Ryan Burge published a post on September 14, 2026, that I have been waiting years for someone to write. It is called “The Religious Behavior Index: How Engaged Are Americans, Really?” and my first reaction on reading it was one word: "finally".
People most at risk of violence for their faith are... the ones who do something with their religion.
Here is why a survey index of American churchgoing matters to anyone who cares about religious freedom. Since undertaking my doctoral fieldwork in Mexico I have been arguing that the people most at risk of violence for their faith are usually not the ones who belong to a religious minority but the ones who do something with their religion.
Furthermore, because nobody measures what people do, the violence against them goes unrecorded, unexplained, and unaddressed. Burge has just shown that the doing can be measured. That is the first half of a problem I will come back to in a second post. This one is about why the distinction matters and what it costs to get it wrong.
For two decades the study of religion has talked about “lived religion,” the idea that what matters is not the box people tick on a census but what they actually do with their faith. The concept is everywhere in the literature. Turning it into something you can measure has been slower.
Combined measures of religiosity are not new: Pew researchers have long used a religious commitment index that folds attendance and prayer together with how important people say religion is to them, and the World Values Survey has measured religious participation for decades. But those instruments mix behavior with belief and self-assessment, and the behavior they do capture is almost always attendance.
Burge’s index is made of behavior and nothing else.
Burge’s index is made of behavior and nothing else, it reaches past attendance into prayer groups, scripture reading, and interfaith dialogue, and it does so on a sample large enough to slice by tradition, age, and region. Using the Pew Religious Landscape Study, a survey of roughly 37,000 Americans, he combines six behaviors into a single scale from 0 to 10: weekly in-person attendance, online attendance, personal prayer, participation in a prayer or scripture group, reading scripture or devotional material, and interfaith dialogue. A zero means none of it. A ten means all of it, at the highest frequency.
The list is not one anyone would design from scratch. Attendance appears twice, once in person and once online, and interfaith dialogue is a strange companion to the other five. But the six are the behaviors the Pew survey happens to ask about, and Burge is working with what exists.
Religion... a complex of practices rather than a set of beliefs.
Christian Smith, whose Religion: What It Is, How It Works, and Why It Matters (Princeton, 2017) defines religion as a complex of practices rather than a set of beliefs, works through a list of kinds of religious practice that runs to eighteen, and his point is that religious behavior is so varied that any list has to stop somewhere arbitrary. Six items will not exhaust lived religion. They are still six more than most surveys measure.
The results are sobering for anyone who still thinks of the United States as a religiously active country. The weighted mean score is 3.73. Only 3.7 percent of Americans score nine or above. The bottom fifth of the population scores between 0 and 0.6, which in practice means they do nothing at all. Jehovah’s Witnesses and Latter-day Saints top the table at around 6.6. Catholics and mainline Protestants sit at 3.8, barely above the national average. People who identify as “nothing in particular” average 1.5, and atheists and agnostics fall below 1.
None of this is surprising to people who study religion. What is new is that it can now be stated as a number, which means it can be compared, tracked over time, and, in principle, set alongside what else we know about the same people, including what happens to them. That last possibility is the one I care about.
Why identity is the wrong variable
Religious behavior, not religious identity, is what determines the vulnerability of religious groups to human rights abuses.
One point I have made repeatedly in my own work is that religious behavior, not religious identity, is what determines the vulnerability of religious groups to human rights abuses. The argument grew out of fieldwork rather than theory.
In my doctoral dissertation, published as The Specific Vulnerability of Religious Minorities, I applied a vulnerability assessment tool to three case studies: actively practicing Christians in the cartel-controlled states of northeast Mexico, converts to Protestantism among the Nasa indigenous people of southwest Colombia who broke with the traditional practices of their community, and Christians in Cuba.
Being a nominal Christian did not get anyone into trouble.
In all three, the finding was the same. Being a nominal Christian did not get anyone into trouble. Starting a football team to keep boys out of the cartels did. Refusing to take part in the rituals required by the indigenous council did. Criticizing the Communist government from the pulpit did. The vulnerability tracked what people did because of their convictions, not what they called themselves.
Christians who witness... draw the persecuting reaction, not the ones who keep their heads down.
Anyone who has worked in a hostile setting will find this obvious. Nik Ripken’s The Insanity of Obedience made the same point from the missionary side years ago: it is the Christians who witness who draw the persecuting reaction, not the ones who keep their heads down. What practitioners know from experience has never made it into the instruments that count victims, and that gap is the subject of this post.
A continuum, not a category
Think... of religion as a continuum running from passive identity to active behavior, with five categories along it.
