Ed Royce, the Republican who represents California’s 39th congressional district in Orange County, may be the most vulnerable member of Congress that you’ve never heard of. And that’s despite holding the high-profile position of chair of the House Foreign Affairs Committee; he keeps a low profile and manages not to say intemperate, attention-grabbing things, to the extent that he sometimes get misidentified as a "moderate" (which is insane, since DW-Nominate recently identified him as the 411th most-liberal member of the House, adjacent to a lot of Freedom Caucus members). He’s been there 1993 without facing serious opposition at any point, kept secure by Orange County’s dark-red status as California’s anchor of conservatism.
What’s threatening Royce this time is that Orange County’s stronghold status is crumbling before our eyes; Hillary Clinton was the first Democrat to win the county since FDR. Much of what’s driving that is the fact that Orange County, once a uniform slice of suburban white bread, is now much more diverse (42 percent non-Hispanic white, compared with 62 percent nationwide) and more educated (38 percent college-educated, compared with 31 percent nationwide) than the country at large. As we talked about a few weeks ago, both those demographic factors are more closely linked to voting Democratic than at any point in recent history. And Royce’s district has even more pronounced numbers (30 percent non-Hispanic white, 41 percent college-educated) than the broader county.
This isn’t the usual way to talk about how vulnerable a Congress member is; the most common data point that someone in my shoes reaches for is what the presidential numbers were, in that district or state, in the last election. By that measure, Royce’s district, which Clinton won 52-43 in 2016 (and which was won by Romney, 51-47, in 2012) isn’t the bluest district held by a Republican; with Ileana Ros-Lehtinen having already bailed out on FL-27 (which Clinton won 59-39), the GOP incumbent with the bluest district is now FL-26’s Carlos Curbelo, at 57-41. Once we get closer to the election, we start talking more about fundraising, candidate strength, and finally district-level polling, if that’s ever available.
But demographics is an important consideration as well: more than anything, it tells us where we're going, more so than where we are right now. More attention to demographic crosstabs in national polls that showed terrible numbers for Democrats among white non-college voters, for instance, might have been a useful warning light that something might have been off with those statewide polls (that failed to weight by education) that showed Clinton with significant (and incorrect) leads in Midwestern states. With that in mind, let’s take a look at the 2018 House battlefield, purely from a demographic standpoint without looking at presidential numbers.
If you go back and take another look at my article from two weeks ago, you’ll see that if you break the nation’s counties down into quintiles, and arrange them from most diverse to least diverse, the more diverse counties have a clearly stronger Democratic lean that the less diverse counties, and the more educated counties have a stronger Democratic lean the less educated counties. When you create a grid where you overlap the counties, that helps us pinpoint the places where there’s a synergy between race and education, which not coincidentally moved sharply in the Democratic direction in 2016, and the places where there’s not much of either of those characteristics, which were the ones that decisively moved in the Republican direction in 2016.
Counties are a very useful unit of analysis because they’re granular (except for the nation’s most populous counties, unfortunately; Los Angeles County, for instance, has more people than most states), and also because their boundaries usually stay fixed, so you can do apples-to-apples analysis across the decades. However, electoral power isn’t distributed according to county lines, so talking about counties has kind of a “so what?” aspect to it; congressional districts are in some ways more important, in that their gerrymandered boundaries directly shape what Congress is going to look like, and also because they let us peer further under the hood in mega-counties like Los Angeles or Orange Counties.
So, with that in mind, I’ve performed the same feat of data-wrangling with the nation’s 435 CDs, breaking them up into quintiles (based on the population needed to represent 1/5th of the nation, rather than, say, merely 87 of each; their mostly-uniform population, however, means that most quintiles do, in fact, have 86 to 88 CDs in them). Despite the way that CDs get sliced-and-diced for political advantage while counties are, in theory, a more neutral construct, the break-points between the quintiles still falls in almost the same places in both analyses.
