The power of demographics in the 2018 vote

From left: Election judges Geraldine Berger, Marlise Riffel and Wayne Christiansen discuss election-related information Tuesday, November 6, 2018 at the Miners Memorial Building in Virginia, Minn.   Photo: Derek Montgomery for MPR News

From left: Election judges Geraldine Berger, Marlise Riffel and Wayne Christiansen discuss election-related information Tuesday, November 6, 2018 at the Miners Memorial Building in Virginia, Minn.

Photo: Derek Montgomery for MPR News


By Craig Helmstetter, Managing Partner

We (pretty much) saw that coming

The power of demographics in the 2018 vote 

Democrats took the U.S. House in this fall’s elections and women made historic gains. But who sent them there? Now that the election results are finally in, here is what the data from our Representing US project are telling us.


Did the results of the historic 2018 congressional election reflect, magnify, or overturn the demographic voting patterns from the 2016 elections?

In August we published an analysis of the previously elected House that confirmed much conventional wisdom concerning district demographics and their voting patterns: the older, less educated, more employed, and, especially, Whiter, a district is, the more likely it was to be represented by a Republican. And vice versa.

This time around we benefit not only from new election results, but also new and improved U.S. Census Bureau data. 2017 American Community Survey data, released in September, included several new tables on the citizens of voting age within each congressional district (which we promptly fed into our interactive Demographics Tool).

These new data show that the 2018 House elections are just as strongly related to district demographics as were the prior elections. Maybe more so:

  • Racial and ethnic diversity remained very strongly related to vote patterns. Ninety-three percent of the 98 districts where potential voters are “majority minority” voted Democrats in to the House, while 70 percent of districts where non-Hispanic whites make up at least 8 in 10 potential voters went for Republicans.

  • Age remained strongly related to party preference as well. Considering median age, nearly two-thirds of the nation’s “youngest” congressional districts voted in Democrats, while over half of “older” districts voted in Republicans.

  • Education and income appeared more strongly related to the party of the newly-elected representative party than was the case in the prior election. About three in five districts that have either lower educational attainment or lower median income voted in Republicans, while more than three in four districts with either a more highly educated or highly compensated electorate opted for a Democrat.

Note: Median age is for all district residents; all other characteristics data refers to citizens of voting age (potential voters). As of publication, North Carolina’s 9th District was undecided.  Data from 2017 American Community Survey and 2018 voting data

Note: Median age is for all district residents; all other characteristics data refers to citizens of voting age (potential voters). As of publication, North Carolina’s 9th District was undecided.

Data from 2017 American Community Survey and 2018 voting data


What about the suburbs?

One of the narratives that quickly emerged after the elections was that the nation’s suburbs handed the Democrats their new House majority. Although others have pointed out that Democrats’ gains were much more widespread than just the suburbs, the gains that actually flipped control happened predominantly in the suburbs.

Of the 169 districts designated as suburban by CityLab, 90 are currently held by Democrats, but as a result of the November elections that number will climb to 119. This swing of the suburbs, which tend to be more highly educated and have higher incomes than rural districts, largely explains the strengthened relationship between these markers of social class and party preference in the 2018 elections.

Note: As of publication, North Carolina’s 9th District was undecided.  APM Research Lab analysis of voting data, and CityLab’s Rural, Urban, and Suburban classifications.

Note: As of publication, North Carolina’s 9th District was undecided.

APM Research Lab analysis of voting data, and CityLab’s Rural, Urban, and Suburban classifications.

In the average suburban district 33 percent of potential voters have a bachelor’s degree or higher education, which is much closer to the educational profile of the average urban district than the average rural district (36 percent and 24 percent with bachelor’s degrees, respectively).

However, suburban districts are closer to rural than urban districts in terms of their racial demographics. An average of 65 percent of suburban potential voters identify as non-Hispanic white, compared with 79 percent in rural districts and 45 percent in urban districts.


Year of the woman

As we have already reported, a record-setting 102 congresswomen are headed to the upcoming 116th Congress’ House of Representatives. Who is sending them there?

In the 2018 elections women tended to be elected by more diverse, more educated, and more urban districts. About 40 percent of “majority minority” districts elected women. Many of those districts are urban districts and, when the new House is seated in January, women will represent nearly half of all the districts CityLab designates as urban.

If those characteristics sound similar to the characteristics of districts which elected Democrats, that is no mistake: Democrats nominated women in 184 districts, and nearly half of them won; whereas Republicans nominated women in 52, and only one-quarter won.

Note: Characteristics data refers to citizens of voting age (potential voters). (*) As of publication, North Carolina’s 9th District was undecided.  APM Research Lab analysis of 2017 American Community Survey data, 2018 voting data, and CityLab’s Rural, Urban, and Suburban classifications

Note: Characteristics data refers to citizens of voting age (potential voters). (*) As of publication, North Carolina’s 9th District was undecided.

APM Research Lab analysis of 2017 American Community Survey data, 2018 voting data, and CityLab’s Rural, Urban, and Suburban classifications


What does all of this hold for 2020 and beyond?

Current demographic patterns, including the nation’s rapidly increasing racial and ethnic diversity, slow-but-steady increases in educational attainment, and continued population growth in urban and suburban over rural areas, would seem to favor the election of even higher numbers of Democrats and women in years to come.

At the same time, however, this election saw many extremely close races: not counting North Carolina’s still-contested ninth district, 42 seats were won by less than a 5-percentage point margin, compared with only 20 in the prior election. Additionally, coalitions can and do shift in American politics.

While it is far too soon to prognosticate about the outcome of the 2020 elections, one thing is certain: voter demographics, the leanings of the suburbs, and the ascendance of women candidates will continue to be closely watched.

-Craig (@c_helmstetter)

Reactions? Please email us your thoughts or join the conversation on Twitter or Facebook. 


Craig Helmstetter