Note that theses maps have not been updated to account for post-2016 redistricting in North Carolina and Pennsylvania.
Religion is one of the driving forces in American culture; it always has been, and continues to do so even as American religiosity, in general, has declined. It tends not to loom as large as race in the American consciousness, but it’s still one of the main ways that Americans tend to identify themselves, and to divide themselves along “us” and “them” boundaries.
More specifically, religion is something that tends to correlate with, and be predictive of, one’s politics. This is especially the case in an era where cultural issues supplant economic issues in many people’s ideological framework. If you genuinely believe, for instance, that abortion means killing people, that often creates a moral imperative against which, say, your marginal tax rate pales in comparison.
You probably have a vague idea of which parts of the country are more religiously active, and where religion is more suffused into local politics (usually taking it in a more conservative direction), versus which parts of the country are more laissez-faire about religion and about social issues in general. You may also have a general sense of, say, which parts of the country are more likely to be Southern Baptist or Catholic or Lutheran; you may have even seen state or even county-level maps that depict this. What I’m doing in this article that hasn’t been done before, though, is to quantitatively show those differences at the level of congressional districts, where much of our politics actually happens, so we can get a better sense of how religiosity corresponds with people’s electoral choices.
I’ve put together a spreadsheet that covers every CD, and what percentage of each CD’s residents belong to which religious tradition and denomination. One additional advantage of using district-level data is that it, as seen above, it lets us use the Daily Kos Elections 435-CD cartogram map, which breaks out the nation’s metropolitan areas into proper focus, and de-emphasizes the nation’s empty spaces (which has the effect of not making the country seem more evangelical, or more Mormon, than it actually is). As we continue, we’ll talk more about the methods and limitations behind these maps, and then zoom in on some additional data for certain denominations.
As a general rule, white evangelical Protestants tend to be more likely to be Republican; non-religious persons and members of non-Christian religions are disproportionately Democratic. Catholics and mainline Protestants are more of a mixed bag. For one thing, the frequency of church attendance, just as much as religious tradition, often indicates a person’s politics. A Catholic who attends mass multiple times per week is more likely to be politically conservative than one of the Easter-and-Christmas variety. Also, the ways that race and religion interact are very important. Hispanic Catholics are more likely to be Democrats than white Catholics, and African-Americans who attend churches with an evangelical framework are still quite likely to vote Democratic.
Still, people who write and think quantitatively about politics and demographics don’t often write much about religion. Part of that may simply be that sense that religion is something that’s personal and considered somewhat of an impolite line of questioning. (Though, in some parts of the country, what church you attend is often something that comes up in conversation right away when meeting someone new—though that may be more prevalent in areas where there’s more of a religious monoculture, and one’s denomination might simply be assumed.) But more than anything, it may be because the U.S. Census Bureau, unlike some other countries’ censuses, doesn’t ask anything about religion. Go to the Census website and you can vacuum up endless information about not just race and income, but education, fertility, commuting, and housing costs—but nothing about religiosity.
A number of polling organizations do ask regularly about religion, often as separate in-depth studies instead of just tacked on questions in a political poll. PRRI's American Values Atlas and Pew Research's Religious Landscape Study are the most prominent (and detailed) examples of this in recent years, though Gallup has also been asking more rudimentary questions about religiosity for decades. The problem with their output is that it functions like a poll, not a census. It only contacts thousands of people and generalizes from there, enough to provide numbers for states or large metro areas, but not at a greater level of granularity, and with the requisite margin of error.
Luckily, an entirely separate non-profit organization, the Association of Statisticians of Religious Bodies (or ASARB) does release something like an informal religious census every 10 years (their data is hosted by the Association of Religion Data Archives or ARDA). They don't do so by contacting individual Americans, but by contacting churches themselves and getting membership numbers from them. This is superior to what Pew and PRRI are offering in several ways. One, it’s a much greater level of granularity, and you can get data at the county level (and then you can map it—just from the last few months, for example, see Lyman Stone’s maps of Reformation churches and Matthew Isbell’s maps of Judaism in America). And two, it theoretically encompasses everyone; people who aren’t members of a particular congregation are left as “unclaimed,” in ARDA’s terminology (which unfortunately makes them sounds more like lost luggage than people), but everyone is accounted for.
There are still some problems with the ARDA’s approach. For one thing, it doesn’t measure the religiosity of the people who are “claimed” by a particular congregation. The numbers reported by a particular church might mean only dues-paying members who attend regularly, or it might include sporadic attendees, or it might even include people who showed up once, signed the guest book, and never came back. It might include people who have an agnostic worldview, but are members of a church anyway, for family or social reasons or just out of inertia. And conversely, it misses people who are ardently religious in their beliefs but, for whatever reason, aren’t members of any particular congregation. It also sometimes misses entire congregations that have opened since the previous Census years (or includes one that have closed), and potentially misses informally organized congregations that they don’t know exist, because they aren't part of any national organization’s denomination and may not even be in the phone book.
