I cannot wrap my head around why so many people in the US seemingly(?) vote for different parties (Democrats vs Republicans) in the elections for the House vs the Senate?

In the Senate the Republicans had historical midterm record gains, but in the House they were still reduced quite a bit?

Especially in this time of supposed "division" and "polarization", where many people say that the Midterm votes are driven by "how well the president did in his fast half term" and "hate against Trump" etc, how can so many people still vote Republican for the Senate but then Democrat for the House? What am I missing?

Or is it all rather due to some strange technicality of the voting system that the results become like this, even though most people in reality voted for the same party both for the Senate and the House?

  • Thanks for all the great answers! I selected the one that I deemed most complete and clear, but the other ones were also very good! Commented Nov 8, 2018 at 22:03
  • Were you really asking why people "split their votes" between parties instead of voting all Republicans or all Democrats?
    – RonJohn
    Commented May 27, 2019 at 17:54

4 Answers 4


In examining the possible causes of your confusion, it is important to understand some key factors

  • All 435 seats for the House of Representatives were up for re-election for the midterms. Key point: these seats are up for election every cycle.
  • Only 1/3 of the Senate was up for re-election this midterm cycle. These seats are only voted on every 6 years.
  • There were a few special elections for the Senate this Year.
  • While Senators are elected State wide, Representatives are limited to their voting District.

While it is possible some people split their ticket, voting for both Democrats and Republicans, it is more likely that the sum of votes for Democrat Representatives in a State would equal the votes for the Democrat Senator (in States where a Senator was up for re-election).

As an example, consider Missouri. Claire McCaskill, a Democratic Senator, was up for re-election this cycle in a State won by Trump in 2016 (57% to 38%). Missouri has 8 seats in the House of Representatives.

Senate Results
- Hawley: 1,245,732
- McCaskill: 1,101,377
- Other: 73,720

Total House Votes
- Republican: 1,318,481
- Democrat: 1,016,095
- Other: 58,793

- 27,460 MORE Votes were cast for the Senate Seat
- Hawley underperformed relative to the House Votes
- McCaskill overperformed relative to the House Votes

As a counterexample, consider West Virginia, and Joe Manchin. Another Democratic Senator up for re-election in a state won by Trump (68% to 26%). West Virginia has 3 House Seats.

Senate Results
- Morrisey: 269,872
- Manchin: 288,808
- Other: 24,231

Total House Votes
- Republican: 335,791
- Democrat: 232,856
- Other: 6,213

- 8,051 MORE Votes were cast for the Senate Seat
- Morrisey underperformed relative to the House Votes
- Manchin GREATLY overperformed relative to the House Votes

What could be a contributing factor in the different outcomes between these two results? Manchin voted for Supreme Court Justice Kavanaugh's confirmation, while McCaskill voted against it. Yet, even this isn't a full proof indication. Turn to Montana, with Democratic Senator Tester up for re-election. Fortunately for the math, Montana only has one seat in the house.

Senate Results
- Rosendale: 223,687
- Tester: 237,986
- Other: 13,563

Total House Votes
- Gianforte: 243,436
- Williams: 218,635
- Other: 13,373

- 208 fewer Votes were cast for the Senate Seat

Here, it appears to be strength of the incumbents that held Tester through the election. Similar analysis could be performed for the other 30 states that had a Senator on the ballot this year.

  • Nitpick: Montana numbers seem to be incomplete. The race was called with 97% of precincts reporting. Commented Nov 8, 2018 at 7:28

Members of the House serve two year terms while members of the Senate serve six year terms. This means that only 1/3 of the Senate is up for election every two years, while the entire House is up for a vote.

The distribution of Senate seats also disproportionately benefit states with low populations (Republican states), and the fact that only 1/3 of the chamber is up for election at one time means that the distribution of seats this year was particularly favorable to Republicans even beyond that.

