WASHINGTON, D.C. — The principle of religious freedom is important to most Americans. But as President Donald Trump touts his support for it during his reelection bid, there are notable fault lines among people of different faiths and political ideologies over what it truly means.

About 8 in 10 Americans said religious freedom issues are at least somewhat important to them, with 55 percent saying they are very important, according to a newly released poll conducted by The University of Chicago Divinity School and The Associated Press-NORC Center for Public Affairs Research.

While 35 percent of U.S. adults overall said they believe their own religious freedom is threatened at least somewhat, conservatives were more likely than liberals to say so. Across the largest U.S. religious denominations, evangelical Protestants were especially likely to perceive risks to their freedom to worship.

A divide over religious freedom for Muslims was particularly apparent in the poll. About half of Americans said U.S. Muslims’ freedom to worship is threatened at least somewhat, including 7 in 10 atheists and agnostics, about 6 in 10 Catholics, and about half of white mainline and nonwhite Protestants. Only about 3 in 10 white evangelicals, however, said Muslim Americans’ religious freedom is at least somewhat threatened.

The poll’s findings suggest that, as Trump leans into religious freedom as a touchstone of his outreach to devout voters, conservatives and evangelicals who make up the core of his base hold distinctly different perspectives on the topic than other Americans. Whether those divergent views can be joined and harnessed to make progress on the issue for a wide variety of faiths remains an open question.

“No one’s religious freedom is an island, and if the government is empowered to take religious freedom away from Muslims or other religious minorities, the government is going to be empowered to take away religious freedom from other religious groups,” said Luke Goodrich, senior counsel at the nonpartisan Becket Fund for Religious Liberty.

Becket has racked up legal victories for Christian clients, including last month’s Supreme Court ruling that upheld the ability of religious schools and other institutions to make certain employment decisions without facing discrimination claims.

But the nonprofit firm represents members of all faiths, and Goodrich said “it’s vitally important, for the health of religious freedom over the long term, that Americans care for religious freedom for those with whom they deeply disagree.”

In the poll, 77 percent of liberals said Muslims’ freedoms were threatened at least somewhat, compared with 32 percent of conservatives. Liberals also were roughly three times more likely than conservatives to perceive threats to atheists and Buddhists, and somewhat more likely to perceive threats to Jews, 56 percent to 41 percent.

By contrast, by roughly two to one, more conservatives than liberals said evangelical Protestants, Catholics and other Christians face threats to their religious freedom.

Andrew Lewis, an associate political science professor at the University of Cincinnati who focuses on religion, said a “polarization of religious freedom” has developed over the last two decades, with potentially negative consequences.

“Religious freedom has always been a contested question in America,” Lewis said, “but the fact that it’s entered into our politically partisan landscape is bad for thinking about how we protect the rights of people.”

The poll was conducted in February, before the coronavirus wreaked havoc in the country, but has not been made public until now. And some of its findings have resonated more sharply since the pandemic began.

For example, 46 percent of evangelicals said their religious freedom was under threat, compared with 27 percent of mainline Protestants, 36 percent of Catholics and 40 percent of Americans affiliated with other religions.

Another UChicago Divinity School/AP-NORC poll conducted in the early weeks of the pandemic found white evangelicals were particularly likely to say in-person worship should be allowed in some form, and 46 percent said that barring those services would violate religious freedom.

David Nirenberg, dean of The University of Chicago Divinity School, pointed to “communications bubbles in certain faith communities” that often reinforce to believers the assertion that their religious liberty is imperiled.

“We’re seeing some of that now mobilized again in the public health response” to the virus, Nirenberg added, including restrictions on in-person worship and other behavior that are “represented in some of those same communities as an attack on their religious liberties.”

Trump’s reelection campaign has appealed to evangelicals in part by seeking to portray Democrats as opponents of religious freedom, with surrogates often citing state and local restrictions on in-person worship during the pandemic.

Religious freedom lawyer Asma Uddin connected the poll’s finding that evangelicals are less likely than other groups to believe Muslims’ freedom to worship is threatened to their broader perspective on religious freedom.

Evangelicals can view religious freedom “essentially as a shield to protect the ‘in’ group,” added Uddin, a fellow at the Aspen Institute think tank. As a result, it “becomes something you can’t defend for the ‘out’ group.”

However, the poll’s findings of a gulf in perceptions about religious freedom wasn’t limited to evangelicals. About half of atheists and agnostics said evangelical Protestants’ claims to religious freedom threaten others’ rights at least somewhat, and about 4 in 10 said the same about other Christians.

By comparison, 16 percent of Protestants perceived evangelicals and other Christians as threatening the rights of others with their claims.

Roughly comparable shares across religious affiliations — about 4 in 10 — said Jews’ freedom of religion is being threatened at least somewhat. Similarly, roughly comparable shares across religious affiliations — all under 2 in 10 — said Jews’ claims were threatening to others’ rights.

Religious groups that make up a small share of the U.S. population could not be analyzed separately in the poll because of sample size.

The AP-NORC poll of 1,015 adults was conducted Feb. 13-16 using a sample drawn from NORC’s probability-based AmeriSpeak Panel, which is designed to be representative of the U.S. population. The margin of sampling error for all respondents is plus or minus 4.2 percentage points.

Schor reported from New York.