This inability to account for people who would report to a pollster that they belong to a particular religious tradition, but who aren’t involved in a formally organized congregation, oddly enough, makes the ARDA’s data seem like the country is less religious than it actually is. For comparison, let’s take a look at the difference between their findings and the findings of the Pew Religious Landscape study, when you zoom out as far as possible and look at membership by religious tradition at the national level. (Religious “tradition” is a category that religion scholars use to classify different denominations within Christianity, such as whether a particular Protestant denomination is evangelical or mainline. A helpful comparison for the secular among us might be from biological taxonomy: if you think of Christianity as being like “kingdom” and a specific denomination like the Southern Baptist Convention as “species,” then tradition is something in between, like “order” or “class.”)
The differences especially stand out when looking at the number of evangelicals when compared with the number of Catholics. Consider that the Catholic Church is a large bureaucracy that, if nothing else, is good at keeping records, while many evangelicals belong to small, loosely-organized non-denominational churches, or, for instance, were raised in that tradition and might consider themselves “born again” even while not regularly attending church.
ARDA Pew
Catholic 19.1 percent Catholic 20.8
Evangelical Protestant 16.2 Evangelical Protestant 25.4
Mainline Protestant 7.3 Mainline Protestant 14.7
Other Religion 4.3 Other Religion 8.7
Black Protestant 1.6 Black Protestant 6.5
Orthodox Christian 0.3 Orthodox Christian 0.5
“Unclaimed” 51.2 Religious “nones” 23.4
(Note here that I condensed some of Pew’s categories to match ARDA’s categories. ARDA, for instance, includes Mormons and Jehovah’s Witnesses in the “Other” category along with non-Christian religions; Pew breaks them out separately under the “Christian” heading. Maybe most importantly, Pew’s equivalent to ARDA’s “unclaimed” is much more specific and thus much smaller; it only includes people who specifically identify as atheist, agnostic, “nothing in particular,” or “don’t know.”)
So, unfortunately, when two perfectly valid methods of gathering data result in such widely varying results, you have to take any findings with a few grains of salt; in other words, you’ll need to accept that the results aren’t precise and approach it more like we’re describing a range of possible results. When a person reports to the Census Bureau that he is 30 years old and has a bachelor’s degree, for instance, there’s no subjectivity there; even with the social construct of race, there are enough cues and markers there that most people feel like they can answer that question without much ambiguity. There’s a lot more ambiguity with religious beliefs, though, where considering yourself to be a believer in X tradition or participant in Y denomination is largely a state of mind, and doesn’t necessarily have much to do with whether you’re on a formal membership list somewhere.
TRADITIONS
With those caveats in mind, let’s start looking at the actual data of what religious traditions are most prevalent in what congressional districts. Start by taking a look at the map at the very top of the article (click here for a larger version). This map shows what the most common tradition is in each CD—excluding “unclaimed,” which is so prevalent that, if I treated it as a tradition unto itself, it would be the plurality in almost every CD, except for a few intensely religious parts of the country (such as Utah, parts of the Deep South, and a few heavily Catholic suburbs in the Northeast).
There’s no theological significance to coloring the predominantly Catholic parts of the country blue and the predominantly evangelical parts of the country red. But it was still a very intentional decision on my part, because it underscores how much a religious map of the U.S. looks like a map of Democratic and Republican strength in various parts of the country. In other words, the parts of the country where there are more Catholics—the Northeast, the West, and the more urban parts of the Midwest—also tend to be the parts of the country likelier to vote Democratic; in the South and in certain rural parts of the Midwest, you see both more evangelicals and more Republican party strength.
In fact, if you look closely, you can see certain states where there’s a clearly defined geographical split between religious traditions, which is also mirrored in that state’s split in political preferences. In Texas, for instance, you have a clear distinction between the counties in the Rio Grande Valley reaching over to El Paso, which are mostly Catholic, versus most of the rest of the state, which is predominantly evangelical; not coincidentally, those Catholic areas are the most heavily Democratic parts of the state (though, of course, that also has to do with the fact that those areas have a Latino majority).
You can also see that in, say, Florida, where the metro areas of the south part of the state have a Catholic plurality (and, again, where Latinos are more prevalent, as well as white residents who’ve moved there from the Northeast or Midwest), versus the more traditionally “southern” Panhandle, which is both much more evangelical and more conservative than the rest of the state. You can see it in Virginia, where you have the D.C. metro area (which, as with Florida, has a Catholic plurality primarily thanks to in-migration) vs. an evangelical plurality in the rest of the state. You can see it in Michigan, which has a Catholic, Democratic east and a largely Dutch evangelical, Republican west; you can even see it in Louisiana, which definitely has an evangelical north and a French-ancestry, Catholic south, and where as recently as the 1990s the Catholic vote was usually Democratic-leaning.
This isn’t to say that there’s a direct causal relationship; that, say, Massachusetts is so strongly Democratic because of how Catholic it is, and Alabama is so strongly Republican because of how evangelical it is. Racial composition and education levels are already strongly correlated with political preferences, and they each seem to have more explanatory power than does religion. For example, here are correlations, at the county level, between percentage of non-Hispanic whites and percentage of bachelor’s degree holders (from the 2011-2015 American Community Survey), and percentage of evangelical, mainline, black Protestant, and Catholic adherents (from the 2010 ARDA census), versus the Democratic vote share in the 2016 presidential election. (The way that correlation coefficients work is that the closer you are to 1 or -1, the stronger the relationship is. A zero means no relationship at all, only random noise.)
