Government Research Intern: Michael Mazidi

Title: The Supreme Court and Statutory Interpretation: Are Justices Judges of Law or Politicians in Robes?

Author: Michael Mazidi


Justices on the Supreme Court for many years now have debated how to utilize ‘intent’ (more particularly, the legislative record) and the “plain language” of a statute in interpreting a law passed by Congress. This is fundamental for two reasons:

a)      The Court must come to a proper understanding of any statute before considering whether to strike it down.

b)      more than 60% of the Court’s docket each year is dealing strictly with statutory interpretation.

However, some political scientists suggest that these “legal” arguments are nothing more than a ruse. In reality, Justices vote according to their policy preference. Thus, political scientists studying SCOTUS have been at odds with each other – how much of the Court’s decisions are based on “law” and how much of their decisions are based on “policy preferences”?

While attitudinalists provide ample evidence for their inferences, they have yet to go to the two most fundamental aspects of legal thinking – text and intent. This study aims to get at the heart of the Justice’s behavior by studying statutory interpretation.

I created three variables that I coded for:

  1. “Plain Language”: When does a Justice in an opinion attempt to understand the statute via the “plain language”?
  2. “Congressional Intent”: When does a Justice attempt to divine the “intent” of Congress?
  3. “Legislative Record”: When does a Justice attempt to divine the intent of Congress by dipping into the legislative record?

I limited my study to cases from the 1986 term to the 2000 term, and examined only cases that dealt with a question of statutory interpretation.

The Plain Language variable has a “constraining” effect on the policy attitudes of Justices. As an example, when Stevens and Scalia do not use this judicial tool, their opinions are skewed towards the policy direction the attitudinal model predicts. However, when they use the “plain language” variable, both move towards the middle – Stevens becomes less “liberal” and Scalia becomes less “conservative”. The intent variables however have little effect.

These variables have a negligible effect on Majority opinions, but this is a result of the opinions already having a normal distribution (about 50/50 liberal conservative). Yet when we examine dissenting opinions, we see that Plain language once again has a “constraining” effect (taking the liberal skew towards the middle). We also see (yet again) that use of the legislative record has a negligible impact. In sum, the plain language variable (when used) tends to mitigate, or at least “constrain” the attitudinal skew.

While the attitudes of Justices certainly have a measureable influence, so too do legal variables (particularly plain language). However, caution must be exercised on two points:

1. We cannot necessarily conclude that the legal variables for intent (intent and intent with record use) are meaningless constraints. A problem with measuring these legal aspects empirically is that they do not measure “good” uses and “bad” uses of a certain type of judicial tool or legal reasoning. Thus, it might be that the legislative record is an effective constraint if Justices are only referring to certain aspects of the legislative record (such as committee reports and not floor remarks). This coding scheme would not pick up on that.

2. The chicken and the egg problem still exists. Which is primary and comes first – the attitudes of the Justices, or the legal variables that constrain them? One could still argue that Justices are attitudinal actors who are (to a degree) constrained by law. Or one could argue that Justices are legal actors, who decide on the basis of policy only when the law is ambiguous