With the recent passing of Justice Antonin Scalia, a vacancy has opened up on the Supreme Court in the middle of an election cycle that could have drastic ramifications on the ideological makeup and leaning of the highest court in the country. Controversy is swirling as many Republicans are crying afoul over outgoing President Obama being able to appoint a new justice (who has already appointed two justices during his tenure as president). Their argument is that the “lame duck” president (though technically, having been elected through November 2016, he’s not a lame duck until the period of time following this year’s election before the winner assumes office in January) is on his way out, and should not be able to politically appoint such a powerful figure that serves a lifelong tenure.
Many Republicans believe Scalia’s seat should remain unfilled until our next president is elected, burdening the country with an 8 member court that is split ideologically down the middle, where a tie decision would end up dismissing or overruling any case they tackle. Because our next elected president wouldn’t be able to appoint a justice until they take office in January of 2017, that would leave an open seat on the court for at minimum a full year, the longest since 1969 when it took Lyndon B. Johnson a little more than a year to fill a vacancy. With all eyes on potential nominees and their political philosophies, ideologies and judicial temperaments, researchers at the Discovery Analytics Center at Virginia Tech are utilizing artificial intelligence to deliver insight into all of this court talk.
The team proposed a way to model the voting behavior of the Supreme Court by employing computer-based machine learning in order to create a pretty accurate assessment of each justice’s views on legal and political issues. This model, dubbed the Supreme Court Ideal Point Miner (SCIPM), is a data-powered framework that can actually learn the behavior and judicial preferences of each justice in order to answer theoretical questions about how a justice would vote on a given case. Predicting insights on their political alignments and views on separate issues, the researchers were even able to identify which justices might be a swing vote in different cases.
Still, it’s a tricky task being able to predict the behavior of a justice, as a healthy amount of information including prehearing filings, posthearing meetings, written opinions, voting records, formal decisions and the dissenting arguments in each majority decision must be taken into account for each justice, especially since many cases the court takes on involve more than one issue, where each justice might have differing stances on a given issue. The SCIPM can augment and consolidate all of these preferences and opinions from these texts, specifically counting the number of words relating to a given issue and assigning a weight to that issue in order to create a fairly accurate prediction of behavior.
The model aggregated and analyzed various opinions from 2010 to 2014 in order to paint a picture of how each justice feels about a given issue. When it came to those likely to be swing voters, the model identified those by looking at how a given justices’ views vary across multiple issues, particularly in cases decided with a 5-4 margin. A justice whose voting record had a larger range across the issue spectrum is more likely to be a swing voter than one who more consistently falls on one side of the ideological spectrum for a given issue. In particular, the model singled out Justice Kennedy as the most likely to swing because his positions are more varied than the other justices’. With an accuracy level of nearly 80 percent, the SCIPM model is an effective and enlightening way to highlight the Supreme Court decision process and predict the outcomes of future cases based on the past decisions and behaviors from the current crop of justices.