Correlation with non-Hispanic white: -0.59
Correlation with bachelor’s degree or higher: 0.43
Correlation with black Protestant adherents: 0.35
Correlation with evangelical Protestant adherents: -0.31
Correlation with Catholic adherents: 0.22
Correlation with mainline Protestant adherents: -0.15
The positive numbers next to black Protestants and Catholics indicate that the more there are of them in a county, the more likely that county will vote Democratic — but it’s still not as strong a predictor as how many people with college educations are there. And, likewise, the negative numbers next to evangelical and mainline Protestants indicate that the more there are of them, the more likely a county is to vote Republican — though, again, it’s not as strong a predictor as the percentage of white residents in general. (In fact, the coefficients for Catholics and mainline Protestants are low enough they’re barely suggestive of any relationship at all.)
Even if there isn’t a statistically strong relationship, though, it’s still pretty visually striking when you see it on a map: the parts of the country where there are more Catholics than evangelicals tend to also be the “blue” parts of the political map, and the parts of the country where there are more evangelicals than Catholics — or the few parts of the country where either mainline Protestants or Mormons (who account for most of the “other” category here) are dominant — are usually the “red” parts of the political map. You might say that religion interacts with race and education, and other factors like urban-ness and average age levels, in interesting ways to produce the political map we’re used to.
This map, however, doesn’t give specific percentage levels of each type of tradition, and only shows which tradition makes up the plurality in each CD. So let’s zoom in, more specifically, on which CDs ave the highest percentages of each type of religious tradition. Let’s start with the most evangelical:
| CD |
EVANGELICAL % |
2016 vote |
| AL-04 |
53.7 |
17/80 |
| TX-13 |
46.7 |
17/80 |
| KY-01 |
46.6 |
24/72 |
| OK-05 |
46.3 |
40/53 |
| OK-02 |
46.1 |
23/73 |
| TN-03 |
44.9 |
30/65 |
| TN-02 |
44.0 |
30/65 |
| MS-03 |
43.8 |
37/61 |
| MS-01 |
43.7 |
32/65 |
| AL-06 |
43.5 |
26/71 |
I’ve added the 2016 vote share in this particular table, to highlight just how red these districts are. In fact, the two reddest districts in the nation—Alabama’s 4th district, in the rural area north of Birmingham, and Texas’s 13th district, in the Texas panhandle—are also the two most evangelical districts in the nation! One other thing you might notice is that all of these districts are in the former Confederacy, and that shouldn’t be a surprise either; the largest evangelical denomination, by far, is the Southern Baptist Convention, who split from the Northern Baptists—today the mainline denomination known as the American Baptist Churches—in 1845 specifically over the question of slavery.
(If you’re wondering what the least evangelical district in the nation is, you might be thinking that it’s a district in Upper New England or the Pacific Northwest where there just aren’t a lot of religiously active people, period. It’s not, though: the answer is Utah’s 3rd district, at 1.4 percent! As we’ll see shortly, this is the CD that comes closest to being a religious monoculture; in this case, most residents are Mormon, to the near-exclusion of both Catholic and Protestant churches.)
Now, as a comparison, let’s look at the most Catholic CDs:
| CD |
CATHOLIC % |
2016 VOTE |
| NY-01 |
58.9 |
42/55 |
| NY-02 |
56.8 |
44/53 |
| MA-04 |
51.0 |
59/35 |
| NY-04 |
50.6 |
53/44 |
| NY-03 |
50.0 |
52/46 |
| PA-08 |
49.9 |
48/48 |
| MA-05 |
47.5 |
69/26 |
| MA-07 |
46.9 |
84/12 |
| MA-06 |
46.5 |
56/38 |
| MA-08 |
45.7 |
60/34 |
Here, the two bluest districts in the nation (New York’s 15th and 13th districts) most certainly aren’t the most Catholic districts in the nation. (What NY-15 and NY-13 are is among the least white districts in the nation.) In fact, the two most Catholic districts in the nation, in Suffolk County, New York on Long Island, are ones that Hillary Clinton outright lost (though Barack Obama won them narrowly in 2012). One thing you might be noticing is that these districts are in the Northeast, but they’re predominantly in the Northeast’s suburbs rather than the cities themselves (with the exception of MA-07, which contains most of Boston). These suburbs, not coincidentally, are heavily Irish and to a lesser extent Italian.
