# What was the tipping point state in the 2018 House elections? [closed]

If there were a Presidential election in 2018, and everyone voted for the same party for the Presidency as they did for the House, what state would provide the 270th electoral vote for the winning candidate?

• This doesn't make any sense. House elections are done at the district level while presidential elections happen at the state level. It's nonsensical to ask how much the the Democrats won the Arizona House vote by, since it's not a single election but many completely separate elections with completely different candidates Sep 1, 2020 at 15:16
• Can you stop talking like its the 1980s? Yes there are different candidates but most people vote for the party not the person in federal elections. And, to back this up, how close was the House result in AZ to the Senate? And what about Texas too? Sep 1, 2020 at 15:24
• I've edited your question because this is not the correct way to self-answer a question here. You should instead ask the question above, and then answer it below. Sep 1, 2020 at 20:15

Assuming each district cast the same number of votes in a presidential election with the current allocation of electoral votes, the tipping-point state would have been Wisconsin, with its 10 electoral votes taking the Democrats from 263 EVs to 273.

This is calculated by aggregating the votes at the state-level from the district-level results here, manually adding the results from D.C., sorting the states in order of Democrat margin of victory, and adding up electoral votes as we go down the list until we reach the state which would tip the party's candidate over 270 electoral votes.

Below is the data sorted as described above. Note also that if the states were to vote in this way, it would be a record - there has never been a presidential election where the states voted the exact same way as in the previous midterms, as explored in this question.

``````State,Electoral Votes,Republican,Democrat,Other,Dem Victory Margin (pp)
D.C.,3,9700,199124,19945,82.8
Massachusetts,11,497953,1943597,43531,58.2
Hawaii,4,87348,287921,7063,52.5
Vermont,3,70705,188547,13199,43.3
New York,29,1859054,3990483,92231,35.9
California,55,3973396,8010445,200681,33.1
Maryland,10,737906,1493047,55331,33.0
Rhode Island,4,129838,242575,867,30.2
Delaware,3,125384,227353,0,28.9
Washington,12,1048712,1888593,84646,27.8
Connecticut,7,520521,849341,9946,23.8
Illinois,20,1754449,2757540,27715,22.1
New Jersey,14,1198664,1856819,43260,21.2
New Mexico,5,264701,404026,24584,20.1
Oregon,7,702531,1061412,83703,19.4
Maine,4,250119,343635,29670,15.0
Virginia,13,1408701,1867061,37209,13.8
Minnesota,10,1125533,1420769,30695,11.5
New Hampshire,4,249685,310269,10517,10.6
Pennsylvania,20,2206260,2712665,10950,10.3
Michigan,16,1853459,2175003,126241,7.7
Wisconsin,10,1172964,1367492,21641,7.6
Iowa,6,612338,664676,39634,4.0
Arizona,11,1139552,1179193,22525,1.7
North Carolina,15,1846041,1771061,46224,-2.0
Texas,38,4135359,3852752,214597,-3.4
Georgia,16,1987204,1814468,683,-4.5
Montana,3,256661,233284,14476,-4.6
Ohio,18,2291333,2082684,32341,-4.7
Florida,29,3675417,3307228,38831,-5.2
Mississippi,6,471162,398770,68971,-7.7
Kansas,6,563190,464380,22752,-9.4
South Carolina,9,927494,758340,23458,-9.9
Indiana,11,1247978,1000104,8300,-11.0
Missouri,10,1330975,1027969,59469,-12.5
West Virginia,5,337146,234568,6277,-17.7
Alabama,9,975737,678687,5471,-17.9
Louisiana,8,835715,553184,71694,-19.3
Tennessee,11,1279655,846450,33719,-20.1
Kentucky,8,935304,612977,21517,-20.5
Utah,6,617307,374009,61190,-23.1
South Dakota,3,202695,121033,12237,-24.3
North Dakota,3,193568,114377,13587,-24.6
Oklahoma,7,730531,428452,19853,-25.6
Idaho,4,367993,207303,20428,-27.0
Arkansas,6,556339,312978,19981,-27.4
Wyoming,3,127963,59903,13379,-33.8
``````
• Not too bad. Only one point more Republican. That makes me happy to hear. Sep 1, 2020 at 15:43

@CDJB does a good job of compiling and analyzing the raw data in the model that the question implicitly assumes.

This said, the model is profoundly unsound.

The problem is turnout. If a Congressional District race is close, turnout will be high.

If a Congressional District race is not close (and only a pretty small minority of Congressional Districts are close, by design through gerrymandering), turnout is largely a function of what else is on the ballot. In a state that happens to have a Governor's race or U.S. Senate race or high profile ballot issue that election, turnout could be high even in uncompetitive Congressional Districts. On the other hand, if there aren't close, high profile races elsewhere on the ballot, and there are not close and highly salient ballot issues on the ballot, turnout can be quite low in a safe seat with an incumbent running.

Also, in general, turnout is consistently much lower in midterm elections, like 2014 and 2018, than in Presidential elections like 2016 and 2020. And since higher turnout tends to favor Republicans over Democrats (on average), midterm elections have a more "red" electorate than Presidential elections.

Some pundits, such as Cook, acknowledging this fact, instead, look to a variety of historic election outcomes that are more reliable, to assign a partisan index to a state or to a Congressional District, rather than looking at outcomes from one set of House of Representative outcomes for a single midterm election. And, Congressional District races could help in formulating such an index. But you'd need to adjust for factors like turnout and incumbency effects.

Uneven turnout makes any statewide estimate based upon votes cast for each party in each Congressional District a fundamentally unsound tool to measure one party or the other's political support at the state level.

Certainly, a single midterm election is a better estimate than a blank slate. But 2016 Presidential results, for example, are probably a better comparable if you want to predict election outcomes. Even voter registration data would be more useful.

Something that makes this intuitively clear at a macro-level is that fact that is is not uncommon in U.S. history for two of the electoral prizes at the federal level (the U.S. House of Representatives, the U.S. Senate, and the Presidency) to be controlled by one party, and one to be controlled by the other.