Very high diversity: 3.1-39.9 percent non-Hispanic white for CDs; 0.9-41.1 percent for counties
High diversity: 39.9-60.2 percent for CDs; 41.2-57.6 percent for counties
Average diversity: 60.3-71.2 percent for CDs; 57.7-71.3 percent for counties
Low diversity: 71.3-82.2 percent for CDs; 71.4-84.1 percent for counties
Very low diversity: 82.4-96.0 percent for CDs; 84.2-99.8 percent for counties
Very high education: 71.5-38.7 percent for CDs; 78.8-37.7 percent for counties
High education: 38.6-31.0 percent for CDs; 37.6-30.9 percent for counties
Average education: 30.9-26.1 percent for CDs; 30.8-27.0 for counties
Low education: 26.0-21.6 percent for CDs; 26.9-19.8 percent for counties
Very low education: 21.5-9.1 percent for CDs; 19.7-1.9 percent for counties
As with what we saw when I looked at counties, the CDs that have high diversity and high education tend to be the ones that are likeliest to have voted for Democrats at both the presidential and House levels. In CDs where the results were different in 2012 and 2016, most likely the CD went for Romney in 2012 but Clinton in 2016, so they’re moving in our direction. The House, however, tends to be a bit of a trailing indicator; in other words, in the high diversity/high education quadrant, the Romney/Clinton CDs (like CA-39, as well as other Orange County districts like CA-45 and CA-48, and Atlanta or Dallas-area suburban districts like GA-07 and TX-32) are still electing GOP House members … for now.
The unfortunate flip side, of course, is that in the low diversity/low education quadrant, most of the CDs are already solidly GOP. (When I say “solid,” the criteria is they voted for the same party in both the 2012 and 2016 presidential elections, and the 2014 and 2016 House elections.) Some of them were historically Democratic as recently as 2010, but haven’t been since then. In addition, of the ones where the results are mixed, in that quadrant the usual move is from voting Dem in 2012 or 2014, to voting GOP in 2016.
Below, let’s take a look at what districts go in each bucket. You’ll notice they’re arranged by diversity quintile first, and education quintile second. I’ll break in after each diversity quintile with some further commentary.
VERY HIGH DIVERSITY, VERY HIGH EDUCATION
10 CDs: 9 solid Dem (90 percent); 1 mixed (10 percent)
Solid Dem: CA-13 (Lee), CA-14 (Speier), CA-15 (Swalwell), CA-17 (Khanna), CA-27 (Chu), DC-AL (Holmes Norton), GA-05 (Lewis), IL-07 (Davis), NY-16 (Engel)
Dem in 2016 prez; GOP in 2012 prez, 2016 and 2014 House: CA-39 (Royce)
VERY HIGH DIVERSITY, HIGH EDUCATION
14 CDs: 13 solid Dem (93 percent); 1 mixed (7 percent)
Solid Dem: CA-19 (Lofgren), CA-37 (Bass), CA-53 (Davis), HI-01 (Hanabusa), MD-04 (Brown), MD-07 (Cummings), NJ-08 (Sires), NJ-09 (Pascrell), NY-06 (Meng), NY-07 (Velazquez), NY-09 (Clarke), NY-13 (Espaillat), PA-02 (Evans)
Dem in 2016 and 2012 prez; GOP in 2016 and 2014 House: FL-27 (Ros-Lehtinen)
VERY HIGH DIVERSITY, AVERAGE EDUCATION
12 CDs: 11 solid Dem (92 percent); 1 mixed (8 percent)
Solid Dem: CA-06 (Matsui), CA-20 (Panetta), CA-47 (Lowenthal), GA-04 (Johnson), GA-13 (Scott), HI-02 (Gabbard), IL-01 (Rush), MI-14 (Lawrence), NJ-10 (Payne), NM-03 (Lujan), NY-08 (Jeffries)
Dem in 2016 and 2012 prez and 2016 House; GOP in 2014 House: FL-10 (Demings)
VERY HIGH DIVERSITY, LOW EDUCATION
