Donald Trump’s July 14 tweets telling four women of color who are House members to “go back and help fix the … places from which they came” may be remembered as one of the defining moments of his presidency, one that laid bare the racist nature of his project to such an extent that even major media outlets had to call it what it was. It was also an important moment, though, in terms of inspiring many people to think a little harder about what it means to be a citizen of a nation where everyone, to some extent, is originally from somewhere else.
The tweets take on an added note of absurdity when you look beyond the surface outrage and start pondering more deeply what it means to be from somewhere, and if you were to go back to someplace else, how you would know where you’re supposed to go back to … and how long your ancestors have to have been here before you can simply say that you’re from here. (If you’re a person of color, the answer to that last question may be “never.”)
If Alexandria Ocasio-Cortez, for instance, were to “go back” to where her ancestors came from, she’d be going back to Puerto Rico … which (perhaps unbeknownst to Trump) is still part of the United States. Ayanna Pressley’s ancestors have been in the United States far longer than Trump’s ancestors, and records are not likely to exist that would help her to determine which specific country in Africa she should move to and start fixing.
Then there’s the question of Trump himself: He was full of criticisms of the country and of its then-president back in 2016, so maybe he should have gone back to where he came from and tried to fix that country instead. His mother was born in Scotland; his paternal grandparents emigrated from Germany. Assuming he was going to take his own advice, would he “go back” to Germany or to Scotland? Does the most recently emigrated-from country take precedence? (If so, congratulations, United Kingdom: It’s time for some fixing.)
Trump’s story isn’t that unusual, though. In 2017, 82 million Americans, over one-quarter of the population, reported having multiple ancestries; in other words, instead of being “from” one country of origin, they had ancestors of varying origins who married across national and ethnic lines. And Pressley’s story isn’t unusual, either. In 2017, 60 million Americans reported no ancestry at all, for either lack of documentation or a lack of family oral history that identifies a particular national origin (or maybe just because it was perceived as too invasive, difficult, or vague a question).
How do we know these arcane numbers? From the census, or more specifically, from the American Community Survey, a massive ongoing annual survey that takes the place of the more detailed questions that used to be asked on the now-defunct census long form (including the now notorious citizenship question). While the decennial short-form census itself asks simply about race (for instance, whether you’re white, black, or Asian), the ACS is where you would get further into the nitty-gritty and tell the Census Bureau whether you’re, say, of Polish, Ethiopian, or Hmong descent.
There’s a wealth of information about Americans’ various national or tribal origins in there, but unfortunately, it’s scattered across a variety of tables on the census’s byzantine website. (That Polish number, for instance, is on a totally different chart than the Hmong number; you have to find different charts beyond that if you’re looking for a Navajo number or a Samoan number or a Peruvian number.) But we at Daily Kos Elections have synthesized all those tables into one spreadsheet, which has the added advantage of letting us see what the most prevalent ancestry is in any particular congressional district, county, or state. We’re also using this new data set to test-drive my colleague Daniel Donner’s fantastic new congressional district hex map.
In fact, let’s start by taking a look at the map above (large version here), which shows the most prevalent ancestry or origin in each congressional district, with each district having an approximately equal area. This map isn’t a choropleth (a map that’s scaled to show the intensity of one particular characteristic). Instead, it uses a spectrum of colors to show whatever is the “most” in each district, regardless of whether it’s close to a monoculture (Texas’s 34th District in the Rio Grande Valley is the closest to unanimity; 81% of its residents are of Mexican origin) or is one in which a bunch of different ancestries are jostling for the top spot (take New Jersey’s 9th District as an example, which shows bright green for Italian, but is 11% Italian, 10% African American, 10% Dominican, and 8% Puerto Rican).
You’ll notice that in nearly half of the nation’s 435 congressional districts, the most prevalent ancestry is a non-European one. In 97 districts, African American ancestry is most common (yellow), while in 90 others, Mexican (red) is most common (no doubt you’ll notice that African American is heavily concentrated in the South, while Mexican is heavily concentrated in the Southwest). If you add in the small number of districts in which the most prevalent origin is Cherokee (1), Chinese (6), Cuban (3), Dominican (2), Indian (1), Japanese (1), Native Hawaiian (1), and Puerto Rican (4), the total is 206. That may seem inordinately large, but keep in mind that currently, in the U.S., only 61.5% of the population is non-Hispanic white, so that’s not far off.
