Snap Judgment: Implicit Perceptions of a (Political) Court

Baird and Gangl (2006) and Ramirez (2008) utilize decision-specific analogs to this approach and ask subjects whether particular Court decisions were reached in a nonpolitical manner.

We do not differentiate between “high” and “low” politics (Gibson and Caldeira 2011).

“Good” terms include “marvelous,” “superb,” “fantastic,” “glorious,” and “wonderful.” The bad terms are “horrible,” “terrible,” “awful,” “unpleasant,” and “dreadful”.

IAT studies frequently rely on online samples (e.g., Banks and Hicks 2016; Trawalter et al. 2012). Project Implicit (implicit.harvard.edu/implicit) has collected an enormous amount of data using online participants. The MTurk survey was conducted on December 9 and 10, 2014. There were no particularly salient Court cases decided or argued near these dates, which decreases any concern that these data were influenced by a short-term reaction to a Court decision. Respondents who completed our studies were paid $1.25.

For example, 27% of the sample are 40 years old or older, 24% identify as Republican (59% identify as Democrats), 50% are women, 22% are non-white (8% African American and 5% Hispanic/Latino), and 53% do not have a four-year college degree.

Since only approximately 1/4 of our participants (n = 372) received both the Court-Congress politics IAT and the Court-Congress, good-bad IAT, including Implicit Court Preference in our statistical models leads to a substantial decline in sample size and statistical power. We thus exclude this variable from the models reported here. Auxiliary analyses suggest that Implicit Court Preference does not affect either diffuse or specific support. Adding Implicit Court Preference to the models in which Implicit Perception is included, if anything, increases the absolute value of the coefficient for Implicit Perception but also increases the size of the standard error, due to the dramatic reduction in sample size. See the online appendix for details.

Specifically, respondents were asked the extent to which they agree (on a five-point scale) with each of the following statements: (1) “When deciding cases, justices base their decisions on the law and the Constitution,” (2) “When deciding cases, the justices base their decisions on their own political views,” and (3) “The justices’ party affiliations have little to do with how they decide cases before the Supreme Court.” Gibson and Caldeira (2011) also ask whether justices are just “politicians in robes” and claim that this question taps perceptions of politicization, as compared to perceptions of legal realism. It is also possible, however, that this question differs primarily in its particularly negative valence and we thus excluded it from the battery.

Cronbach’s alpha for this scale is a relatively unimpressive 0.652, which is not surprising given that Gibson and Caldeira (2011, footnote 21) report that there may be more than the usual measurement error associated with the items, as they are “demanding for ordinary people.”

Court Knowledge is the total number of correct answers to the following three questions: (1) “Who is the current Chief Justice of the U.S. Supreme Court?”; (2) “Can the U.S. Supreme Court declare an act of Congress unconstitutional?”; and (3) “How are Supreme Court justices selected?” Education is measured on a six-point scale with higher values representing higher levels of educational attainment. Democrat and Republican are dummy variables that equal one if the respondent identifies with the party or is an Independent that leans toward one of the two parties. So-called “pure” Independents serve as the baseline. Ideology is measured on a seven-point scale with positive (negative) values for conservatives (liberals). Respondents who do not place themselves ideologically are coded as zero (i.e., moderate) but are also indicated by a dummy variable which is included in all the models. Income is measured on an 11-point scale, with those who decline to answer assigned to the lowest category and indicated with a dummy variable which is also included in all the models. White and Female equal one for respondents who identify as such. Age is measured in years.

As is customary in the literature, we use self-reported, explicit attitudes. There may be ways to capture implicit attitudes for these items, but we are interested in exploring whether implicit perceptions of the political nature of the Court affect diffuse and specific support, two dependent variables of long-standing importance to scholars and practitioners.

Specifically, we ask respondents whether they agree (on a five-point scale) with (1) removing the Court’s right to hear certain controversial types of case, (2) removing judges who rule at odds with what the public wants, (3) the statement that the Court is too independent and needs to be reined in, and (4) that the power of judicial review should be eliminated.

For Citizens United, respondents were told that “The U.S. Supreme Court recently decided that corporations and unions can spend as much money as they want to help political candidates win elections.” The summary for Graham is: “The U.S. Supreme Court recently decided that juveniles cannot be sentenced to life in prison without parole for any crime other than murder.”