19 CDs: 17 solid Dem (89 percent); 1 mixed (5 percent); 1 solid GOP (5 percent)
Solid Dem: CA-31 (Aguilar), CA-34 (vacant, ex-Becerra), CA-38 (Sanchez), CA-43 (Waters), FL-09 (Soto), FL-20 (Hastings), IL-02 (Kelly), IL-04 (Gutierrez), LA-02 (Richmond), NY-05 (Meeks), NY-14 (Crowley), OH-11 (Fudge), PA-01 (Brady), TN-09 (Cohen), TX-09 (Green), TX-16 (O’Rourke), TX-20 (Castro)
Dem in 2016 and 2012 prez; GOP in 2016 and 2014 House: FL-26 (Curbelo)
Solid GOP: FL-25 (Diaz-Balart)
VERY HIGH DIVERSITY, VERY LOW EDUCATION
32 CDs: 29 solid Dem (91 percent); 2 mixed (6 percent); 1 solid GOP (3 percent)
Solid Dem: AL-07 (Sewell), AZ-03 (Grijalva), AZ-07 (Gallego), CA-09 (McNerney), CA-16 (Costa), CA-29 (Cardenas), CA-32 (Napolitano), CA-35 (Torres), CA-40 (Roybal-Allard), CA-41 (Takano), CA-44 (Barragan), CA-46 (Correa), CA-51 (Vargas), FL-05 (Lawson), FL-24 (Wilson), GA-02 (Bishop), MI-13 (Conyers), MS-02 (Thompson), NV-01 (Titus), NY-15 (Serrano), SC-06 (Clyburn), TX-15 (Gonzalez), TX-18 (Jackson-Lee), TX-28 (Cuellar), TX-29 (Green), TX-30 (Johnson), TX-33 (Veasey), TX-34 (Vela), TX-35 (Doggett)
Dem in 2016 and 2012 prez; GOP in 2016 and 2014 House: CA-21 (Valadao)
Dem in 2016 prez; GOP in 2012 prez and 2016 and 2014 House: TX-23 (Hurd)
Solid GOP: NM-02 (Pearce)
One thing that you might have noticed in my county-level article was that there was a real drop-off in Democratic vote share within the “Very High Diversity” column as you went down education levels; in the Very High Diversity/Very High Education bucket, the Dem vote share in 2016 was 74 percent, but it was “only” 58 percent in the Very High Diversity/Very Low Education bucket, where you have lots of people of color but pretty low turnout. That’s not so much the case here with congressional districts, with Dems winning around 90 percent of the seats in all buckets in this column … but I don’t think that tells us that “education doesn’t matter” here.
Instead, what’s happening is that the Very High Diversity/Very Low Education CDs (and the Low Education ones as well, in many cases) are the ones that are heavily gerrymandered to produce pro-Democratic results; they concentrate many of a state’s black or Hispanic voters in one place, and the flipside is that makes it even easier for Republicans to win the low diversity CDs that cover the rest of the state.
Keep in mind, though, that isn’t necessarily the most nefarious sort of gerrymandering at work; these CDs are, for the most part, compelled by the Voting Rights Act, which compels the creation of districts as necessary to maximize minority representation. However, you still don’t hear the GOP complaining too much about this requirement, because, in fact, concentrating non-white voters in a few CDs makes it easier to bleach the remaining CDs in a state.
There are also a few outliers here, especially the three mostly Cuban districts in south Florida (FL-25, 26, and 27); these districts, you’ll notice, are steadily moving in the Dems’ direction anyway, as younger Cuban-Americans tend to be less motivated by anti-Communist rhetoric than their ancestors were, as well as these districts having a larger variety of non-Cuban Hispanic immigrants moving in. There are also a few mostly Hispanic districts (CA-21, TX-23, NM-02) where Hispanic turnout is especially low, combined with reflexive conservatism among those districts’ white minorities.