Let’s look at the districts with the most mainline Protestants:
| CD |
MAINLINE % |
2016 VOTE |
| MN-07 |
30.4 |
31/62 |
| IA-04 |
29.6 |
34/61 |
| ND-AL |
29.3 |
28/64 |
| MN-01 |
26.3 |
38/53 |
| SD-AL |
24.1 |
32/62 |
| NE-03 |
23.0 |
20/75 |
| IA-01 |
21.7 |
45/49 |
| PA-15 |
18.7 |
44/52 |
| PA-04 |
18.3 |
37/59 |
| IA-03 |
18.2 |
45/49 |
While these are all districts that went GOP in the 2016 election, many of them were ones that were narrowly won by Barack Obama in 2012 but had some of the reddest swings from 2012 to 2016 of any CDs. Note that most of these districts are in the Upper Midwest (Minnesota, Iowa, the Dakotas) and, as we’ll discuss later, are heavily Lutheran. The only exceptions from the top 10 are several districts in Pennsylvania. Also, if you refer to the map at the top, you’ll note that a few of these districts, despite being in the top 10 for mainline Protestants, still don’t have a mainline Protestant plurality, because there are even more Catholics present. (This aberration also goes the other way too; there are some districts, like Maryland’s 1st or Ohio’s 6th, where there is a mainline Protestant plurality, but not a high enough mainline Protestant percentage to make this top 10 list. That’s probably because those districts have a higher rate of “unclaimed” residents.)
Now let’s look at the districts with the most members of historically black Protestant denominations:
| CD |
BLACK PROTESANT % |
2016 VOTE |
| TN-09 |
16.3 |
78/20 |
| AL-07 |
14.6 |
70/29 |
| MS-02 |
11.2 |
64/35 |
| AL-06 |
9.3 |
26/71 |
| SC-06 |
8.9 |
67/30 |
| GA-02 |
8.5 |
55/43 |
| DC-AL |
8.4 |
91/4 |
| AL-02 |
8.3 |
33/65 |
| TN-08 |
8.2 |
31/66 |
| AR-04 |
7.6 |
31/64 |
Here, as you’d expect, most of the districts in the top 10 list are among the African-American-majority districts in the South. The actual percentages might seem low at first glance, and that’s partly because of the way the ARDA sets up its data. This category doesn’t simply measure all African-Americans who are Protestant; it only considers people who are members of the major historically black churches (the largest of which are the National Baptist Convention, USA and the African Methodist Episcopal Church).
That winds up leaving out not only African-Americans who are devout but not members of a particular congregation, but also African-Americans who are members of a congregation that’s officially non-denominational. (In other words, a Protestant church that’s independent and not affiliated with a national organization—this can take the form of anything from a suburban mega-church with thousands of members to a storefront church with dozens of members. ARDA categorizes all non-denominational Protestant churches as “evangelical”—and while most, indeed, are evangelical, that umbrella category vacuums up a lot of churches that have a more liberal worldview, that are majority-African-American, or both.) It also neglects to count African-Americans who are members of denominations that aren’t “historically black,” whether mainline or evangelical (nearly 20 percent of Southern Baptist Convention members, for instance, are people of color, despite that denomination’s historical reason for existing).
Now let’s look at the ARDA’s “Other” category, a catch-all which cobbles together both Christian denominations that don’t fall in either the Catholic or Protestant traditions (most notably, the Church of Jesus Christ of Latter-Day Saints), and all non-Christian religions, including Judaism, Islam, and Buddhism.
| CD |
OTHER % |
2016 VOTE |
| UT-03 |
72.9 |
23/47 |
| UT-01 |
69.8 |
22/50 |
| UT-02 |
67.3 |
32/46 |
| UT-04 |
64.5 |
32/39 |
| ID-02 |
40.9 |
30/55 |
| NY-09 |
15.9 |
84/14 |
| NY-08 |
15.5 |
85/14 |
| NY-17 |
14.9 |
59/38 |
| NY-07 |
14.3 |
87/10 |
| NY-10 |
13.4 |
78/19 |
You can see a very clear division here within the “Other” category: half of the top 10 are districts in the interior West where Mormons are dominant (Utah and southern Idaho). The other half are the districts with the largest concentrations of Jewish residents, especially, for purposes of this list, Orthodox Jews (who are heavily concentrated into Brooklyn and several suburban counties in the New York metro area—Reform and Conservative Jews, by contrast, are more widely and thinly distributed among a number of different large metro areas). If you look at the map above, you’ll notice there are two other CDs with an “Other” plurality: Arizona’s 1st and Wyoming’s at-large district, which have significant Mormon populations—though nowhere near the same levels as Utah or Idaho—and not a lot of anyone else. (Hawaii’s 2nd district also narrowly misses the cut on an “Other” plurality, thanks to an unusual combination of a lot of Mormons and a lot of Buddhists.)
I should also mention the nation’s Orthodox Christians (in other words, Eastern Rite churches like Greek Orthodox or Russian Orthodox), whom ARDA breaks out as a separate tradition instead of filing them under “Other.” At only 0.3 percent of the nation’s population, they’re a very small part of the overall religious landscape. The CD with the highest percentage of Orthodox Christians is the District of Columbia, at 2.3 percent. It’s narrowly followed by Indiana’s 1st, Florida’s 13th, and Alaska’s at-large districts (which have significant Serbian, Greek, and Russian populations, respectively).
