medical journals.”

Gill considers one of these interdisciplinary pieces particularly impactful. In “Controlling Phlebotomy Volume Diminishes PICU Transfusion: Implementation Processes and Impact,” published in 2017 in the journal Pediatrics, he and co-authors identified safeguards against overdrawing blood from sick infants.

“My co-authors worked in the pediatric intensive care unit (PICU) at Washington University, St. Louis,” he said. “I actually spent a lot of time there . . . and it's a super powerful experience. The attending physicians work across three rotations, and any of these may order a blood test. Also, if the nurse draws too little blood, the lab will ask for a redraw, so they tend to overdraw blood to prevent that. The outcome was screamingly obvious. You just don't want to draw too much blood out of sick babies.”

Gill’s heavy CV also includes “A Methodological Manifesto for Public Administration,” written with SPA Distinguished Professor Ken Meier, which called out methodological mistakes in public administration scholarship.

At SPA, Gill teaches graduate-level statistics courses to a mix of social science PhD students and data science master students, and they all get the same advice.

“A really great intellectual combination, whether you want to work in industry or academia, is to combine a serious, deep technical skill with some deep social science interest: some aspect of human behavior, human biology, human interaction, or human cognition,” said Gill. “That is the winning career combination for the 21^st century.”

While passion and ideology may bring young people to D.C., it alone won’t keep them there, he argued.

“You can't stay in this city unless you marry your passion with some technical skill,” Gill explained. “Sure, you're passionate about economic development, women's reproductive rights, foreign policy, or gun control, and I love that. But you can't stay here unless you have something else to offer. Employers aren't going to keep you. Take what you love and put a technical skill on top of it, and you will thrive in highly competitive Washington.”

Gill has supervised around 30 PhD students over the course of his career, and left them all as applied statisticians, with skill sets in statistics, data science, AI, mathematics, or computation.

“My PhD students in the government department here take every [math or computer science] machine learning class they can fit into their schedule,” said Gill. “Then I encourage them to study whatever they want in humans, biomedically, socially, or politically.”

Gill encourages his students to pick up as many methodological tools as possible, for two reasons. First, scholars should learn to use these tools before they are on the tenure clock and needing to produce papers, to avoid the need for on-the-fly training. Secondly, competence in a variety of these tools can widen their potential base of coauthors.

Antisemitism Decoded

Sometimes Gill’s work reveals the darker side of human behavior. As part of the AU Signature Research Initiative, Gill, in partnership with colleagues in the computer science and journalism departments, received funding to create tools for decoding antisemitic language online. The team has produced four papers, an app, two large surveys, and heavy qualitative work, to provide tools to be used by the U.S. Holocaust Museum and the Southern Poverty Law Center, among others.

“These antisemitic terms tend to have a life cycle,” Gill explained. “They're initially used by a small in-group. Then the group widens, and one of two things happens. Sometimes they go into the public lexicon, and sometimes they just die off.”

For example, the phrase “cultural Marxism,” first limited to obscure online spaces, has grown so common that a current member of Congress has published a book with the term in the title.

“This project is fascinating because it combines really cool, very different methodologies,” said Gill. “It's qualitative. It's quantitative. It's computational. It involves a lot of wonderfully motivated grad students and postdocs.

“It makes you feel bad about humanity, to be honest. But it needs to be studied.”