“Unpresidented!” or: What happens when the president attacks the federal judiciary on Twitter

ABSTRACT

There are many things that make Donald Trump a controversial figure in American politics, not least of which is his prominent and unorthodox use of Twitter to attack his opponents. The federal judiciary is a frequent target of the president’s ire: during his campaign and tenure as president, Trump has not hid his disdain for federal judges who disagree with him, and has on occasion taken to Twitter to express these feelings. This unprecedented form of going public provides the opportunity to shed light on the malleability of support for politically insulated institutions. We evaluate the hypothesis that Trump’s statements erode specific and diffuse support for the federal judiciary using a survey experiment in which real tweets written by the president are manipulated as the experimental treatment. We find that specifice, but not diffuse, support for the Court changes in response to his attacks, but those changes are conditioned on respondents' preexisting support for democratic values.

KEYWORDS:

• Supreme court
• Trump
• Twitter
• public opinion
• experiments

Notes

1. This argument is bolstered by findings from research on authoritarianism. Authoritarianism is not the same as having low support for democratic norms, but the concepts are similar. As Hetherington and Suhay (2011) define the concept, “authoritarianism is, at root, a need for conformity to group norms” (547). Whereas democratic norms emphasize tolerance, the rule of law, and individual liberty, authoritarians prize conformity. Authoritarianism deemphasizes the individuality that democratic norms seek to protect. The overlap between support for democratic norms and authoritarianism suggests that observable behavior of authoritarians, such as deference to an executive and skepticism toward federal courts, who protect the civil liberties and rights of outsiders that authoritarians abhor, might also be exhibited by those with low support for norms.

2. Undoubtedly, some students of American politics will recognize our negative manipulation as depicting an attack on a federal district judge, while our dependent variables are industry-standard measures of attitudes about the Supreme Court. We think that most respondents will perceive the tweet as a generalized attack on judges who overstep the judicial role by encroaching on the authority of law enforcement. The tweet does not mention Judge Robart by name or official position, and general public ignorance of the identities of Supreme Court justices has been well documented (but see Gibson & Caldeira, 2009b). In any event, we believe that the presence of respondents in our sample who are unlikely to base an evaluation of the Supreme Court on negative presidential messaging because of preexisting knowledge about the context of the tweet would make our hypothesis tests conservative ones.

3. Two other items, dealing with whether the Court can be trusted to make the right decisions and whether the Court should have the right to interpret the Constitution, were asked but dropped from this factor analysis. When included, the loadings were weak – both less than.2. Furthermore, each of these two items share little variance with the four items that were ultimately included ($R^2<.06$ in both instances).

4. Brant tests support the proportional odds assumptions for each model presented in Table 4. For Model I, ${\chi }^2=29.75$ and $p=.12$; for Model II, ${\chi }^2=32.85$ and $p=.17$; and for Model III, ${\chi }^2=31.06$ and $p=.23$. Additionally, the results of the ordered logit analysis presented here are substantively similar to those when conducting an OLS regression.

Additional information

Notes on contributors

Christopher D. Kromphardt

Christopher D. Kromphardt is a Lecturer in the Doctor of Law and Policy Program at Northeastern University.

Michael F. Salamone

Michael F. Salamone is an Associate Professor of Political Science in the School of Politics, Philosophy, & Public Affairs at Washington State University.