HIGH DIVERSITY, VERY HIGH EDUCATION
29 CDs: 19 solid Dem (66 percent); 5 mixed (17 percent); 5 solid GOP (17 percent)
Solid Dem: CA-11 (De Saulnier), CA-12 (Pelosi), CA-18 (Eshoo), CA-28 (Schiff), CA-30 (Sherman), CA-52 (Peters), CO-01 (De Gette), FL-23 (Wasserman Schultz), MD-03 (Sarbanes), MD-06 (Delaney), MD-08 (Raskin), MA-07 (Capuano), NJ-12 (Watson Coleman), NY-04 (Rice), NY-17 (Lowey), NC-04 (Price), VA-08 (Beyer), VA-11 (Connolly), WA-09 (Smith)
Dem in 2016 prez; GOP in 2012 prez and 2016 and 2014 House: CA-45 (Walters), CA-48 (Rohrabacher), TX-07 (Culberson), TX-32 (Sessions)
Dem in 2016 and 2012 prez and 2016 House; GOP in 2014 House: IL-10 (Schneider)
Solid GOP: GA-07 (Woodall), TX-02 (Poe), TX-03 (Johnson), TX-22 (Olson), TX-24 (Marchant)
HIGH DIVERSITY, HIGH EDUCATION
22 CDs: 16 solid Dem (73 percent); 3 mixed (14 percent); 3 solid GOP (14 percent)
Solid Dem: AZ-09 (Sinema), CA-07 (Bera), CA-24 (Carbajal), CA-26 (Brownley), FL-14 (Castor), FL-21 (Frankel), FL-22 (Deutch), IL-08 (Krishnamoorthi), IL-11 (Foster), MD-02 (Ruppersberger), MD-05 (Hoyer), NJ-06 (Pallone), NM-01 (Lujan Grisham), NC-12 (Adams), PA-13 (Boyle), TN-05 (Cooper)
Dem in 2016 and 2012 prez and 2016 House; GOP in 2014 House: FL-07 (Murphy)
Dem in 2012 prez and 2016 House; GOP in 2016 prez and 2014 House: NV-03 (Rosen)
Dem in 2012 prez; GOP in 2016 prez and 2016 and 2014 House: NY-11 (Donovan)
Solid GOP: NC-09 (Pittenger), TX-10 (McCaul), TX-31 (Carter)
HIGH DIVERSITY, AVERAGE EDUCATION
13 CDs: 7 solid Dem (54 percent); 2 mixed (15 percent); 4 solid GOP (31 percent)
Solid Dem: CA-05 (Thompson), IL-03 (Lipinski), MO-01 (Clay), NC-01 (Butterfield), OH-03 (Beatty), VA-03 (Scott), WI-04 (Moore)
Dem in 2016 prez; GOP in 2012 prez and 2016 and 2014 House: CA-25 (Knight)
Dem in 2016 and 2012 prez and 2016 House; GOP in 2014 House: VA-04 (McEachin)
Solid GOP: CA-50 (Hunter), OK-05 (Russell), TX-06 (Barton), TX-17 (Flores)
HIGH DIVERSITY, LOW EDUCATION
8 CDs: 2 solid Dem (25 percent); 1 mixed (13 percent); 5 solid GOP (63 percent)
Solid Dem: CA-03 (Garamendi), IN-07 (Carson)
Dem in 2016 and 2014 House; GOP in 2016 and 2012 prez: AZ-01 (O’Halleran)
Solid GOP: CA-22 (Nunes), CA-42 (Calvert), FL-15 (Ross), GA-01 (Carter), TX-14 (Weber)
HIGH DIVERSITY, VERY LOW EDUCATION
13 CDs: 1 solid Dem (8 percent); 2 mixed (15 percent); 10 solid GOP (77 percent)
Solid Dem: CA-36 (Ruiz)
Dem in 2016 and 2012 prez; GOP in 2016 and 2014 House: CA-10 (Denham)
Dem in 2016 and 2012 prez and 2016 House; GOP in 2014 House: NV-04 (Kihuen)
Solid GOP: CA-08 (Cook), CA-23 (McCarthy), GA-12 (Allen), LA-04 (Johnson), LA-05 (Abraham), TX-05 (Hensarling), TX-11 (Conaway), TX-19 (Arrington), TX-27 (Farenthold), WA-04 (Newhouse)
In the higher-education buckets within the “High Diversity” column, you see mostly Dem-friendly CDs in suburban areas of the northeast (and California and Florida). The few exceptions tend to be those Romney-to-Clinton CDs in Orange County, and the Atlanta, Dallas, and Houston suburbs, that I was talking about previously.
In the lower-education, buckets, though, you start seeing more rural areas that are reliably GOP. These places tend to be about half-white and half-minority, but low turnout among the minority half of the electorate hurts Dem chances here.