With those six traditions accounted for, we should also give some serious discussion to the “tradition” that actually encompasses the majority of the nation, according to ARDA: the “unclaimed” population. They aren’t uniformly distributed throughout the nation, though, just as with every other tradition. This choropleth map shows where the highest and lowest levels of “unclaimed” persons are; they’re highest in the Pacific Northwest and upper New England, and, interestingly, also in states like Florida, Nevada, and Arizona where the population is rapidly growing and there are a lot of people who, having recently uprooted their lives to move there, haven’t affiliated with a particular congregation yet. The levels are lowest in places where one tradition is particularly dominant, like in Utah or the rural Deep South. (The least “unclaimed” district is the same one that’s the least evangelical, by virtue of how solidly Mormon it is: Utah’s 3rd, with only 16 percent unclaimed.)
I'm willing to bet that some of the people from Oregon or Washington who are reading this article took one look at the map at the top of the article and said “Hey, wait a minute! Why is the Northwest red? There aren’t a lot of evangelicals here!” Well, that’s exactly right; there aren’t a lot of evangelicals there, and the region’s politics reflects that. Keep in mind that the map at the top doesn’t show how many evangelicals are there; it just shows that there are more evangelicals than Catholics there. There aren’t many evangelicals in the Northwest (there’s very little penetration by the Southern Baptists into that part of the country; the evangelicals there tend to be in Pentecostal churches like the Assemblies of God or non-denominational mega-churches), but there are even fewer Catholics there. The “unclaimed” map adds a helpful bit of context here; most Northwesterners aren’t members of any church (regardless of whether they think of themselves as religious or as non-believers).
Here are the most “unclaimed” districts:
| CD |
“UNCLAIMED” % |
2016 VOTE |
| OR-04 |
74.2 |
46/46 |
| AZ-04 |
73.9 |
28/68 |
| ME-02 |
73.7 |
41/51 |
| FL-12 |
73.4 |
39/57 |
| ME-01 |
71.1 |
54/39 |
| WA-06 |
70.3 |
52/40 |
| NV-02 |
69.8 |
40/52 |
| CA-01 |
69.7 |
37/56 |
| WA-03 |
69.7 |
43/50 |
| WA-02 |
69.5 |
57/35 |
As you can see, there’s no clear pattern between how “unclaimed” a district is and its politics. (That’s especially true when you look for a correlation. The correlation coefficient between how “unclaimed” a county is, and the 2016 Democratic vote share, is a meager 0.16.) A lot of the top 10 CDs are swing districts, but there are also some fast-growing areas with a lot of retirees that are dark red.
DENOMINATIONS
Finally, let’s drill down a little deeper, to the level of denomination (for instance, asking not just which CD has the most mainline Protestants, but, say, which has the most members of the United Methodist Church). Whether or not a specific denomination is prevalent in a region doesn’t usually have much to do with that region’s politics; if only because there are simply so many Protestant denominations, and in many places there aren’t large numbers of any particular one of them. In other words, what’s important, in predicting how conservative an area is, has more to do with simply how evangelical it is overall, rather than whether those evangelicals are mostly Southern Baptist or mostly Pentecostal.
Oddly enough, this denomination-centered analysis also generates a less interesting map. (It does tell you a lot about a region’s cultural flavor, though, and if you do a breakdown of the top 10 CDs for a particular denomination, you get another level of nuance that helps you understand what parts of the country major denominations are concentrated in.)
The map of what is the plurality denomination in each CD just winds up making Catholicism look even more dominant throughout the country. For instance, if you look closely, you’ll see that there’s no longer any red in the Pacific Northwest; Catholicism is the most common denomination there. The same is true of the parts of the Midwest that show up as evangelical on the map of traditions, like Indiana, southern Illinois, or western Michigan.
The reason for this is an artifact of the way that Protestantism is broken down into dozens of different denominations with different theological frameworks—but Catholicism is, for all practical purposes, one huge unitary organization. (There are certain breakaway, non-Papist Catholic churches, and the ARDA actually includes one of them, the United Catholic Church, as a separate entity under the “Catholic” tradition. It has only several thousand U.S. adherents, though, so 99.999 percent of the 59 million Americans that ARDA considers “Catholics” are members of the standard Roman Catholic Church.) So, to return to Oregon as an example, there are more evangelicals total in Oregon than there are Catholics, but there are more Catholics in Oregon than just, say, members of the Assemblies of God, or of the Seventh-Day Adventists.
In most of the southern states, the vast majority of evangelicals are members of the Southern Baptist Convention, so that stays in the plurality in most southern districts. (The same is true of the Mormon pluralities in Utah, Wyoming, and Idaho.) There are only a handful of CDs where there’s a different plurality Protestant denomination; several have a dominant mainline denomination (the Evangelical Lutheran Church in America in Minnesota’s 7th, and the United Methodist Church in Indiana’s 2nd and West Virginia’s 2nd and 3rd districts). And in Georgia’s 5th congressional district, which is primarily in Atlanta, members of non-denominational churches are most common. (Keep in mind that here, it’s probably black-majority churches that aren’t part of the historically black Protestant denominations, more so than your stereotypical white suburban mega-churches.)
