README: 'Epic' Impact
By Ben Hancock
This GitHub repository is part of a data journalism project by ALM Media data editor Ben Hancock and National Law Journal labor law reporter Erin Mulvaney to assess the impact of Epic Systems Corp. v. Lewis, a May 2018 decision by the U.S. Supreme Court that centered on the enforceability of mandatory workplace arbitration agreements. In collaboration with Casetext, we identified and categorized nearly a hundred U.S. district and appellate court decisions that cited Epic in calendar year 2018, in order to quantify the scope of Supreme Court's impact on this issue in the months after it was handed down. The story was published on the NLJ website in February 2019, and was the cover story for the NLJ's March 2019 print magazine.
In this repository, you will find the data set we assembled to report this story, a data dictionary explaining the different variables and codes (DICTIONARY.md), and a Jupyter Notebook that shows how we did our analysis. Our goal in presenting this to our audience is to "show our work" and give greater transparency into the approach we took in reporting an issue with wide public interest and significance.
We reviewed nearly a hundred decisions by U.S. federal courts that were identified by Casetext’s legal research tool as citing Epic Systems in 2018. If there was more than one such decision handed down in the same case, we excluded all but one in favor of the decision that dealt most closely with arbitration. In total, we identified 17 circuit court decisions and 75 district court decisions, some of which were marked as unpublished.
In order to analyze this corpus of cases, we manually categorized the opinions along some key metrics. These included whether the case was a proposed class or collective action, case type, and whether arbitration was the main issue before the court in the decision. Selecting a case type was often the most challenging task, as many cases involve a slew of legal claims. For example, some cases involving allegations of gender discrimination may also allege general wage-and-hour law violations. For such cases, our rule was that if discrimination or sexual harassment claims were present, they would prevail for the purpose of categorization.
A brief note about what we can and cannot learn about the impact of Epic strictly from the data we gathered: First, we cannot conclude that Epic directly caused cases to be compelled to arbitration at a higher rate than before. We did not gather data on how similar cases were decided prior to Epic, and so do not purport to have a baseline for comparison. It’s also difficult to generalize the degree to which Epic was the sole or the most important factor in a body of cases being compelled to arbitration, given the unique fact-patterns in much of the litigation.
But what we can say is that in the majority of cases where arbitration was the key issue before the court in a case citing Epic, arbitration won out, and that for cases involving workplace-related claims, that statistic was even more pronounced.
Questions & Comments
If you have questions or comments about our categorization or overall methodology, please feel free to contact ALM data editor Ben Hancock at firstname.lastname@example.org.