Are Christians More Likely to Invoke Rfra—and Win—than Other Religions Since Hobby Lobby

• Jurisdiction: United States,Federal
• Publication year: 2021
• Citation: Vol. 72 No. 2

Are Christians More Likely to Invoke RFRA—And Win—Than Other Religions Since Hobby Lobby

Stephen Cranney

[Page 585]

ARE CHRISTIANS MORE LIKELY TO INVOKE RFRA—AND WIN—THAN OTHER RELIGIONS SINCE HOBBY LOBBY?

by Stephen Cranney^*

I. Introduction

Religious freedom has become a flashpoint in the culture wars, and religious freedom rhetoric is often framed differently by different political sides. Political conservatives are typically perceived as emphasizing Judeo-Christian institutions; progressives are typically perceived as emphasizing minority faiths such as Islamic and Native American religious practices. This framing, in turn, affects perceptions of religious freedom legislation.

Arguably, the most significant piece of religious freedom legislation in recent history is the Religious Freedom Restoration Act (RFRA),¹ a federal law that heightens the legal standard required for the federal government to interfere with the free exercise of religion. This fight over the meaning and purpose of religious freedom in the United States has extended to RFRA, and much academic and popular discourse has framed RFRA as a tool of Christian institutions trying to be granted special privileges.² However, the actual effect of RFRA is an empirical question, as opposed to its initial perception. Has RFRA been primarily used as a legal tool of majority Christian institutions? Or should it be

[Page 586]

conceptualized as a historically useful asset for smaller, less popular religious movements that lack the level of sociocultural support afforded to Christians in the United States?

Despite the relevance of this empirical question to the framing of legislation at the crux of current hot-button political and ideological debates, very little research has empirically investigated who is bringing and winning RFRA claims in court. However, some prior research has investigated the question of who is raising and winning religious liberty cases generally. Qualitatively, a recent monograph tracks the history of Native American religious liberty court cases and rhetorical strategy.³ The RFRA is particularly relevant to these religious communities. More quantitatively, an early analysis of state and federal free exercise cases from 1946-1956 and 1970-1980 found that minority faiths were responsible for a plurality of such cases, with mainline Protestants and Catholics only accounting for 8% of claims from 1970-1980.⁴

Similar results were found by another paper examining the 1981-1997 timeframe that found minority religious groups (including marginal Protestant fundamentalist groups) constituted 16% of the religiously identified population in America. During that timeframe, the minority religious groups brought more than half of all free exercise claims in the courts between 1981 and 1997.⁵ Finally, in the 1990-2015 window, a 2017 study found that only 33% of free exercise decisions made at the federal level, using the strict scrutiny standard, involved Christian claimants, with Muslims constituting about 16% and Native Americans constituting 17% of claims.⁶ Finally, Goodrich & Busick mined Tenth Circuit court decisions from 2012-2017 and extracted 118 relevant religious liberty cases.⁷ From this sample they found that Christian groups are proportionally underrepresented in religious freedom cases.⁸ Indeed, when contraceptive mandate cases are excluded (as there was a spat of them during this time because of the Affordable

[Page 587]

Care Act's requirements), Christian cases were not even a plurality, and tied for second with Muslim cases.⁹

While prior research has consistently found that historically religious minorities are proportionately more likely to rely on religious freedom protections in the courts, the literature on whether they are more likely to be successful is more ambiguous. Some of this variation undoubtedly stems from variations in definitions of what constitutes a minority or marginal faith, as pairwise comparisons between religious traditions vary.¹⁰ Different groups have enjoyed different successes across time, and the results are sensitive to which groups are placed together, and which groups act as the comparative reference. It is likely that religious affiliation tracks with distinct legal issues, and as such, any investigation of differential successes by religious groups in the court is incomplete without acknowledging and addressing the intersection of statutory particulars, germane legal issues, and the legal needs of particular faiths. Contraception cases are not generally brought by Sikhs, and prison facial hair cases are not generally brought by Catholics, as the theological particulars of their respective faiths do not bring them into conflict with the law on these issues. Despite this fact, there is a lack of research about particular statutes and how they may interact with religious particulars.

[Page 588]

Abrams helped fill this lacuna by systematically collecting a dataset of all post-Hobby Lobby¹¹ federal district court cases that made a merits decision on a RFRA issue. Abrams then applied multivariate logistic regression analysis in order to test predictors of success, finding in the analysis that Christians are more likely to win RFRA cases.¹²

Here, I reanalyzed Abram's data, and did not find evidence of a bias towards Christianity, but found strong evidence of the negative secular effect that Abrams discovered. The discrepancies found are probably attributable to lack of clarity in variable operationalization, statistical overfitting, or too many variables per observation. Additionally, I investigate not only who is winning RFRA cases, but who is invoking RFRA relative to their proportion of the United States population, and what types of cases (such as prisoner and pro se) cluster with different religious traditions.

II. METHODOLOGY

A. Categorizing Religious Groups

Because of the small cell sizes involved for some religions, different religious groups are sometimes combined in order to create a sample size large enough to perform statistical comparisons. As noted above, these risks overlooking potentially important differences, but the reality of the sample sizes leave little choice. It is difficult to draw a generalized conclusion about Amish success when there has only been one case that has met the criteria for inclusion. However, when possible, different permutations are used in order to show the robustness of results to alternative categorizations.

This fraught categorization issue did make replicating Abrams' results difficult. In the regression, Abrams categorized religions into the groups of Christian, Catholic, Jewish, Muslim, Rastafarian, Unspecified, Secular, and Sikh.¹³ While Abrams was not clear about what categories in the original data fit into the categories delineated in the paper, based on my replication of the summary statistics it appears that the "non-Catholic Christian" group referred to in the paper (12.2% of the sample) only included generic Christians, and did not include Baptists (2% of the sample) or Pentecostals (1%).¹⁴ Additionally,

[Page 589]

throughout the paper "Christian" is conflated with "non-Catholic Christian."¹⁵ This becomes theoretically problematic because while the non-Catholic Christian coefficient is significant in the logistic model, the Catholic coefficient is null. This mitigates against Abrams' theory that the Christian effect is attributable to contraception cases, since that effect would presumably also be seen among Catholic litigants (indeed, probably more so).

B. Replication of Results

Despite trying several iterations of their model, I was unable to replicate Abrams' significant finding for Christian groups (in most iterations, I was...