To get a better sense of what the most common denominations within each Protestant tradition (and within the “other” category) are, let’s look at how ARDA breaks them down. (Note that, for purely practical purposes, this list stops at denominations that make up around one-third of one percent of the nation’s population. ARDA categorizes hundreds of denominations, many with only a few thousand members nationwide.)
Catholic 19.1 percent
Evangelical Protestant 16.2 percent
Southern Baptist Convention 6.4 percent
Non-denominational 4.0 percent
Assemblies of God 1.0 percent
Lutheran Church — Missouri Synod 0.7 percent
Churches of Christ 0.5 percent
Christian Churches 0.5 percent
Seventh-Day Adventist Church 0.4 percent
Church of God (Cleveland, TN) 0.4 percent
Church of the Nazarene 0.3 percent
Mainline Protestant 7.3 percent
United Methodist Church 3.2 percent
Evangelical Lutheran Church in America 1.4 percent
Presbyterian Church (USA) 0.8 percent
Episcopal Church 0.6 percent
American Baptist Churches in the USA 0.5 percent
United Church of Christ 0.4 percent
Christian Church (Disciples of Christ) 0.3 percent
Other 4.3 percent
Church of Jesus Christ of Latter-Day Saints 2.0 percent
Muslim 0.8 percent
Orthodox Judaism 0.3 percent
Reform Judaism 0.3 percent
Mahayana Buddhism 0.2 percent
Conservative Judaism 0.2 percent
Black Protestant 1.6 percent
National Baptist Convention, USA 0.6 percent
African Methodist Episcopal Church 0.3 percent
Church of God in Christ 0.2 percent
So, let’s look at the top 10 districts for a few of the most common Protestant denominations. Let’s start with the most numerous one, the Southern Baptists.
| CD |
SOUTHERN BAPTIST % |
2016 VOTE |
| AL-04 |
39.1 |
17/80 |
| KY-01 |
36.4 |
24/72 |
| MS-03 |
35.5 |
37/61 |
| MS-01 |
34.7 |
32/65 |
| TN-02 |
33.2 |
30/65 |
| NC-11 |
33.1 |
34/63 |
| SC-03 |
32.3 |
29/67 |
| TX-04 |
31.9 |
22/75 |
| TX-13 |
30.8 |
17/80 |
| LA-05 |
30.8 |
34/64 |
You’ll notice that this list is pretty similar to the list of most evangelical districts, which isn’t surprising since nearly half of all the nation’s evangelicals are Southern Baptists. The Oklahoma districts, however, fall off the list (Oklahoma, less identifiably a part of the Confederacy than most of these other states, isn’t as monoculturally Southern Baptist as the other states; it does, however, still have a lot of Assemblies of God and non-denominational church members as well.)
| CD |
NON-DENOMINATIONAL % |
2016 VOTE |
| GA-05 |
10.3 |
85/12 |
| TX-03 |
10.0 |
41/55 |
| IN-01 |
9.9 |
54/42 |
| GA-13 |
9.4 |
71/27 |
| OK-01 |
9.2 |
33/61 |
| TX-32 |
8.9 |
49/47 |
| MD-04 |
8.9 |
77/20 |
| TX-30 |
8.8 |
79/18 |
| TX-02 |
8.5 |
43/52 |
| TX-07 |
8.5 |
49/47 |
The most non-denominational districts are an interesting mixed bag; the common thread is that they are mostly suburban districts, with the accompanying recently founded mega-churches that aren’t affiliated with a particular denomination. Some of these suburban districts are mostly black and heavily Democratic (like Maryland’s 4th district, in Prince George’s County, and Georgia’s 13th district). The districts in the Dallas (TX-03 and TX-32) or Houston (TX-02 and TX-07) suburbs are, however, likely the white upper-middle-class professionals that you might expect when visualizing suburban mega-churches. (It’s notable, though, that these are some of the Never Trump-iest districts that swung most significantly in the Dem direction in 2016. Contrast them with the districts where Southern Baptists are most prevalent, which tend to be more rural, less educated, and which got even more intensely Trumpy in 2016.)
| CD |
ASSEMBLIES OF GOD % |
2016 VOTE |
| MO-07 |
5.9 |
25/70 |
| WA-06 |
3.6 |
52/40 |
| MO-08 |
3.1 |
21/75 |
| FL-10 |
3.1 |
62/35 |
| AR-03 |
3.0 |
31/62 |
| MN-02 |
3.0 |
45/47 |
| OK-02 |
2.9 |
23/73 |
| WA-10 |
2.9 |
51/40 |
| FL-19 |
2.7 |
38/60 |
| MI-07 |
2.7 |
39/56 |
The Assemblies of God are the nation’s largest Pentecostal denomination. (Like other evangelical denominations, Pentecostals believe in biblical inerrancy, but there’s also an added dimension in the form of belief in “spiritual gifts,” such as speaking in tongues.) Pentecostalism is usually associated with the rural South, but the Assemblies of God is pretty widely, if inconsistently, distributed throughout the nation. Its headquarters are in Springfield, Missouri, and that’s where the heaviest concentration is, but there are also clusters in the suburbs of various major cities of the West, Midwest, and South. (The second largest Pentecostal denomination, the Church of God, is much more heavily concentrated in the rural South, especially in Tennessee and Alabama.)
