The Evolution of Consensus in the U.S. Supreme Court

Abstract

For small, collegial bodies such as the U.S. Supreme Court, interpersonal dynamics play a key role in the decision making process. Previous work on the Court has focused on the existence of "consensual norms," examined only the presence of concurring and dissenting opinions, and rested entirely on aggregate statistics. As a result, that work is limited in its ability to uncover inter-justice conflict during the 19th and early 20th centuries and fails to explain whether changes in the Court's collegiality are due to shifts in issues, personnel, or the larger political environment. A sentiment analysis of all opinions -- majority, concurring, and dissenting -- issued by the Court between 1792 and 2000 allows us to measure the degree of consensual and discordant content for every opinion, justice, and term of the Court. Leveraging data at the opinion level also allows us to exploit an identification strategy for uncovering "chief justice effects." Our findings suggest a permanent increase in overall levels of valenced language in dissenting opinions corresponding roughly to the ascension of Chief Justice Taft, and a similar increase among majority opinions under Chief Justice Burger. However, though we uncover substantial justice-level differences in the positivity or negativity of opinion content, we find little evidence of systematic variations in the polarity of language associated with the regimes of particular chief justices.