Interview by Kelundra Smith
When Google collects data about you and sells it to other companies, does this practice damage competition and consumers? Does Amazon’s broad involvement in e-commerce increase prices for consumers in the long run? These are the types of legal questions that Susan Navarro Smelcer sought answers to as an attorney at an antitrust firm in D.C. However, the demands of being a lawyer made it hard to do a deep dive into these matters. When she came across a new faculty position in the College of Law Legal Analytics & Innovative Initiative, she jumped at the opportunity to share her love of the law and interest in data with students.
At her firm, Axinn, Veltrop and Harkrider, she applied her quantitative skills to analyze price and production data in antitrust class actions. She also developed computer programs to simplify answering complaints in matters involving many plaintiffs who file nearly identical complaints, such as class action lawsuits. This fall, she is teaching Legal Analytics I, where students will learn the basics of empirical research design and how to apply quantitative methods to text. She will co-teach Legal Analytics II in the spring.
Here, she discusses why there’s no getting around big data.
How scalable is legal analytics in the field? Will it only be possible for big firms to use this technology?
E-discovery is dominated by vendors that use algorithmic methods and machine learning techniques for identifying relevant documents. Keyword searches are still useful, but you don’t have to be a big firm to use e-discovery tools. Computing power is so cheap that law firms will be able to use techniques to automate substantive things like case review, or in non-legal ways such as billing.
Do you ever worry about putting yourself out of a job?
No. You can only automate so much. There is always going to be a need for attorneys to do the things that require professional judgement. I can write a program that simplifies answering complaints, but a computer can’t write the initial answer itself. There’s always going to have to be an attorney at the back end. The tools just make us better attorneys.
What made you say yes to Georgia State College of Law?
The opportunity to combine my interest in the law with my love of using data to think about the law is explicitly the job here. My research expectations are interdisciplinary, which is what I love. Plus, I love Atlanta. I spent a lot of time here [when I was earning my Ph.D. in political science at Emory]. I’m looking forward to teaching and getting to know the students. I’ve been designing this class in my head for three years and now someone is paying me to do it.
What area of research are you most interested in?
I’m interested in looking at how courts’ analyses of data in assessing mergers doesn’t account for missing data that might bias the analysis. One example is the Sysco/U.S. Foods merger—they’re basically the only two companies with contracts for national broadline food distribution. In their sales management software, their sales reps record when they get a sale, when they don’t get a sale and who the major competitor is. But, what if they don’t record every competitor, because that competitor is a smaller regional player or a cash-and-carry system for a local restaurant? So, you end up with data that says that Sysco’s major competitor is U.S. Foods, but that may not always be true.
Another area I’m thinking about, but haven’t really explored yet, is disparate impact in Title VII cases. Employees can allege that they were discriminated against if they are a part of a protected class, but you have to prove that it disproportionately affects a certain group. But what if you have bad records? For example, what if managers are not entering their disciplinary data for women of color in the workforce? Say the allegation is that women of color are held to different time standards in the workplace– then you’re going to conclude that there was no discrimination. That comes down to a data cleanliness issue. The courts often treat data like [the gold standard], but we have to think about how the data were generated.
What do you think is the most important thing for students to take away from the law school experience?
My goal is for them to be much more skilled critical readers of opinions and cases. I want them to think more about data. I want them to think about the assumptions that go into data collection so that they can make better arguments. To think like a lawyer is not so different than thinking like a social scientist. You have to recognize assumptions built into an argument and evaluate those assumptions.
I hope that they come out of law school like they have a good foundation to keep learning. Sometimes you work really long hours or for difficult people, but at the end of the day you get paid to read really interesting things.