| CD |
LCMS % |
2016 VOTE |
| NE-03 |
8.3 |
20/75 |
| NE-01 |
7.6 |
36/58 |
| MN-07 |
6.6 |
31/62 |
| IA-04 |
6.4 |
34/61 |
| MN-01 |
6.2 |
38/53 |
| WI-07 |
6.1 |
37/58 |
| WI-06 |
5.6 |
39/56 |
| WI-08 |
5.6 |
39/56 |
| MI-05 |
5.0 |
50/46 |
| IN-03 |
4.9 |
30/65 |
Here’s where things get a little confusing. The Evangelical Lutheran Church in America isn’t evangelical: they’re the liberal, mainline branch. (Think the fictional residents of Lake Wobegon.) The Lutheran Church — Missouri Synod, however, are evangelical … but they aren’t primarily found in Missouri, instead especially in rural parts of the Upper Midwest like Wisconsin, and the Great Plains states. On top of that, there’s also the Wisconsin Evangelical Lutheran Synod, which in fact is evangelical (and generally even more conservative than the LCMS), though it’s considerably smaller than ELCA and LCMS.
Now let’s switch over to some of the nation’s largest mainline denominations, starting with the most numerous one, the United Methodist Church:
| CD |
METHODIST % |
2016 VOTE |
| IN-02 |
14.1 |
36/59 |
| IA-04 |
10.1 |
34/61 |
| OK-01 |
9.9 |
33/61 |
| NC-13 |
9.9 |
44/53 |
| PA-05 |
9.8 |
34/62 |
| NE-03 |
9.7 |
20/75 |
| WV-01 |
9.4 |
26/68 |
| KS-01 |
9.4 |
24/69 |
| MD-01 |
9.2 |
33/62 |
| GA-06 |
8.6 |
47/48 |
Unlike the other mainline denominations, which tend to be concentrated in one part of the country (for instance, ELCA is mostly in the Upper Midwest, Episcopalians and the United Church of Christ are mostly in the Northeast, and Presbyterians are in the Mid-Atlantic states in a band from Pennsylvania to the Carolinas), Methodists are pretty widely and evenly distributed everywhere but the West. In fact, they tend to be the default second-most-common denomination in both the Northeast and Great Lakes states (behind Catholicism) and in the South (behind the Southern Baptists).
| CD |
ELCA % |
2016 VOTE |
| MN-07 |
25.8 |
31/62 |
| ND-AL |
24.3 |
28/64 |
| MN-01 |
19.3 |
38/53 |
| SD-AL |
13.8 |
32/62 |
| MN-08 |
13.5 |
39/54 |
| WI-03 |
13.1 |
45/49 |
| MN-02 |
13.0 |
45/47 |
| IA-04 |
11.1 |
34/61 |
| MN-05 |
10.7 |
74/19 |
| MN-03 |
10.5 |
51/41 |
The Evangelical Lutheran Church in America (which, despite its name, is the mainline branch of Lutheranism) is heavily concentrated in the Upper Midwest, especially in Minnesota and the Dakotas. Several overwhelmingly Lutheran CDs—Minnesota’s 1st and 7th, and Iowa’s 4th—manage to be in the top 10 for both ELCA and the evangelical branch of Lutheranism, the LCMS. In general, though, the more conservative LCMS is located more in the Great Plains states and border states further south. Despite ELCA’s doctrinal liberalism, though, the rural CDs where they’re most prominent still tend to be either already pretty conservative, or in some cases (like Minnesota’s 1st and 8th) moved pretty sharply in the GOP direction in 2016.
| CD |
PRESBYTERIAN % |
2016 VOTE |
| PA-03 |
4.7 |
35/61 |
| NC-12 |
4.2 |
68/28 |
| PA-18 |
3.8 |
39/58 |
| PA-14 |
3.5 |
66/31 |
| PA-12 |
3.1 |
38/59 |
| VA-06 |
3.0 |
35/60 |
| NC-08 |
2.6 |
41/56 |
| NC-09 |
2.5 |
43/54 |
| SC-04 |
2.4 |
35/60 |
| NC-13 |
2.3 |
44/53 |
The Presbyterian Church (USA), which is not to be confused with the Presbyterian Church in America (an evangelical branch that’s primarily in the South that, much like the Southern Baptists, split off and did their own thing—though the split happened in the 20th century and wasn’t, directly at least, about slavery), is widely distributed throughout the Northeast and South, though most heavily concentrated in western Pennsylvania and upland parts of the Carolinas. The Presbyterian Church has its roots in the Church of Scotland, and so it’s most prevalent in parts of the country where the first waves of Scottish and Scots-Irish immigration took hold.
