The Roberts court is often treated as especially protective of the First Amendment. It is true that the court has reshaped free speech law across such areas as campaign finance, student speech, government speech, and online platforms. It has also transformed religious doctrine through expanding the free exercise clause and the establishment clause, and in several religious-accommodation cases.
But these two First Amendment stories are not the same. The court’s speech cases are often favorable towards speakers, but speech claimants do not win uniformly. Religion is more consistent. Across the Roberts court, religious claimants have done unusually well. That pattern predates Justice Amy Coney Barrett, but it has become sharper since she joined the court. And when cases sit at the boundary between speech and religion, the claimant often fares especially well.
The takeaway is that the Roberts court, particularly post-Barrett, is not simply pro–First Amendment. It is more precise to say that the court is often pro-speech, strongly pro-religion, and most consistent when speech overlaps with religious identity or conscience. That distinction is fundamental for understanding the court’s current direction, including its decision to hear the case of St. Mary Catholic Parish v. Roy, which challenges long-standing religious precedent.
The basic pattern
In seeking to understand the Roberts court’s First Amendment jurisprudence, I separated cases into three categories.
First are formal religious cases, which include free exercise, establishment clause, Religious Freedom Restoration Act, Religious Land Use and Institutionalized Persons Act, ministerial-exception (a doctrine which bars ministers from suing churches and other religious institutions for employment discrimination), religious-accommodation, and religious public-benefit cases.
Second are “pure” speech cases, which include campaign finance, retaliation, government speech, student speech, commercial speech, public-employee speech, donor disclosure, trademark, compelled-speech, and online platform cases.
Finally are what I call religion-adjacent speech cases, or speech cases that involve religious speakers, religious institutions, conscience-based claims, or religiously inflected expression; that is, speech claims connected to religious identities or religious institutions.
Religious claimants triumphant
The win-rate contrast is the article’s cleanest empirical point. Before Barrett joined the court, from 2005-2019, religious claimants or parties won 10 of 12, or 83%, of formal religious cases – a very high, although not perfect, percentage of cases. (If one does not count Zubik v. Burwell as a religious case, in which certain religious institutions challenged the Affordable Care Act’s birth control provisions, this number rises to 10 of 11, or 91%.) In the Barrett era, formal religious claimants are six for six.
The broader religion-related category is even stronger. When religion-adjacent speech cases are included, pre-Barrett claimants won 12 of 15, or 80%, of such cases. In the Barrett era, they have won 10 of 10 of these.
By contrast, speech-primary claimants won 25 of 46, or 54%, pre-Barrett cases and 10 of 19, or 53%, Barrett-era cases. Pure speech cases, excluding religion-adjacent disputes, are even more mixed: 23 of 43, or 53%, pre-Barrett claimant wins, and 6 of 15, or 40%, Barrett-era claimant wins.
These numbers show a court whose speech jurisprudence is thus a good deal more conditional than its religion jurisprudence.
Religion as the more consistent track
As noted above, the majority’s siding with religion-related claimants began before Barrett joined the court. In the 2006 case of Gonzales v. O Centro, the court protected sacramental religious practices, including the use of hallucinogens, under RFRA. In 2012’s Hosanna-Tabor v. EEOC, it recognized the ministerial exception, protecting religious employers from employment discrimination suits. In the 2014 case of Burwell v. Hobby Lobby, it applied RFRA to closely held corporations objecting to the Affordable Care Act’s contraceptive mandate. In 2015’s Holt v. Hobbs, it protected prisoner religious exercise under RLUIPA. In the 2017 case of Trinity Lutheran v. Comer, the court held that a church could not be excluded from a public-benefit program because of its religious status. Espinoza v. Montana Department of Revenue then extended that logic to religious schools, and Our Lady of Guadalupe v. Morrisey-Berru strengthened religious institutional autonomy.
The Barrett-era cases make the pattern more categorical. 2020’s Tanzin v. Tanvir allowed damages under RFRA against federal officials for violating persons’ free exercise rights. Ramirez v. Collier, decided in 2021, protected religious touch and prayer in the execution chamber. 2022’s Carson v. Makin protected religious schools from exclusion when it came to state tuition assistance. That same term, Kennedy v. Bremerton School District protected a coach’s ability to pray at a public school event. Catholic Charities Bureau v. Wisconsin Labor and Industry Review Commission, decided last term, rejected a state’s narrow definition of when an entity operates for “religious purposes.” And Mahmoud v. Taylor, also decided last term, sided with religious parents challenging a school policy that denied opt-outs for readings of “LGBT+-inclusive storybooks.”
This shift may be described as a move from “accommodation” to “priority.” Accommodation implies limited carve-outs from general rules; that is, drawing narrow lines to protect religious institutions but not reconsidering the rules themselves. Priority better captures the current pattern: when religious identity, religious institutional autonomy, or religious participation in public programs is at stake, the court increasingly supports the legal framework of the religious claimant’s position, questioning the rules themselves rather than simply how they have been applied.