In my dissertation, and later with Marlies Glasius in a 2022 article in Human Rights Quarterly, “Vulnerability and Active Religious Behavior: Christians and Crime Syndicates in Mexico” (open access copy here), I proposed thinking of religion as a continuum running from passive identity to active behavior, with five categories along it.
At the passive end sits religious self-identification, the census category: what people say they are when asked. Next comes religious participation, meaning institutionalized religious activity such as attending services or belonging to a congregation.
Then religious lifestyle, meaning behavioral choices that flow from conviction and that may set a person apart from social norms, such as refusing to pay bribes, abstaining from alcohol, or declining to take part in a ritual everyone else observes.
I call these two the semi-active forms, not because they are unimportant but because they stop at the person: attending a service or refusing a bribe makes someone visible as religious without any attempt to change anyone else.
At the active end are missionary activity, the effort to make converts, or to call nominal believers to a more serious practice of their faith, and civic participation, meaningful engagement in charitable work, civil society, or politics as a result of religious convictions.
The further along the continuum a person moves, the more likely their behavior is to collide with whoever holds power locally.
The further along the continuum a person moves, the more likely their behavior is to collide with whoever holds power locally, whether that is a drug cartel, a traditional authority, or a communist party.
Northeast Mexico shows how this plays out. In the three states I studied, Nuevo León, Tamaulipas, and San Luis Potosí, roughly 96 percent of the population self-identifies as Christian, which is about the national figure. On identity alone there is no minority to study and no variation to explain.
Move one step along the continuum and the picture changes. RIFREM, a Mexican network of scholars of religion that runs a national survey of religious beliefs and practices, put weekly church attendance in the north and northeast at about 54 percent, and Pew’s 2014 figure for Mexico as a whole was 45 percent, with a sharp gap between Catholics at 44 percent and Protestants at 76 percent.
Move to religious lifestyle, using RIFREM’s self-description as a “believer by conviction” or “practicing believer,” and the share drops to roughly 37 percent. At the active end the numbers fall off a cliff. Pew found that about 10 percent of Mexican Christians share their faith with others at least weekly, and about 13 percent of regular churchgoers had organized or joined a protest for the rights of the poor in the previous year.
In a country that is 96 percent Christian, the actively religious are a minority of somewhere between one in ten and one in eight.
So in a country that is 96 percent Christian, the actively religious are a minority of somewhere between one in ten and one in eight, and nobody looking at religious identity would ever notice them. When something happens to one of them, the incident is noticed, but as a crime, not as an attack on religion. (The figures are from my dissertation, drawing on Pew Research Center, 2014, the RIFREM survey, 2016, and the World Christian Database, 2017.)
That minority is the one that carries the risk, and the type of risk tracks the type of behavior. Religious participation, the mere act of gathering in a church, exposed people to the cartels’ protection rackets, to kidnapping for ransom, to curfews on evening services, and in some cases to the forced installation of shrines to Santa Muerte, the skeletal folk saint venerated by many cartel members, inside church buildings.
These threats are real, but most of them are not aimed at religion as such; churches are targeted because they are visible, easy to enter, and assumed to hold cash, and any business in the same street faces the same racket.
Religious lifestyle raised the stakes: pastors who refused on principle to pay the derecho de piso, the cartels’ “floor fee” for the right to keep operating, were kidnapped and in at least one case killed, not because they were Christians but because their convictions made them refuse.
Missionary activity produced threats that only religious people face.
Missionary activity produced threats that only religious people face. Former cartel members who converted and left disappeared. Pastors who evangelized in cartel-controlled neighborhoods received death threats or cancelled campaigns after warnings of gunfire.
Civic participation drawn from religious conviction, whether a football team to keep boys out of recruitment, a drug rehabilitation program, or a sermon denouncing extortion, put religious leaders in direct competition with the cartels for influence over the same people, and they were treated accordingly. The football team’s founder had threatening notes left on the church door and calls made to his wife. A rehabilitation clinic in Villa Nueva was attacked and fifteen patients and social workers were killed.
The small group at the active end is the one that matters for anyone trying to understand who is vulnerable and why.
The pattern is the same one Burge’s index reveals in the United States, except that in Mexico the far end of the distribution is where people get hurt. Most of the self-identified population does little. A shrinking share does more. And the small group at the active end is the one that matters for anyone trying to understand who is vulnerable and why.
Two Luthers show what an identity lens misses. Martin Luther was a Catholic monk when he nailed his theses to the church door. So, on paper, he was a member of the majority. Martin Luther King Jr. described everything he did in civil rights as part of his ministry, but his vulnerability did not stem from being a Baptist. It stemmed from what he did as a Baptist. A framework built on identity captures neither man.
Why nobody sees the violations
I am convinced, that this distinction is the main reason so many people believe that Latin America has no religious freedom problem. Ask most diplomats, journalists, or academics, including many of my colleagues in political science, and the answer is the same: the region is overwhelmingly Christian, the constitutions guarantee freedom of worship, and whatever violence exists is criminal or political, not religious.
