What happens when we make arrest records or property records readily available? How do we design analyses of health records to respect the work of health practitioners, recognize environmental threats to health, and accurately capture health information? An NSF funded workshop organized by Susan Sterett of CPAP and Kelly Joyce, a sociologist of science at Drexel University, will invite conversations about thorny questions like these during a two-day discussion of data analytics.
Advances in computing power and analytics, the capability to make government administrative records public, and the data that people generate as they move through their lives with smartphones and in cities with sensors have all turned us toward the promise of data to solve problems. Excitement about data analytics that can manage very large quantities of information has upended many expectations about how we analyze people’s lives. The White House has issued reports on the promise of new data analytics for better evidence-based policymaking, while cautioning us to be aware of the risks of exacerbating inequality, or threatening privacy. prioritized a national discussion around big data analytics. The National Science Foundation established regional hubs to coordinate big data analytics, which Virginia Tech is part of. The workshop will include participants from the South Data Hub as we work together and across disciplines to identify challenges and opportunities in relation to big data.
However, more access to greater computing power and more records do not automatically result in improvements for society. Indeed, sometimes this data may unintentionally exacerbate inequalities by reinscribing sexism, racism, ageism and classism into data mechanisms or outputs. We need to make sure we ask good, theoretically informed questions, we need to make sure the data link to those questions, and analysts and users need to make sure that they understand what they are doing with analyses. The more we think about data analytics, the more questions multiply. How are data produced? What does that tell us about analyses?
Hosting a Vibrant Conversation amongst Diverse Thinkers
Virginia Tech’s Center for Public Administration and Policy is co-hosting a two-day conference on Collaboration as Big Data Ethics with Drexel University’s Center for Science, Technology and Society (STS).
The conference will bring together leading scholars and and practitioners in data analytics, including those who work in government or health care, computer scientists, and social scientists.
“How we do big data analytics will affect people’s lives substantially,” says Susan Sterett, principal investigator for the workshop. “We need to work together to ask good questions and create valid models. Social scientists work on what organizations do, which includes processes for producing data. These processes matter: predictive analytics for policing, for example, rely on data gathered in a system that polices more aggressively in poorer communities. ig data analytics need to be structured in line with our commitments to equality, for example.
“While we can see examples of problems, we can also assume that databases can be linked to illuminate complex problems when practitioners know they cannot be. We can only learn by relying on good theorizing, and the different kinds of knowledge people with different expertise bring to bear. ” says Sterett.
Sterett and Kelly Joyce will open the two-day workshop with a presentation on Big Data, Disciplinary Expertise, and Building Community for Empirical Ethics.
Kelly Joyce has been collaborating with an information science colleague to understand how data scientists create and work with algorithms and software to create big data sets and the individual, professional, cultural, and institutional values and incentives that drive this work. As Dr. Joyce notes, “People are excited about the promise of big data to answer meaningful questions about society, but we need to be reflective about the disciplinary expertise that shapes the questions asked and the data content.”
Creating Big Data Ethics in a Rapidly Changing World
One challenge of developing ethical principles around big data analytics is that the field is changing so rapidly. “One way to view ethics is as a fixed set of rules that individuals learn once and then practice,” says Sterett. “Another is to work out collective ethical practices, to include the environments in which decisions are made, and the pressures to make good or bad decisions”
Crafting a flexible, collaboration-based system could lead to better theorizing, awareness of the promise and limits of data and data bases, and ethical principles against collecting and publicizing more data just because we can. ciologists of science arguehat collaboration across different kinds of expertise makes for better decisions.
Susan Sterett is a professor at the Center for Public Administration and Policy within the School of Public and International Affair.