Weekly News

Can Any One Demographic Group Predict the House Elections?

September 13,2018 20:21

Every election cycle, political junkies fixate on one question: “Which demographic is the key for [insert my political party] winning [insert important election] this time?” And, as we get closer to Election Day, people will inevitably try to zero in ...


September 13, 2018 at 4:34 AM
Probably not.
Every election cycle, political junkies fixate on one question: “Which demographic is the key for [insert my political party] winning [insert important election] this time?” And, as we get closer to Election Day, people will inevitably try to zero in one the One Demographic Group that explains everything about the 2018 House elections.
That’s not the best approach.
Don’t get me wrong: I love granular analysis of demographic groups. I do that sort of analysis all the time, and I think it deepens our understanding of how America works. But if you really want to understand the 2018 House battleground, you can’t just look at one demographic, geographic or political group.
There’s no clear demographic pattern in the battleground districts
The 2016 election proved that education and race are some of the most important dividing lines in our politics. But competitive House districts have a cookie-cutter racial or educational profile.

This scatter shows all 435 House districts. The “competitive” ones are shown in black, and the uncompetitive races are shown in blue. A race is deemed competitive if it has an average handicapper rating of 1.5 to -1.5 (basically between Leans/Likely Democratic and Leans/Likely Republican). The Average Handicapper is simple—I devised it by looking at the three major handicappers (Cook Political Report, Inside Elections and Sabato’s Crystal Ball), translating their race ratings into numerical scores (e.g. if Cook rates a race as "solid Democratic" it gets a -3, "likely Republican" gets 2, "toss-up" gets zero, etc. with the same scale applied to all three handicappers) and averages the scores from each handicapper. The vertical axis is the percentage of adults over 25 with a college degree, and the horizontal axis is the white percentage.
The basic takeaway is that competitive races (i.e. races somewhere between likely/leans and toss-up for either party) are demographically scattered across the map. Some competitive House districts are largely white, some aren’t. Some are well-educated and others aren’t. Obviously not every type of district is competitive. I don’t know of any majority-black districts that are up for grabs, and most rural Appalachian districts are looking uncompetitive. And looking at the district-level demographics won’t tell the whole story (e.g. the gender gap is important even though men and women are relatively evenly distributed across districts).
But the point is that the battle for the House can’t yet be reduced to a GOP effort to keep new blue -ollar white voters in the fold or a Democratic war in the suburbs. The list of competitive districts includes Colorado’s 6 th District (Republican-held suburbs that voted for Hillary Clinton last time and Barack Obama twice before that), Minnesota’s 8 th (a blue-collar district with numerous Obama-to-Trump converts), Kentucky’s 6 th District (typically red central Kentucky), California’s 39 th (racially diverse Orange County), New York’s 22 nd District (upstate New York) and more. One party may end up overperforming with a specific demographic group, but the battleground currently encompasses a wide variety of districts.
There’s no broad regional pattern either
The battle for House control probably won’t be decided by one region, either. I’ve divided the House districts up by Census region (the North, South, Midwest and West divisions I used here) and tallied up the number of competitive districts in each. Each region had a roughly similar number of competitive districts

The message of the table is simple: There are roughly a dozen districts in each Census region that can be reasonably counted as competitive (i.e. using the same definition listed above) at this point according to my handicapper measure (again, it's capped at 1.5). The South has fewer competitive districts relative to its size than other regions. But overall, there’s still football to be played in every major region of the U.S.
Obviously it’s possible for a party to run up the score in one region and win the 2018 House elections. If Democrats win every competitive seat in California and New Jersey (plus winning a few seats due to the new Pennsylvania map) and hold their current seats, they’re about halfway to a majority.
But we don’t know how the elections are going to turn out. And ignoring any region could lead to significant, maybe unexpected defeats.
And not every swing district is perfectly purple
And, interestingly, not every competitive district is an evenly split, typically purple battleground

This plot shows districts that are, on average, rated as “likely Republican” “likely Democratic” or more competitive according to our average rating system on the vertical axis and Trump’s percentage of the vote there on the x-axis (color corresponds to which party currently holds the seat). The spread here is real. There are a few outliers (e.g. Republican Rep. Carlos Curbelo represents a Florida district that went heavily for Clinton and West Virginia’s 3rd District—ancestrally Democratic but Trump-loving coal country—both look competitive), but most of the districts gave Trump roughly 40 percent to 60 percent of the vote. That means they range from as red as Arkansas to as blue as Connecticut.
Presidential voting obviously matters—there’s a real relationship between Trump’s vote percentage and the race’s average rating (i.e. that as Trump vote goes up, the rating tends to get more Republican). But there’s real noise in that graphic. Incumbency, candidate quality, fund-raising, conflicting down-ballot party loyalty (i.e. areas that have a history of voting for Republicans or Democrats and still do on the local level but now vote for the other party at the top of the ticket), state and local factors, demographics, geography and more all push races in different directions.
It’s possible that in the end the results will fall perfectly on presidential lines—that is, Democrats will win all districts that gave Clinton X percentage of the vote or more and Republicans will every seat that gave her less than that threshold. But for the time being, it’s worth watching a wide variety of districts only some of which looked purple in 2016.

elections 2018 elections in hungary elections in hungary 2018 elections in sweden elections 2019 elections in spain elections usa elections in poland elections turkey 2018 elections in turkey

Share this article

Related videos

The 2018 Senate Elections if Certain Demographic Groups Voted
The 2018 Senate Elections if Certain Demographi...
The 2020 Election If the Democrats Don't Win the House + Senate in 2018
The 2020 Election If the Democrats Don't Win th...

DON'T MISS THIS STORIES