Meanwhile, among the 229 of the 435 congressional districts (that’s 52%) where the plurality ancestry is European, by far the most common ancestry is German. It’s most prevalent in 152 districts, concentrated across much of the nation’s northern tier in the Midwest, the Northwest, and Pennsylvania. (In fact, German is the most common ancestry in the country as a whole: 14.1% of the nation claims either whole or partial German ancestry. That’s followed by African American at 11.9%, Mexican at 11.1%, Irish at 10.1%, and English at 7.5%, according to data from the Census Bureau’s 2012-2017 American Community Survey.)
That’s followed by Irish (24 districts), Italian (22 districts) (both almost entirely in the suburbs of cities in the Northeast), and, somewhat confusingly, American in 22 districts in the rural South. (American ancestry isn’t a mistake, though. It’s a real thing, according to demographers, that is heavily concentrated in Appalachia and other upland parts of the South; it’s considered to be something of a proxy for Scots-Irish ancestry, though it’s often used as a stand-in for wherever there tends to be a lot of white people who’ve been here long enough that they’ve lost any sense of any other country of origin.)
Despite the traditional description of the U.S. as an Anglo-Saxon nation, only seven districts have English as the most prevalent ancestry, and those are heavily concentrated in the most heavily Mormon parts of the country, in Idaho and Utah. One French district—Louisiana’s 1st District—and one Armenian district—California’s 28th District, in the Los Angeles area—round out the count.
Much of the disparity between a country that’s 61% white and a country where only 52% of its congressional districts have a plurality ancestry that’s European shows up because it’s something of an apples-to-oranges comparison. A number of districts, for instance, have a white majority (or plurality) when you look at race data, yet wind up with African American or Mexican as their most prevalent ancestry or origin. There’s a simple explanation for that: The white residents of a particular district may be split up into a variety of different European ancestries, but taken as a whole as white people are more numerous than other groups.
(One other detail to keep in mind, in terms of the apples-to-oranges comparison, is that when looking at race data, the total always adds up to 100% of the population; everybody has to choose something, even if it’s “some other race” or “two or more races.” When looking at ancestry data, in many places, the total number of selected ancestries is greater than the total population, because people with multiple ancestries can choose more than one. In some other places, the total number of selected ancestries is less than the total population, because, of course, people can also fill in no ancestry at all.)
By contrast, there aren’t any further subdivisions among, for instance, people who identify as Mexican. In a district that’s 50% non-Hispanic white and 40% Hispanic, you might find that when you zoom in further to look at ancestries and origins, the same district is 35% Mexican, but 15% German, 10% Irish, and 7% English, meaning that a district that’s majority-white when looked at through race is also plurality-Mexican when looked at through ancestry.
The same is true of the African American category, which requires a bit of methodological explanation. The census’s ancestry section includes a variety of options for various sub-Saharan African or West Indian ancestries, including just listing your ancestry as African. However, very few black respondents seem to use any of these options (even African); in mostly black parts of the country, there’s very little use of the ancestry question at all. (Case in point: If you look at Georgia’s 5th Congressional District, which overlaps largely with the city of Atlanta, but limit yourself to the conventional ancestry table, it would tell you that English is the most prevalent ancestry, even though that represents only 5% of the district population. English is probably not the first thing that comes to mind when most people visualize Atlanta.)
So our solution to this problem is to include the African American population from ACS race data as essentially another ancestry category—but to then subtract people who selected specific sub-Saharan African or West Indian ancestries from that number, so as not to double-count them. This is, admittedly, an imprecise method, but the results pass the smell test more effectively. The Atlanta-based 5th District, for instance, turns out to have an African American plurality at 56% of its population (in other words, 450,000 people who choose African American as their race, minus the 31,000 people who also chose an overlapping ancestry such as Ethiopian, Jamaican, or Nigerian, for a result of 419,000 people out of a total population of 751,000).
Let’s take a closer look at this phenomenon, using Alabama’s seven congressional districts as an example. If you go by race data alone, Alabama has six majority-white districts and one that is majority-black.
|
1st |
% |
2nd |
% |
3rd |
% |
4th |
% |
| AL-01 |
White |
65 |
Black |
28 |
Hispanic |
3 |
Asian |
1 |
| AL-02 |
White |
62 |
Black |
31 |
Hispanic |
4 |
Asian |
1 |
| AL-03 |
White |
68 |
Black |
25 |
Hispanic |
3 |
Asian |
2 |
| AL-04 |
White |
84 |
Black |
7 |
Hispanic |
6 |
Native Am. |
1 |
| AL-05 |
White |
73 |
Black |
17 |
Hispanic |
5 |
Asian |
2 |
| AL-06 |
White |
77 |
Black |
15 |
Hispanic |
5 |
Asian |
2 |
| AL-07 |
Black |
63 |
White |
32 |
Hispanic |
3 |
Asian |
1 |
But when you switch over to the aggregated ancestry data, many of the districts are now have an African American plurality; you also get a more nuanced look at who the white residents in each of these districts are.
|
1st |
% |
2nd |
% |
3rd |
% |
4th |
% |
| AL-01 |
African Am. |
28 |
American |
16 |
Irish |
8 |
German |
7 |
| AL-02 |
African Am. |
30 |
American |
19 |
English |
7 |
Irish |
7 |
| AL-03 |
African Am. |
25 |
American |
16 |
Irish |
8 |
English |
7 |
| AL-04 |
American |
20 |
Irish |
11 |
African Am. |
7 |
English |
7 |
| AL-05 |
African Am. |
17 |
American |
14 |
Irish |
10 |
English |
10 |
| AL-06 |
American |
15 |
African Am. |
15 |
Irish |
12 |
English |
10 |
| AL-07 |
African Am. |
62 |
American |
11 |
Irish |
4 |
English |
3 |
In Alabama’s two whitest districts (the 6th, in Birmingham’s suburbs, and the 4th, in the rural uplands north of Birmingham but south of Huntsville—the southernmost tip of Appalachia), there are enough white residents that American ancestry is most prevalent, despite the large pool of white residents being broken down into a variety of ancestries. But in the Huntsville area and in the southern half of the state, African American becomes the plurality ancestry. (The Voting Rights Act-compelled 7th, of course, remains heavily black either way.) Interestingly, the 4th is white enough that Irish becomes the second-most prevalent ancestry, with African American in third place. Not coincidentally, the combination of its whiteness and its Southern ruralness also makes AL-04 both the reddest district in the country at the presidential level and the most heavily evangelical district in the country.
As you can see from the chart of Alabama’s ancestries, you can go deeper and deeper, asking, for instance, what the second-most prevalent ancestry in each congressional district is. Things get a little more complicated-looking here, though. When you look at the map of second-most prevalent ancestries, you don’t see the same level of dominance of Mexican or African American. Instead, in the Southwest, you see a lot of German across the rural parts of both California and Texas. You get Asian ancestries in second place in many of California’s cities, while it’s African American in many of Texas’ cities.
Irish ends up in second place throughout most of the nation’s northern tier where German was in first place, with the exception of Minnesota and the Dakotas, where Norwegian moves up in dominance. In the Northeast, Italian and Irish simply trade places in many areas, though some of the most urban areas tend to have African American or Puerto Rican as a second category. You also start seeing some less-typical ancestries pop up in parts of the country where they’re a big part of the local fabric, such as Navajo in parts of Arizona and New Mexico, Polish in the Chicago and Detroit areas, Dutch in Western Michigan, Portuguese in coastal Massachusetts, and Haitian in parts of the Miami area.
By the time you drop down to the map of the third-most prevalent ancestry in each congressional district, it becomes harder to spot patterns, and instead the map starts looking like more of a crazy quilt. You do, however, still see how frequently Irish and English appear as third options, and the large number of Poles in Illinois and Wisconsin comes into sharper relief.