AVERAGE DIVERSITY, VERY HIGH EDUCATION
21 CDs: 11 solid Dem (52 percent); 4 mixed (19 percent); 6 solid GOP (29 percent)
Solid Dem: CA-02 (Huffman), CA-33 (Lieu), CT-04 (Himes), IL-05 (Quigley), IL-09 (Schakowsky), MN-04 (McCollum), MN-05 (Ellison), NY-03 (Suozzi), NY-10 (Nadler), NY-12 (Maloney), WA-07 (Jayapal)
Dem in 2016 prez; GOP in 2012 prez and 2016 and 2014 House: CA-49 (Issa), VA-10 (Comstock)
Dem in 2016 and 2012 prez; GOP in 2016 and 2014 House: CO-06 (Coffman)
Dem in 2016 House; GOP in 2016 and 2012 prez and 2014 House: NJ-05 (Gottheimer)
Solid GOP: AZ-06 (Schweikert), GA-06 (vacant, ex-Price), GA-11 (Loudermilk), SC-01 (Sanford), TX-21 (Smith), TX-26 (Burgess)
AVERAGE DIVERSITY, HIGH EDUCATION
22 CDs: 9 solid Dem (41 percent); 3 mixed (14 percent); 10 solid GOP (45 percent)
Solid Dem: CO-07 (Perlmutter), CT-01 (Larson), CT-03 (De Lauro), CT-05 (Esty), KY-03 (Yarmuth), MA-03 (Tsongas), NY-25 (Slaughter), PA-14 (Doyle), RI-01 (Cicilline)
Dem in 2016 prez; GOP in 2012 prez and 2016 and 2014 House: AZ-02 (McSally)
Dem in 2012 prez and 2016 and 2014 House; GOP in 2016 prez: NY-18 (Maloney)
Dem in 2016 and 2012 prez; GOP in 2016 and 2014 House: WA-08 (Reichert)
Solid GOP: AZ-05 (Biggs), FL-16 (Buchanan), FL-19 (Rooney), NC-02 (Holding), OH-01 (Chabot), SC-02 (Wilson), TX-25 (Williams), VA-01 (Wittman), VA-02 (Taylor), VA-07 (Brat)
AVERAGE DIVERSITY, AVERAGE EDUCATION
19 CDs: 5 solid Dem (26 percent); 3 mixed (16 percent); 11 solid GOP (58 percent)
Solid Dem: DE-AL (Blunt Rochester), MO-05 (Cleaver), NJ-01 (Norcross), NY-26 (Higgins), WA-10 (Heck)
Dem in 2016 and 2012 prez and 2016 House; GOP in 2014 House: FL-13 (Crist)
Dem in 2014 House; GOP in 2016 and 2012 prez and 2016 House: FL-18 (Mast)
Dem in 2012 prez; GOP in 2016 prez and 2016 and 2014 House: NY-02 (King)
Solid GOP: AK-AL (Young), AZ-08 (Franks), AR-02 (Hill), CO-03 (Tipton), FL-03 (Yoho), LA-06 (Graves), NC-13 (Budd), OK-01 (Bridenstine), SC-04 (Gowdy), TX-08 (Brady), TX-12 (Granger)
AVERAGE DIVERSITY, LOW EDUCATION
13 CDs: 1 solid Dem (8 percent), 1 mixed (8 percent), 11 solid GOP (85 percent)
Solid Dem: IN-01 (Visclosky)
Dem in 2012 prez; GOP in 2016 prez and 2016 and 2014 House: NJ-02 (Lo Biondo)
Solid GOP: AL-01 (Byrne), GA-03 (Ferguson), GA-10 (Hice), MS-03 (Harper), NV-02 (Amodei), NC-03 (Jones), NC-06 (Walker), NC-07 (Rouzer), NC-08 (Hudson), OK-04 (Cole), SC-05 (vacant, ex-Mulvaney)
AVERAGE DIVERSITY, VERY LOW EDUCATION
12 CDs: 1 solid Dem (8 percent), 11 solid GOP (92 percent)
Solid Dem: OH-09 (Kaptur)
Solid GOP: AL-02 (Roby), AL-03 (Rogers), GA-08 (Scott), LA-03 (Higgins), MS-01 (Kelly), MS-04 (Palazzo), OK-02 (Mullin), SC-07 (Rice), TX-01 (Gohmert), TX-13 (Thornberry), TX-36 (Babin)
The high-education CDs within the Average Diversity column tend to be pretty evenly split between the Democrats and the GOP, but if you look closely, you start to notice a regional pattern becoming pretty pronounced: the ones in the northern half of the country are Dem-friendly, while the ones in the southern half of the country are very Republican. And as you move to the low-education end of the column, you’ll notice that almost all of the CDs are southern, in the first place. The few remaining northern ones tend to stay Democratic even at the low-education end (like IN-01 in northwestern Indiana and OH-09 in Toledo), though that’s partly to residual organized labor strength there.
