Can Supreme Court Convince Americans? — An Experiment on Court Reasoning

• In Event: Court Opinion Language and Influence

Abstract

The Supreme Court justices are traditionally expected to interpret the Constitution faithfully regardless of the partisanship of the president who appointed them. According to this expectation, when the nine justices vote behind closed doors, a binding ruling arises because its compelling legal reasoning convinces a majority of the justices to support it. But can such reasoning actually convince regular Americans? How does information about the reasoning provided by these rulings change Americans’ stance on the legal issue in question? How do voters’ demographics interact with such an intervention? Previous research on the legitimacy-conferring function of the Court focuses on the causality between the Court’s mere endorsement of a policy stance and greater public support for that stance. Building on this literature, I use a two-by-two factorial experimental design to explore whether priming respondents with not only the Court’s mere endorsement of a policy stance but also its reasoning changes their political opinions.

I will deploy a pre-registered survey experiment on a sample of 1,400 respondents on Lucid Fulcrum Exchange, which provides nationally representative samples that have similar demographic makeup to that of the American public. The first part of the survey is non-experimental and elicits the demographic background of the respondents, including their age, gender, race, sexual orientation, party identification, ideology, education level, and socioeconomic status. The second, experimental part of the survey involves fact patterns on four recent Supreme Court rulings. The existing political science literature on the legitimacy-conferring function of the Supreme Court has provided little insight on which cases to select; in my design, I seek to include four distinct and important fields of debates in law that also involve issues common and controversial in daily political life, so that the respondents can meaningfully engage with these judicial decisions. These cases cover the areas of criminal law, corporate law, personal privacy rights, and discrimination law.

The respondents will be randomly assigned into one of four treatment conditions upon the completion of demographic information. In the control group, the respondent will read a simple policy statement summarizing the position that the Court took in the particular case, with no mentioning of the Court whatsoever; in the second condition, there is a policy statement and the Court’s endorsement of it, simply adding “the Supreme Court ruled recently” after the policy statement; in the third condition, there is a policy statement and the Court’s reasoning in support of the policy. The reasoning appears in the form of a one-paragraph brief in lay language. It does not include any mention of the Court itself, but rather solely focuses on summarizing the reasoning that the Court provided in its majority opinion that became a binding precedent; Finally, the fourth condition presents the policy statement followed by the Court’s endorsement and reasoning. Simple randomization is adopted since the respondents will not be randomized until they finish non-experimental questions on their demographics first. Respondents will be assigned to the same treatment condition across the four cases to avoid interference and speculation of the project’s intent. After each condition is presented, the respondents will indicate their level of support for the particular policy statement. This factorial design will allow me to compare the legitimacy-conferring function of the Court through three different channels — mere endorsement, mere reasoning, or the combination of both — and discern which of them is the most influential in swaying public opinion.

Since the Court’s reasoning is so crucial to the substantive contribution of this project, it warrants extra caution to draft. I recruited current students and recent graduates from Harvard Law School to succinctly summarize the Court’s majority opinion in these four cases into one-paragraph briefs of the reasoning in lay language. They will also proofread each other’s briefs in an attempt to make sure that the briefs represent the Court’s reasoning as accurately and comprehensively as possible, albeit in a shortened form.

Prior to deployment of the experiment onto the 1,400-respondent sample, I will first use a pilot test on 100 respondents for a few reasons. I will test the legibility of these briefs and whether the respondents can infer the true intent of the project. More importantly, to maintain symmetry across the factorial design, the fourth condition involves a policy statement followed by the Court’s endorsement and reasoning brief, but I will not explicitly state that the brief is what the Court claimed as the reason behind their ruling. It is not a far fetch that the respondents will simply presume that this is the case given the way this question is set up, but this will be aptly tested by the pilot study.

Author

• Yiwei Tang, Harvard University