One important feature of these cases is that the Roberts court’s First Amendment shift has not always (or even often) occurred through formally overruling precedent. In the case coding (by the Supreme Court Database) relatively few decisions are marked as openly altering existing case law. The court has more often changed First Amendment law by invalidating statutes, narrowing old tests, or reclassifying the dispute. The result is a court that transforms the field by deciding anew what counts as religious accommodation or exercise that is to be protected from state interference.
Speech is protective but context-dependent
Speech is the larger set of cases, but it has been less uniformly protected. The Roberts court has issued major pro-speech rulings in campaign finance and political spending cases, such as (most famously) Citizens United v. FEC and McCutheon v. FEC. It has also protected donor privacy in Americans for Prosperity Foundation v. Bonta; offensive or controversial expression in Snyder v. United States, Matal v. Tam, and Ianca v. Brunetti; claimants that claimed they were retaliated against for their speech in Heffernan v. City of Paterson, Lozman v. City of Riviera Beach, NRA v. Vullo, and Gonzalez v. Trevino; and student off-campus speech in Mahanoy Area School District v. B.L.
But the court has also rejected many speech claims in a range of areas. Garcetti v. Ceballos limited public-employee speech. Morse v. Frederick upheld discipline for student speech. Beard v. Banks upheld prison restrictions. Pleasant Grove v. Summum and Walker v. Texas Division treated contested expression as government speech, allowing this to be regulated. Williams-Yulee v. The Florida Bar upheld a judicial-campaign solicitation rule. Houston Community College v. Wilson rejected a First Amendment claim based on official censure. City of Austin v. Reagan National Advertising upheld a sign regulation. United States v. Hansen kept an immigration-related statute in place criminalizing “encourag[ing] or induc[ing]” illegal immigration (after narrowing it). Vidal v. Elster upheld a trademark-registration limit. Free Speech Coalition v. Paxton upheld Texas’ adult-content age-verification law. And TikTok v. Garland upheld the federal TikTok divestiture statute against a First Amendment challenge.
These decisions make sense given the court’s current trajectory. The majority of justices are especially skeptical of viewpoint discrimination (that is, bans on a particular opinion or perspective), campaign-finance restrictions, compelled speech, donor-disclosure burdens, and retaliation based on one’s speech. But it is more deferential when the government is an employer, educator, prison administrator, program manager, regulator, or acting on the basis of national security.
That is the key contrast with religion. Speech claimants win often, but their success depends heavily on context. Religious claimants, by contrast, have prevailed across a much wider range of institutional settings.
Barrett, religion-adjacent speech, and the new priority
Of course, Barrett did not create the Roberts court’s religion trajectory. As described, the pre-Barrett court had already decided many major cases for religious claimants. But Barrett’s arrival gave the court a stable six-justice conservative majority, making religion-protective outcomes even more secure in contested cases. Indeed, per term, formal religion cases rose from about 0.80 before Barrett to about 1.00 after Barrett. Religion-adjacent speech cases rose from about 0.20 per term to about 0.67 after Barrett. And the combined religion and religion-adjacent category rose from about 1.00 case per term to about 1.67 after Barrett.
The new majority obviously matters in cases like Carson, Kennedy, and Mahmoud, where the court divided 6-3. But the story extends beyond bloc-based voting. Some religion or religion-adjacent wins have been unanimous or near-unanimous. The result is a court whose religion-protective direction is sometimes ideologically divided and sometimes broadly shared.
St. Mary and the future of Smith
That brings the story to St. Mary Catholic Parish v. Roy. That case concerns Catholic preschools excluded from Colorado’s universal preschool program because they follow Catholic teachings in admissions and related policies. The grant is important because it gives the court another opportunity to address the reach of 1990’s Employment Division v. Smith, a much-debated decision that held the First Amendment’s free exercise clause is not generally violated when a challenged law is not generally applicable; that is, it does not target a specific faith group or religious practice.
The court did not grant review on the broadest question: whether Smith should be overruled. That might seem like restraint. But, as explained, in the Roberts court’s religion jurisprudence, formal restraint often coexists with practical transformation.
That is why St. Mary matters even if Smith survives. For example, the court could rule that Colorado’s program is not generally applicable because it contains exemptions or discretionary features. It could treat the exclusion as discrimination against religious participation involving a public benefit. Or it could blend those approaches. But any of those paths would leave Smith formally standing while making it less important in practice.
And that outcome would fit the Roberts court’s broader First Amendment jurisprudence. Its protection of speech remains significant, but uneven. Religion has become more consistent, more prioritized, and increasingly central to the court’s constitutional identity. Since Barrett joined the court, formal religion claimants and religion-adjacent speakers have prevailed with remarkable regularity. The court may not describe this as a religious-liberty revolution. But the data points only in that direction.
The future of the First Amendment is therefore likely to remain bifurcated. Speech claimants will continue to win in important areas, especially where the court sees censorship, retaliation, compelled speech, or viewpoint discrimination. Yet religious claimants will remain among the most favored litigants before the court, particularly when public benefits, religious schools, parental rights, religious conscience, and institutional autonomy are involved. St. Mary may not kill Smith. But it will likely further demonstrate why the court no longer needs to.