LOW DIVERSITY, VERY HIGH EDUCATION
16 CDs: 6 solid Dem (38 percent), 4 mixed (25 percent), 6 solid GOP (38 percent)
Solid Dem: MA-05 (Clark), MA-06 (Moulton), MA-08 (Lynch), OR-01 (Bonamici), OR-03 (Blumenauer), WA-01 (Del Bene)
Dem in 2016 prez; GOP in 2012 prez and 2016 and 2014 House: IL-06 (Roskam), KS-03 (Yoder), NJ-07 (Lance)
Dem in 2016 and 2012 prez; GOP in 2016 and 2014 House: MN-03 (Paulsen)
Solid GOP: FL-04 (Rutherford), IL-14 (Hultgren), IN-05 (Brooks), MI-11 (Trott), NJ-04 (Smith), NJ-11 (Frelinghuysen)
LOW DIVERSITY, HIGH EDUCATION
17 CDs: 6 solid Dem (35 percent), 4 mixed (24 percent), 7 solid GOP (41 percent)
Solid Dem: CT-02 (Courtney), MA-02 (McGovern), MI-12 (Dingell), NY-20 (Tonko), RI-02 (Langevin), WA-02 (Larsen)
Dem in 2012 prez; GOP in 2016 prez and 2016 and 2014 House: MN-02 (Lewis), NJ-03 (MacArthur), NY-01 (Zeldin)
Dem in 2014 House; GOP in 2016 and 2012 prez and 2016 House: NE-02 (Bacon)
Solid GOP: AL-05 (Brooks), AL-06 (Palmer), CA-04 (McClintock), CO-04 (Buck), CO-05 (Lamborn), MD-01 (Harris), UT-02 (Stewart)
LOW DIVERSITY, AVERAGE EDUCATION
26 CDs: 4 solid Dem (15 percent), 1 mixed (4 percent), 21 solid GOP (81 percent)
Solid Dem: MA-01 (Neal), MI-09 (Levin), OR-05 (Schrader), WA-06 (Kilmer)
Dem in 2016 and 2012 prez; GOP in 2016 and 2014 House: NY-24 (Katko)
Solid GOP: AR-03 (Womack), FL-01 (Gaetz), FL-08 (Posey), FL-12 (Bilirakis), ID-02 (Simpson), IL-13 (Davis), KS-04 (Estes), KY-06 (Barr), LA-01 (Scalise), MI-03 (Amash), MI-06 (Upton), NE-01 (Fortenberry), OH-10 (Turner), PA-15 (Dent), TN-07 (Blackburn), TN-08 (Kustoff), UT-01 (Bishop), UT-04 (Love), VA-05 (Garrett), VA-06 (Goodlatte), WI-01 (Ryan)
LOW DIVERSITY, LOW EDUCATION
16 CDs: 3 mixed (19 percent), 13 solid GOP (81 percent)
Dem in 2014 House; GOP in 2016 and 2012 prez and 2016 House: FL-02 (Dunn)
Dem in 2012 prez; GOP in 2016 prez and 2016 and 2014 House: IL-12 (Bost)
Dem in 2012 prez and 2016 and 2014 House; GOP in 2016 prez: PA-17 (Cartwright)
Solid GOP: CA-01 (La Malfa), FL-06 (De Santis), GA-09 (Collins), KS-01 (Marshall), MI-02 (Huizenga), NC-05 (Foxx), NC-10 (McHenry), OK-03 (Lucas), OR-02 (Walden), PA-04 (Perry), PA-16 (Smucker), TN-03 (Fleischmann), WA-03 (Herrera Beutler)
LOW DIVERSITY, VERY LOW EDUCATION
13 CDs: 2 solid Dem (15 percent), 1 mixed (8 percent), 10 solid GOP (77 percent)
Solid Dem: MI-05 (Kildee), OH-13 (Ryan)
Dem in 2012 prez and 2016 and 2014 House; GOP in 2016 prez: IL-17 (Bustos)
Solid GOP: AZ-04 (Gosar), AR-01 (Crawford), AR-04 (Westerman), FL-11 (Webster), FL-17 (Rooney), GA-14 (Graves), IN-02 (Walorski), SC-03 (Duncan), TN-04 (Des Jarlais), TX-04 (Ratcliffe)
Again, as you look at the “Low Diversity” column, you’ll see that at the high-education levels, there’s about an even split between Democratic and Republican districts (with a noticeable difference coming between northeastern ones and Midwestern ones, which are likelier to go GOP). At the low-education levels, it’s more difficult for the Dems to break through the clutter, though again that’s largely because most of these districts are southern, and the northern ones where there’s an organized labor presence (IL-17, MI-05, OH-13) are the lone exceptions.
