GSA chief Denise Turner Roth joins Harris faculty William Howell and Brett Goldstein in a lively discussion of the future (and limits) of data-driven decision-making

When data analysis drives public policy, the results can be surprising – and surprisingly effective. 

Denise Turner Roth, now the administrator of the U.S. General Services Administration (GSA), first realized the importance of open data early in her career, when a former mentor took on a new and challenging task: leading a foundation that raises money for breast cancer research and treatment for women in North and South Carolina.

Looking for ways to increase the impact of their donations, the foundation chief drilled down into a massive study of breast cancer statistics and noticed something important: Patients who had to travel long distances to reach their doctors or treatment centers were far less likely to survive than similar patients in better-served areas. Armed with that crucial data point, the foundation worked with donors and local agencies to raise money targeted for mobile mammogram detection units that could bring affordable, top-quality care to at-risk women living in rural areas.

“It opened my eyes to the benefit and power of open data,” Roth recalls. 

Today, as head of GSA, Roth works to put that early lesson to use on behalf of taxpayers, focusing on finding smart ways to use data to increase effectiveness, efficiency and transparency in a federal agency that oversees $66 billion worth of annual procurements, helps to manage about $500 billion in federal property, and employs about 12,000 workers.

Turner was the featured guest at a Oct. 26 lunchtime panel discussion, “Can Data and Technology Fix Government?” presented by the University of Chicago Harris School of Public Policy’s Project on Political Reform. Her conversation with Brett Goldstein, Senior Fellow in Urban Science at Harris, was moderated by William Howell, Sydney Stein Professor in American Politics at Harris.

“There’s data everywhere,” Goldstein reminded the overflow crowd of Harris students and faculty. He said that the wealth of data can be used both to create innovative government policies and to enable the private sector to create products based on public data. As an example, Goldstein cited SweepAround, an app that alerts Chicago residents when their neighborhood streets are scheduled for cleaning. 

“Would (building that app) be a good use of tax dollars? Probably not,” commented Goldstein, who previously served as the city of Chicago’s first municipal chief data officer and as the city’s CIO. “Am I super-thrilled that someone built that app? Yes. Did I get a lot of parking tickets before I got that app? Yes.”

But Goldstein cautioned that expanded data access alone is not sufficient to drive good policymaking. “At the University of Chicago, we respect doing your math well,” he said. “You go into a typical office, they have Microsoft Excel, and anyone can do a regression. Well, as super-smart quant-y people, we know you can do bad regressions. So now we’re in a world in which we have big data, high-performance computing, and people developing proprietary algorithms in a black box. When people don’t show their work, we need to worry.” 

Given that political considerations shape government decision-making at every level, Howell raised concerns about elected officials and government employees who may argue against opening datasets to public analysis and public scrutiny. 

“Some people want to resist the push towards open data,” Howell said. “Data may be neutral, but the purposes to which they are put are not. Data can be used for certain policy initiatives that some people disagree with. Strategically, how do you engage with folks who say, ‘We want to keep this stuff under wraps?’”

“Data points can point you in different directions,” Roth acknowledged. “At GSA, we have by default said that open is best. Certainly there are qualifications in terms of security, who has access, things of that nature. We have to guard against public information spills that can affect millions of people, but that doesn’t mean that we go back into our turtle shells. Our focus is opening as much as we can, as quickly as we can. The more eyes you have evaluating data points, at a very agnostic level, the better the conversation.

“The ways Americans are sharing information with each other and building community among themselves have changed,” Roth continued. “We’re talking here about open data access, but I think communities have the power to force what happens next with that data. That’s a conversation for a community to have. Data only gets us so far; we are still human beings who have to make choices about what we do next. We don’t want a world in which we are not innovative because we fear what comes out the other side. That’s where the human part steps back in.”