The Relationship of Justices’ Religious and Party Affiliation, Case-Level Factors and Decisional Era to Voting in Establishment Clause Disputes Involving Public Education, at the United States Supreme Court: 1947-2012

Abstract

We applied logistic regression analyses to the votes cast in 53 decisions involving Establishment Clause disputes in public education rendered by the United States Supreme Court between 1947 and 2012. A binary dependent measure, individual justices’ votes, was selected. The model was set up with two justice-level independent variables [party affiliation and religion] and three case-level independent variables [issue salience, lower court dissent, and inter-court conflict] and a decisional era independent variable [Reagan and later versus pre-Reagan era]. When the Republican and Democratic data bases were combined, the results revealed that with all other variables controlled for the entire period: the odds of justices appointed by a Republican president voting in conservative pro-religion direction were greater than justices appointed by a Democratic president; the odds of Protestant and Catholic justices voting in a conservative pro-religion direction were greater than for Jewish justices; and the odds of Protestant and Catholic justices voting in a conservative pro-religion direction did not differ from one another. The odds of the justices voting in a conservative pro-religion direction during the Reagan and later years were greater than the justices voting in that direction during the pre-Reagan era. Separate examinations of the relationship of the case-level predictors within the Republican, Democratic, Protestant and Catholic justice-groups revealed important distinctions in how votes were cast. Justices nominated by Republican presidents consistently voted in a pro-religion direction, regardless of the judicial era which was studied. The results for justices nominated by Democratic presidents were more subtle. They revealed the odds of Protestant and Catholic justices voting in a conservative pro-religion direction were significantly greater than that of Jewish justices, but the odds of Protestant and Catholic justices voting in a conservative pro-religion direction did not differ from each other within this group. During the Reagan and later years, the odds of Democratic justices voting in a pro-religion direction were greater than during the pre-Reagan era. Analysis of the Protestant justices revealed that the odds of voting in a pro-religion direction were greater for those justices nominated by Republican than by Democratic presidents. The odds of Protestant justices voting in a pro-religion direction were higher in low salience cases than in high-salience ones. The odds of Catholic justices voting in a pro-religion direction were greater during the Reagan and later years, compared to the pre-Reagan era. These and other results were interpreted in terms of the attitudinal and/or legal models in explaining voting at the Court. Because of our findings, we recommend the use of category specific investigations in order to avoid the risk of erroneous conclusions when large undifferentiated data bases are used, with the caveat that this approach may limit the generalizability of the findings. Among the advantages to our approach is the fact that researchers and practitioners are generally interested in answers to questions about specific conflict categories and it may lead to more accurate predictions.