VERY LOW DIVERSITY, VERY HIGH EDUCATION
10 CDs: 3 solid Dem (30 percent), 2 mixed (20 percent), 5 solid GOP (50 percent)
Solid Dem: CO-02 (Polis), MA-04 (Kennedy), WI-02 (Pocan)
Dem in 2016 prez; GOP in 2012 prez and 2016 and 2014 House: PA-06 (Costello), PA-07 (Meehan)
Solid GOP: MI-08 (Bishop), MO-02 (Wagner), OH-12 (Tiberi), PA-08 (Fitzpatrick), UT-03 (Chaffetz)
VERY LOW DIVERSITY, HIGH EDUCATION
12 CDs: 4 solid Dem (33 percent), 2 mixed (17 percent), 6 solid GOP (50 percent)
Solid Dem: ME-01 (Pingree), MA-09 (Keating), NH-02 (Kuster), VT-AL (Welch)
Dem in 2012 prez; GOP in 2016 prez and 2016 and 2014 House: IA-03 (Young)
Dem in 2012 prez and 2016 House; GOP in 2016 prez and 2014 House: NH-01 (Shea-Porter)
Solid GOP: IL-18 (La Hood), OH-14 (Joyce), OH-16 (Renacci), PA-12 (Rothfus), PA-18 (Murphy), WI-05 (Sensenbrenner)
VERY LOW DIVERSITY, AVERAGE EDUCATION
17 CDs: 3 mixed (18 percent), 14 solid GOP (82 percent)
Dem in 2012 prez and 2016 and 2014 House; GOP in 2016 prez: IA-02 (Loebsack), MN-01 (Walz)
Dem in 2012 prez; GOP in 2016 prez and 2016 and 2014 House: NY-19 (Faso)
Solid GOP: KS-02 (Jenkins), KY-04 (Massie), MN-06 (Emmer), MO-06 (Graves), MT-AL (vacant, ex-Zinke), NY-23 (Reed), NY-27 (Collins), ND-AL (Cramer), OH-02 (Wenstrup), OH-15 (Stivers), SD-AL (Noem), TN-02 (Duncan), WA-05 (McMorris Rodgers), WY-AL (Cheney)
VERY LOW DIVERSITY, LOW EDUCATION
31 CDs: 1 solid Dem (3 percent), 6 mixed (19 percent), 24 solid GOP (77 percent)
Solid Dem: OR-04 (De Fazio)
Dem in 2012 prez; GOP in 2016 prez and 2016 and 2014 House: IA-01 (Blum), ME-02 (Poliquin), NY-21 (Stefanik)
Dem in 2016 and 2014 House; GOP in 2016 and 2012 prez: MN-07 (Peterson)
Dem in 2012 prez and 2016 and 2014 House; GOP in 2016 prez: MN-08 (Nolan), WI-03 (Kind)
Solid GOP: ID-01 (Labrador), IL-16 (Kinzinger), IN-03 (Banks), IN-04 (Rokita), IN-09 (Hollingsworth), IA-04 (King), MI-01 (Bergman), MI-07 (Walberg), MI-10 (Mitchell), MO-03 (Luetkemeyer), MO-04 (Hartzler), MO-07 (Long), NE-03 (Smith), NY-22 (Tenney), NC-11 (Meadows), OH-05 (Latta), OH-08 (Davidson), PA-03 (Kelly), PA-05 (Thompson), PA-11 (Barletta), WV-01 (McKinley), WI-06 (Grothman), WI-07 (Duffy), WI-08 (Gallagher)
VERY LOW DIVERSITY, VERY LOW EDUCATION
19 CDs: 19 solid GOP (100 percent)
Solid GOP: AL-04 (Aderholt), IL-15 (Shimkus), IN-06 (Messer), IN-08 (Bucshon), KY-01 (Comer), KY-02 (Guthrie), KY-05 (Rogers), MI-04 (Moolenaar), MO-08 (Smith), OH-04 (Jordan), OH-06 (Johnson), OH-07 (Gibbs), PA-09 (Shuster), PA-10 (Marino), TN-01 (Roe), TN-06 (Black), VA-09 (Griffith), WV-02 (Mooney), WV-03 (Jenkins)
Finally, there is the “Very Low Diversity” column, where the Democrats have trouble winning even at the highest education levels. The ones where they do win tend to be in New England, along with some other districts (CO-02, WI-02) anchored by college towns.
