The advent of big data in social science

May 5, 2021
A Zoom screen with six people talking.

By Juan Siliezar
Harvard Staff Writer

The FAS Division of Social Science held its first ever “Dean’s Symposium on Social Science Innovations” earlier this week, focusing on the advent of big data along with its opportunities, contributions, challenges, and limitations.“[Big data is] an arena in which several of our departments are actively engaged in cutting edge scholarship and developments,” said Dean of Social Science and W. E. B. Du Bois Professor of the Social Sciences Lawrence Bobo. “The topic lends itself to bringing together work now taking place in economics, government, sociology while [pulling] from the history of science and psychology, as well.”

The event centered on three panelists who are harnessing big data to advance their groundbreaking research. They were Raj Chetty, the William A. Ackman Professor of Economics at the Faculty of Arts and Sciences and director of Opportunity Insights; Gary King the Albert J. Weatherhead III University Professor and director of the Institute for Quantitative Social Science; and Mario Small, the Grafstein Family Professor in the Department of Sociology and a visiting professor at the Harvard Business School.

Chetty spoke on upward mobility and using government records to see what’s happening on the ground in neighborhoods across the country. His research applies big-data tools to massive economic, geographic, and demographic data sets and allows analysts to zero in on variables like race, education, and income. Ultimately, Chetty said, the goal is to use big data to understand how children from disadvantaged backgrounds can be given a better chance at succeeding.

“Traditionally, when social scientists are faced with a problem like this, we have had relatively limited data to rely on,” Chetty said. “The type of data I focus on — administrative data from government sources — has really in the past 10 years opened up tremendous possibilities in being able to study questions like this in a much more precise way by unpacking these national trends.”

King presented on how academic researchers can and must partner with large private enterprises to access important sources of big data or risk missing out.

“We have more data than ever before,” King said. “But we have a smaller fraction of data in the world about the subjects that we study than ever before because most of it now is locked up inside industry and other organizations that have done a great job at collecting information about the people, groups, and societies that we study. They have it all locked up, so we have no choice if we’re going to do our jobs [of analyzing data for public good] but to find ways of working with industry.”

To do this, King said it’s not about compromising the integrity or balance of the research but making sure both parties are educated about each other and coming up with creative solutions to problems, such as a company wanting final approval on what research gets published using their data.

Small focused on what big data analysts can learn from qualitative research and how that can be applied to answer critical questions about urban inequality.

He went over a study analyzing limited access to banks in minority neighborhoods and ones with high poverty rates compared with more white and affluent neighborhoods. Small and his colleagues combined government data, private data, and data from apps like Google Maps to break down how many minutes it would take to walk, or drive, or take public transit to the nearest bank or alternative financial institute (like a cash advance store) for every block in the 19 of the U.S.’s largest cities.

The discussion was chaired by Elizbeth Phelps, the Pershing Square Professor of Human Neuroscience, and Peter Galison, the Joseph Pellegrino University Professor, offered comments and questions for the participants.

They also shared what they saw as the impact and role of big data over the next 25 years.

“I think the future is very, very bright with respect to what we can learn,” Smalls said.

King added that the moment marks a kind of jumping off point for social scientists using big data: “This is equivalent to when they first handed out micro microscopes to microbiologists,” he said. “We for the first time, are actually have enough information to be able to solve some or ameliorate some of the major challenges that affect human society. I think we’re beginning to really tackle them one by one.”