That answer is what you get when you look only at identity. The instruments that measure religious freedom, from the Pew indices to the US State Department reports, are built to detect discrimination and violence against religious identity groups, and they do that job reasonably well. But when the victims belong to the majority religion, there is no identity group to flag.
If you go looking for violence against Catholics in Mexico, a country that is overwhelmingly Catholic, you will not find much, because Catholics as a category are not being targeted. If you go looking for violence against actively practicing Christians who run drug rehabilitation programs, offer alternatives to cartel recruitment, or preach against extortion, you will find plenty.
Priests and pastors killed in Mexico over the past decade were not killed for their baptism certificates. They were killed for their behavior.
The priests and pastors killed in Mexico over the past decade were not killed for their baptism certificates. They were killed for their behavior. Leave the behavioral dimension out of the analysis and every one of those killings gets filed under organized crime, with nothing religious about it. That is exactly what has happened, and it is why the standard answer about Latin America is wrong.
The gap shows up even in the best work in the field. Under Caesar’s Sword, the Notre Dame project led by Daniel Philpott and Timothy Shah, spent three years studying how Christian communities respond when their religious freedom is severely violated, with cases in some 25 countries across the Middle East, Africa, and Asia.
It is the most systematic study of its kind, and its case selection was entirely sensible given what the existing instruments show, which is the point: Latin America does not appear among its cases, because on every identity-based measure there was nothing there to study. Serious scholars using the standard tools reached the standard answer. The problem is in the tools.
The cost of that mistake is not academic. A killing filed under organized crime does not appear in any religious freedom report, so no embassy is likely to raise it and no international body to ask about it. Mexico’s federal protection mechanism exists for journalists and human rights defenders; a pastor threatened for running a rehabilitation clinic does not obviously fit either category.
Misclassification is not a bookkeeping error. It decides who gets protected.
Attention and money follow the rankings, and Latin America never tops them. And the next pastor who starts a football team gets no warning that he has moved into a category of risk that nobody has named. Misclassification is not a bookkeeping error. It decides who gets protected.
The contrast with places where identity itself is targeted makes the mechanism visible. A Christian in North Korea, a Copt in Egypt, a Baha’i in Iran, a Yazidi in northern Iraq, or a Christian farmer in Nigeria’s Middle Belt is at risk simply for belonging to the group, before doing anything at all, and the more active forms of religious behavior add to that risk rather than creating it.
In Latin America... the identity-based tools register nothing, and the people who are at risk for what they do as Christians fall through the gap.
Because the vulnerability attaches to the identity category, the identity-based tools pick it up, and those are the countries that fill the top of the religious freedom rankings. That is not a sign that the tools are working well; it is a sign that in those places the thing they measure coincides with the thing that matters. In Latin America the two come apart. Nobody is at risk for being Christian, so the identity-based tools register nothing, and the people who are at risk for what they do as Christians fall through the gap.
For most of Latin American history, as for most of European history, nearly everyone belonged to the same church, so measures of persecution built on religious identity had nothing to grip. That is changing with secularization, the growth of Protestantism, and the rise of new religious movements, and identity-based tools are slowly starting to register the new pluralism.
The actively practicing (Christian) minority has always been distinguishable from the rest, and it has always faced distinct risks.
But even within a nominally homogeneous population, the actively practicing minority has always been distinguishable from the rest, and it has always faced distinct risks. We simply never measured the difference. Burge has shown that it can be measured. What that would take to accomplish outside the United States, and why the other half of the data does not exist, is the subject of my follow-up post available here: https://five4faith.substack.com/p/the-data-that-does-not-exist.
Originally published as a Five4Faith Substack article. Republished with permission.
Dennis P. Petri, PhD is the International Director of the International Institute for Religious Freedom and Founder and scholar-at-large of the Observatory of Religious Freedom in Latin America. He is a Professor in International Relations at the Latin American University of Science and Technology. He is the author of The Specific Vulnerability of Religious Minorities, a book on undetected religious freedom challenges in Latin America.
The International Institute for Religious Freedom (IIRF) was founded in 2005 with the mission to promote religious freedom for all faiths from an academic perspective. The IIRF aspires to be an authoritative voice on religious freedom. They provide reliable and unbiased data on religious freedom—beyond anecdotal evidence—to strengthen academic research on the topic and to inform public policy at all levels. The IIRF's research results are disseminated through the International Journal for Religious Freedom and other publications. A particular emphasis of the IIRF is to encourage the study of religious freedom in tertiary institutions through its inclusion in educational curricula and by supporting postgraduate students with research projects.