As you move further down, there simply aren’t any Democrats in the Very Low diversity/Very Low education bucket at all, which is dominated by CDs in the rural Appalachian arc that runs from central Pennsylvania down to northern Alabama. There are still a few holdout districts in the Very Low diversity/Low education bucket with a long Dem tradition that probably won’t stay Democratic at the House level after the next retirement (like MN-07 and MN-08). Maybe the biggest overperformer, based on demographics alone, turns out to be Peter De Fazio in OR-04, though that district has some unique quirks that partly explain its relatively high Democratic vote share (the presence of two college towns, plus a lot of residual hippie-ness in some of its rural areas).
You might be wondering, in particular, what this all means for the three special elections that are looming on the horizon, which might offer some clues as to how well the Democrats will do in the 2018 midterm. There’s one district of the three that’s particularly well-positioned to move in the Democrats’ direction, and, as you can probably guess, it’s Georgia’s 6th district, which is in the Average Diversity/Very High Education bucket with a lot of other affluent suburbs that went Romney to Clinton. It’s not quite as diverse as its neighboring 7th district (which is in the High Diversity/Very High Education bucket), but with a growing upper-middle-class population that includes both a lot of people of color moving from the city as well as white voters moving from other states, it’s similar to Orange County and Texas’s blue-shifting suburbs. (Darrell Issa’s CA-49 is another huge opportunity in this same bucket.) That’s a big part of why Jon Ossoff is the likeliest, of the three special elections, to be joining the House soon.
By contrast, Montana’s at-large district is in the Very Low Diversity/Average Education bucket, where there aren’t any Democratic Representatives at all. (Comparable CDs are located nearby throughout the west, including WA-05, WY-AL, and ND-AL.) In a way, it’s surprising that we’re doing as well as we are here, with Rob Quist seemingly trailing by only a few points in the latest rounds of polls; that speaks to his skills as a candidate and a good message, as well as a lot of swing-voter dismay with the Trump administration so far. An upset loss here would probably be far more alarming to House Republicans than the somewhat-expected one in Georgia … though one thing that Montana does have (that suburban Atlanta doesn’t have, and that doesn't show up in demographic data) is a long history of ticket-splitting and willingness to vote for Democrats who make a show of their independence.
And finally, there’s the race in South Carolina’s 5th district, where Archie Parnell is our candidate. This could be more demographically promising than Montana, in that it’s in the Average Diversity column (but the Low Education row). The problem here is that the 5th (like its comparable districts in the south that have a mix of suburbs and rural areas, like GA-03, GA-10, and NC-08) is a good example of a very inelastic district, using a concept developed by Nate Silver.
In other words, it’s a very racially polarized district, where conservative white voters will reliably vote GOP and black voters are always going to vote Dem, and the white majority is going to prevail unless the Dem candidate can really juice black turnout and white turnout is badly depressed. (This contrasts with Montana, which is a very elastic state, in that it has a lot of swing voters who’ll respond to candidate quality and national trends.) One other thing worth pointing out, though, is that SC-05 was more recently held by a Democratic House members than MT-AL: conservative Dem John Spratt until 2010. (Though the problem there, of course, is that, facing an Obama White House, white voters in the south rebelled en masse in 2010 against the idea of voting for Blue Dog Democrats any more, and that hasn't bounced back yet.)
(Final boring methodology note for people who care deeply about Census data: The quintiles are based on the 2015 1-year American Community Survey. However, there’s one key quirk, concerning the new districts in Florida, North Carolina, and Virginia that debuted in the 2016 election. Data from those states for the new districts isn’t available through American Fact Finder but only through the Census’s MyCD widget. Unfortunately, MyCD doesn't give ethnicity data beyond simply “Hispanic” and “non-Hispanic,” without, say, telling us how many “non-Hispanic whites” are in each CD. So for those three states, I’ve had to reverse-engineer a “non-Hispanic white” number by starting with total population, subtracting Hispanics, and then subtracting blacks, Native Americans, Asians, Pacific Islanders, and “two or mores,” but not whites or “some others” (who are usually Hispanic). This runs the risk of misallocating “some others” who aren’t Hispanic, and blacks, Asians, etc. who are also Hispanic, but my experimenting suggests these results should be accurate within a one percent tolerance, which is not likely to affect which quintile the districts in these states end up in.)