Other noteworthy items that pop up on the third-most prevalent map that you might not have been aware of include the large numbers of Jamaicans in Brooklyn and Queens, and the interesting prevalence of “Other Hispanic or Latino” in New Mexico. (Many Hispanic New Mexico residents have local roots that go back as far as the 16th century, predating both Mexico and the United States; some identify as Spanish-American, while others identify as Hispano or Nuevomexicano, options that are consolidated by the census.)
While knowing the most prevalent or second-most prevalent ancestry in a particular district is very interesting by itself, we can also slice and dice the data in the opposite direction. For example, you might be wondering, Well, now I know Pennsylvania’s 1st District has more Germans than any other ethnicity … but is it one of the most German districts in the country?
The answer is: No, it’s not, at all. Pennsylvania’s 1st District, in the suburbs north of Philadelphia, is 24.3% German, followed closely by 23.5% Irish and 15.5% Italian. That’s considerably more German than the country as a whole (which is 14% German), but it’s actually a smidge less German than Pennsylvania in its entirety, at 24.5% German. (The German percentages go up significantly as you move further west into rural Pennsylvania, away from Philadelphia.)
But you have to go a lot further west than that to find the most heavily German American districts in the country, to the Upper Midwest, especially the suburban and rural parts of Wisconsin, followed by Minnesota, the Dakotas, and Nebraska. Pennsylvania’s 1st District is down at 63rd place. Here are the top 5 congressional districts by German ancestry:
|
GERMAN |
% |
| WI-06 |
356,452 |
50.0 |
| WI-05 |
341,006 |
47.3 |
| WI-07 |
302,250 |
42.7 |
| WI-03 |
302,167 |
42.1 |
| WI-08 |
303,761 |
41.9 |
By comparison, Pennsylvania’s 1st District is close to the top for Irish ancestry among all the nation’s districts, even though German squeaks ahead of it for most-prevalent within the district. Irish Americans still tend to be found primarily in the Northeast, especially in the suburbs of Boston and Philadelphia.
|
IRISH |
% |
| MA-08 |
225,785 |
29.7 |
| MA-06 |
192,828 |
25.3 |
| MA-09 |
180,561 |
24.5 |
| PA-01 |
166,378 |
23.5 |
| PA-05 |
160,745 |
22.5 |
If you look at the prevalence maps above, you’ll notice that a lot of the same Northeast suburbs that have large Irish populations also have large Italian populations. However, there are some subtle differences when you look at which have the highest Italian percentages: Those are instead in the New York metropolitan area (especially on Long Island), and, if you look further down the list, in Connecticut and Rhode Island.
|
ITALIAN |
% |
| NY-01 |
189,721 |
26.3 |
| NY-02 |
183,175 |
25.4 |
| NY-11 |
180,275 |
24.6 |
| NJ-11 |
181,376 |
24.4 |
| NJ-04 |
167,685 |
22.6 |
“Wait,” you might be thinking, “I thought the Northeast suburbs were full of wealthy WASPs.” Well, they were at one point, but you’re maybe a hundred years behind, as subsequent waves of other ancestries have supplanted them as they joined the middle class and moved out of the cities where they grew up. If you’re looking for the highest percentages of English ancestry, it may surprise you that it’s mostly in the parts of the country that are also the most heavily Mormon. (That's partly because of Mormon proselytizing in England in the late 19th century, but also because many Mormons—who tend to be into genealogical research for religious reasons—can trace back their roots to 17th-century New England Puritans.)
|
ENGLISH |
% |
| UT-03 |
202,339 |
27.3 |
| UT-01 |
175,808 |
23.7 |
| UT-02 |
169,994 |
23.0 |
| UT-04 |
176,287 |
22.8 |
| ME-01 |
141,045 |
20.9 |
Also, as I mentioned earlier, you have a lot of people who probably have some English roots that were many generations ago, and their families didn’t keep track of that. Those tend to be the people who choose American ancestry, which, surprisingly, is the fourth-most common European ancestry, according to the census (after German, Irish, and English, but ahead of Italian, Polish, and French). And, again, they’re most prevalent in Appalachia.