  • 2
    While the first paragraph of this answer correctly addresses the question, I've down voted it for the narrative constructed by the second paragraph. Commented Nov 8, 2018 at 1:34
  • 1
    The benefit to the States by the Senate is not disproportionate, but functioning as designed. Commented Nov 8, 2018 at 1:40
  • 3
    @DrunkCynic If the "narrative" fits with reality, then what is the problem? What in particular are you actually contesting? That the Senate benefits states with low populations? Or that on the whole low population states tend to vote Republican? Or that choosing only 1/3 of the chamber can benefit one party or the other depending on the distribution of seats? All of these are in important elements in explaining how the Senate races differed from the House races.
    – Teleka
    Commented Nov 8, 2018 at 1:41
  • 1
    @DrunkCynic If someone asks why an airplane flies while a random piece of sheet metal does not, saying that the wings provide a disproportionate amount of lift relative to random arrangements of metal would not be wrong. It doesn't matter how it's designed the question was about the results of the election.
    – Teleka
    Commented Nov 8, 2018 at 1:45
  • 1
    @DrunkCynic "disproportionate" and "functioning as designed" are not mutually exclusive :) Commented Nov 8, 2018 at 7:32

I'm not sure based on your question, but it seems like you might not understand the difference between how the two chambers are filled. Every State gets two Senators, regardless of size, for a total of 100. Each State, gets a number of Representatives based on population, with the overall total being limited to 435 by the Apportionment Act of 1911. Seven states have only one Representative. Texas has thirty-six. Overall, Texas is a Republican held State and both of its Senators are Republican. However, of the thirty-six districts, thirteen voted in Democrats for their Representatives, up from eleven. So Texas added two Democrat Representatives while maintaining the same amount of Republican Senators.

Each District votes its Representative in by popular vote, and the Senator gets the popular vote for the whole state so that's how a state can vote in legislators for both parties.

  • If you look at a by sub-state region map of 2016 election (county or district maps), there were only five states had one party win every sub-state region in that particular state. Most states, especially the really big ones, typically have a not to be sneezed at minority region. Doing a comparative map view of districts, county, and states, can really be interesting.
    – hszmv
    Commented Nov 8, 2018 at 22:04
  • @hszmv Indeed. For the most part, the breakdown is urban vs. suburban/rural. Inner-city districts (which tend to be geographically small, but heavily populated) tend to vote blue in every state, whereas rural districts tend to vote red in every state. Suburban districts are in between, but tend to lean red, especially the higher-income suburbs. If you look at a map of House districts, you'll basically see a giant red country with a few specks of blue in the cities, even though said specks of blue happen to currently be a majority of House districts.
    – reirab
    Commented May 28, 2019 at 16:21

In the Senate the Republicans had historical midterm record gains

The Republicans gained at most three seats in the Senate (and possibly none, three races are still counting votes). This is not really a historic midterm gain except that midterms usually see the president's party losing seats. For contrast, the Republicans gained nine seats in 2014 and six in 2010.

In this particular year (2018), Republicans also had an advantage in that ten seats that Democrats held voted for Donald Trump in 2016. Only one seat that Republicans held voted for Hillary Clinton in 2016. So the Democrats were more exposed than the Republicans were. Democrats still won most of the Senate seats that were up for election: twenty-three of thirty-five. That counts the independents as Democrats. Regardless, that's 65% of the seats. Meanwhile, Democrats in the House, even if they win every uncalled seat, can only win 238 of 435 or 55%.

Democrats actually did better in Senate races than in House races. And that disparity may increase. Because if Democrats win the three remaining Senate seats and Republicans win the fifteen remaining House seats, that would be 74% and 51% respectively. 65% and 53% may be more likely, but the higher disparity is still possible.

Looking at the shift in seats makes it look different. The Democrats held a whopping twenty-six out of thirty-five Senate seats that were up for election. They could have lost eight and still won a majority of the seats that were up. But the Republicans hold forty-two of the sixty-five seats that were not up. If we look at the Senate as a whole, it's within 4% of an even split even if Republicans win every uncalled seat.

In the House, it was the Republicans who held a majority of the seats (and all the seats were up for election). So it looks like the Republicans lost a lot. But really, the Democrats remain within 5% of an even split.

The final disparity between the two chambers could be as small as 2%.

  • I've down voted this answer due to the dependency of its reasoning on outstanding races, and the limited coverage of possible outcomes. I intend to reverse this vote in response to more (if-then) clauses that fully cover the eventualities as they stand now, or a substantive edit when the last races resolve. Commented Nov 8, 2018 at 17:21
  • 1
    @DrunkCynic There is no dependency of my reasoning on outstanding races. I quite clearly explain that the results cover a range. And the entirety of the range (from lowest to highest) covers my conclusions. I defy you to point at a single conclusion that I made that could possibly change.
    – Brythan
    Commented Nov 8, 2018 at 17:26

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .