One of the more interesting victories in Democrats’ retaking of the House in 2018 was in New York’s 11th District — mostly Staten Island, meaning that New York City was suddenly entirely blue.

But while interesting, that success wasn’t a representative one. Instead, the Democrats’ achievement was a function of picking up more than two dozen seats in suburban districts. It was a reflection of the political mood of the moment: Moderate voters came out to vote in a rejection of President Donald Trump’s administration.

This year, the likely Republican majority — however narrow it turns out to be — will probably be a function of regaining ground in the same areas.

Four years ago, CityLab’s David Montgomery came up with a useful categorization of congressional districts by population density. It’s a six-category scale, from “pure urban” to “dense suburban” and “sparse suburban” and then on to “pure rural.” After congressional district boundaries were redrawn over the past 18 months, The Washington Post’s Lenny Bronner used Montgomery’s methodology to categorize the new lines in the same way.

This usefully allows us to compare how the two parties fared in each type of district over the past three cycles — including where each gained or lost seats. You can see the Democratic gains in the suburbs in 2018 in the top graph below. And you can see in the bottom graph, the one for 2022, that Republicans picked up at least eight seats in “sparse suburban” districts and 16 in districts categorized as to some extent suburban.

It’s important to note that there are still a number of outstanding districts to be called in 2022, most of them some mixture of suburban. There are also a lot of votes still to count. (The figures used in this article are from Cook Political Report’s running index.)

With that warning in place, we can also see how vote margins in House races shifted in each type of district since 2018. If we simply add up all the votes cast in each type of district for each party and compare them, we see that the overall two-party margin shifted uniformly to the right in House voting from 2018 to 2020 and then from 2020 to 2022.

We’ve also included the 2020 presidential margin in each district, compiled by Daily Kos. Interestingly, the presidential vote in more-rural and more- urban districts was less polarized than the House vote. To some extent, of course, this reflects an incumbent effect. Most of the uncontested races in 2022 were in districts rated “pure rural” or “rural-suburban mix.”

That this is the first election after a redistricting is important for a variety of reasons, including that it means district population sizes have been evened out. Consider the 2020 election: Places that had been magnets for new residents for the preceding eight years were drawn according to their 2010 population, while districts that had been losing residents were as well. The most recent election was (theoretically at least) much closer to distributing power evenly in the House, relative to district populations.

It also meant an election in which district lines were often drawn to specifically advantage one party or the other after the 2020 Census. One result is that the redistricting process itself accounts for the likely Republican majority. Another is that — again using incomplete vote totals — Republicans appear to have more “wasted” votes than Democrats. In other words, the cumulative number of votes cast for Republican candidates beyond what was needed to win was larger than the cumulative excess for Democrats. In 2018 and 2020, the opposite was true.

If you take out uncontested races (defined here as those a party won by at least 50 points), the Republicans still wasted more votes, 12 million to 8 million. Perceptions of what’s likely to happen in an election can inspire incumbents not to seek reelection or candidates to challenge in even unfriendly terrain. The belief that this would be a strong Republican year, in other words, might have contributed to more Republican challengers (reducing the Democratic wasted vote) or caused more Republicans to run unchallenged.

Revisiting the urban-rural split, there’s another interesting consideration. In “sparse suburban” districts, Republicans picked up eight seats — but gave up five. Across all suburban districts (meaning not “pure urban” or “pure rural”), Republicans netted six additional seats, the difference in holding a majority by itself. They gained four in “pure rural” areas, places where there were fewer to pick up.

But that the suburbs were contested, and not just a place where Republicans ran up victories, is the point. This was an election cycle largely dependent on the economy, abortion and concern about the stability of democracy. In the suburbs, those issues yielded a split decision.