|
AMERICAN |
% |
| VA-09 |
178,715 |
25.0 |
| KY-05 |
176,468 |
25.0 |
| TN-01 |
165,769 |
23.3 |
| GA-14 |
142,220 |
20.2 |
| AL-04 |
136,080 |
19.9 |
We can also, of course, use the congressional districts hex map to map which districts have the most of one particular ancestry, rather than which ancestry is most prevalent. In that case, we’d use the traditional choropleth-style map, which is scaled to show the intensity of one particular characteristic. For example, compare these choropleths showing which districts have the highest population of people with Finnish ancestry versus the ones with the most Hmong.
We chose these two rather unusual ancestries, rather than more-common ones, because comparing them makes an important point about how ancestries slowly disperse over time from the places where they clustered when they first arrived in the country.
Both the Finns and the Hmong were heavily concentrated in the Upper Midwest when they first arrived: the Finns over a hundred years ago to work in logging and iron mining, the Hmong arriving much later, mostly in the 1980s, in the wake of the Vietnam War, resettling as refugees. Comparing the two images, they look similar to what happens when you add a few drops of dye to water. With little passage of time, the drops still tend to be clustered in the places where you dropped them (for instance, in the case of the Hmong, that’s largely where they were first resettled, such as in the Minneapolis and Fresno metropolitan areas). With a much longer passage of time, they’ve diffused throughout the vessel (the Finns have moved throughout the Midwest and West, and it also looks like many of them have retired to Florida), though you can still see some distinct remains of the original concentration.
Finally, there’s one other thing you might be wondering, something along the lines of, “This is all kind of interesting, but why is this here at Daily Kos? This is a site about politics, not 23andMe.” Well, the reality is this: All politics, at some level or another, is identity politics. Republicanism is all about identity politics: It’s the politics of whiteness.
This can play out in a crowded primary, where a candidate might be looking for any possible small advantage to cobble together enough of a coalition to win. If so, having the “right” last name can be an important advantage for consolidating support. Why, for instance, do you think the representatives in Michigan’s 2nd and 4th congressional districts are named Huizenga and Moolenaar instead of O’Malley or Vasquez? It’s partly because they are two of the most heavily Dutch American districts in the country.
Knowing a district’s ancestry composition also can add a level of nuance to understanding a district’s unique politics, beyond just the usual demographic markers like race and education. This is especially important when you see ancestries that tend to have their own unusual politics that are a bit different than what’s usually assumed, given their race or education. To keep using the example of Dutch Americans, districts with a high Dutch prevalence tend to be a lot more conservative than you’d otherwise expect, largely for religious reasons. That would tend to explain, for example, Iowa’s 4th District (the district that gives us Steve King), a much redder district than its similarly white, but much swingier, neighbors in Iowa. It also explains Michigan’s 3rd District in the Grand Rapids area, which is redder than other similarly well-educated suburban parts of Michigan.
Another example that you’re likely to be more familiar with is Cuban ancestry. While younger Cubans tend to be more progressive and more motivated by kitchen-table issues, many older Cubans are still driven by anti-communism and hawkish foreign policy. If you simply looked at race data on Florida’s 25th, 26th, and 27th districts in and around Miami, you’d assume they were Democratic strongholds because they have such large Latino majorities.
However, the most prevalent ancestry in those districts is Cuban; as a result, they’re only swingy districts at the presidential level (which is still significant progress from where they were in the 1980s or 1990s), and only two of those House seats flipped into Democratic hands in 2018, with the 25th still held by a Republican. (In fact, the Cuban presence is one of the main factors why Florida, a state with a below-median percentage of white residents, is still often a difficult state for Democratic candidates at the statewide level.)
At any rate, we’ve barely scratched the surface with the maps and tables that we’ve presented in this article. If you click through to the full spreadsheet, you can see the first- through fifth-most prevalent ancestries for every district, as well as the raw numbers for each of over 200 different ancestries, national origins, and Native American tribes. In fact, if you switch tabs, you can access that same level of information for all of the nation’s states, or, if you want to go in an even more granular direction, all of the more than 3,100 counties.
