The Case Salience Index, Public Opinion, and Decision Making on the U.S. Supreme Court

ABSTRACT

Recent work on the U.S. Supreme Court argues that Supreme Court decision making is constrained by several factors, including public opinion. These constraints are not uniform across cases, however, and some suggest that public opinion may only act as a constraint conditional on the salience of the case. Using a dichotomous measure of salience, prior studies finds mixed results as to whether the Court may alter its decisions in cases of low or high salience. By updating our previous measure of salience (Collins and Cooper 2012) and recasting it with slightly different measurement properties, we find that public opinion influences court decisions in cases of very low salience and cases of very high salience. This research has important implications for our understanding of whether the Supreme Court is insulated from public pressures or if it simply reflects the public mood of the day. The analysis introduced here also provides a useful example of a new salience dataset that should be valuable to judicial scholars.

KEYWORDS:

• Supreme Court
• public opinion
• judicial decision making
• salience

Notes

¹ As we detail in the appendix, we do see differences in case coverage from the CSI, New York Times, and CQ Press's list of “landmark decisions.” Both the New York Times front-page coverage and the CQ Press list include a higher percentage of liberal decisions than any level of the CSI. Table B in the Appendix displays a logistic regression model of coverage of the front-page New York Times, CQ Press, and the CSI's highest levels of salience, which shows a potential liberal bias to the New York Times front-page coverage as well as a bias toward cases from New York, as have been suggested by other researchers (Maltzman and Wahlbeck 2003; Unah and Hancock 2006).

² In conducting our analyses, we found that running separate models for each of our nine salience levels (CSI scores of 0 to 8) resulted in a loss of data for some terms. For example, there were no cases in 1981 that received an 8 on the CSI scale, such that we would lose that term from the analysis. However, collapsing the categories did not result in these losses. While we believe this would not be an issue with case-level data, using the full scale of the CSI could pose some issues when examining aggregate data by term.