| CD |
EPISCOPAL % |
2016 VOTE |
| DC-AL |
3.1 |
91/4 |
| SC-01 |
2.5 |
40/54 |
| VA-08 |
2.5 |
74/21 |
| CT-04 |
2.5 |
60/37 |
| VA-04 |
2.3 |
59/37 |
| RI-01 |
1.9 |
61/35 |
| CT-05 |
1.8 |
50/46 |
| RI-02 |
1.8 |
51/44 |
| FL-04 |
1.8 |
34/62 |
| VA-07 |
1.7 |
44/51 |
The Episcopal Church often gets thought of as the WASPy, upper-crust mainline denomination, probably owing to its roots in the Church of England. It may, however, be the most liberal of the mainline denominations (though the United Church of Christ, descended from New England Puritanism, could also lay claim to that). Contrary to what you might think, though, the Episcopal Church isn’t entirely concentrated in New England. It’s actually pretty common in the lowland parts of the South, like South Carolina’s Low Country and Virginia’s Tidewater region: the parts of the country that were first settled in the 17th century by England’s gentry, the ones who were more motivated by commerce than religious purity. Think back to Albion's Seed, if you’ve read it: it’s the church of the Cavaliers, who set up the plantation economy in the flat parts of the South.
At this point, your eyeballs have probably gone numb with chart madness, but we’ve barely scratched the surface of what the ARDA dataset uncovers. I’d encourage you to take a closer look at the full spreadsheet that I’ve put together; if you’ve ever wanted to know how many Amish there are in Alabama, or how many Zoroastrians there are in Wyoming (I’ll save you a click: there are four), here’s your chance! I’ve calculated percentages of each CD’s population for each tradition and for each major denomination; there are also raw numbers for every denomination that ARDA covers.
METHODOLOGY
As a post-script, there’s one last methodological note I should discuss. You might be wondering how I converted ARDA data, which is collected at the county level, into congressional districts. I simply used the same technique that I’ve used for past projects like the Daily Kos media markets-by-CD database. When counties are split into several districts, I’d allocate the percentage of the population based on what percentage of the county is in each district. As a hypothetical example, suppose that County X has 100,000 residents, 60 percent of whom are in district XX-01 and 40 percent of whom are in district XX-02. If County X has 10,000 Catholics, I’d simply assign 6,000 of them to XX-01 and 4,000 of them to XX-02.
Now, that works perfectly well when you have a district that’s composed of a number of rural counties, or even in small metro areas. When you look at counties with millions of residents, though, which are broken down into multiple and very demographically-different CDs, you might run into some significant inaccuracies. Take Los Angeles County and its 10 million residents, for instance. This method assumes that, say, CA-33 and CA-40, which are both entirely within Los Angeles County, have the same ratio of Catholics to Jews. However, it’s unlikely that CA-40, which is heavily Hispanic, has many Jews living in East L.A.; likewise, CA-33, on L.A.’s west side, probably has a disproportionately higher rates of Jewish residents than the county as a whole, and while it still has a lot of Catholics, it probably doesn’t have as high a percentage of Catholics as does the 40th.
There are probably a few other instances in the tables above where those problems seeped in. For instance, in the list of most “Other” districts, it’s likely that NY-10 has a significantly higher percentage of Orthodox Jews than does NY-08, and should be higher up the list. They’re both districts with a large portion of their population in Kings County (i.e. Brooklyn), but the 10th contains the heavily Orthodox neighborhood of Borough Park, while the 8th is mostly African-American neighborhoods further east. ARDA data doesn’t offer a level of granularity that would enable me to make that distinction, though. Similarly, the list of most black Protestant districts contains AL-06, which isn’t a black-majority CD; it’s really the opposite of that, in that it’s the district that covers Birmingham’s white suburbs, and probably doesn’t have a lot of black Protestants in it. However, AL-06 still covers a big portion of Jefferson County, and my method allocates black Protestants at the same ratio as in the adjacent black-majority district of AL-07.
I’m not terribly worried about that level of inaccuracy, though. (If I were, say, writing this as a dissertation, and not just an article for Daily Kos, I’d probably try to find a way around that, maybe by geo-coding the location of every congregation for which I can find a street address.) It’s already a pretty imprecise activity; for instance, simply knowing how many people attend a church in a particular county doesn’t necessarily mean that all of those people live in that county.
One interesting quirk in ARDA data is that usually you’ll notice higher rates of religious adherence in the core counties of metro areas than the surrounding suburban counties. That doesn’t necessarily mean that people in cities are more religious than people in suburbs, though. Instead, it probably means that some people are crossing county lines to go to church; maybe because they belong to a small denomination where there isn’t a suburban congregation, or maybe just because they prefer going to the downtown cathedral instead of a bland, boxy suburban church.
So, in large counties with multiple congressional districts, that problem is even further compounded; you have people not just crossing county lines to go to church, but crossing CD boundaries within the same county. There’s never going to be a way to achieve absolute precision, down to the individual level, with this kind of project; it’s really just intended to be a way to get a general sense of what traditions and denominations are most prevalent at the metro area or regional level, and how that shapes politics at that level.
