Daily Kos Elections extensively uses our calculations of the 2016 presidential election result by congressional district, like those shown above for Virginia. With ticket-splitting rates at historic lows, the presidential results can often be a significant indicator of how future downballot elections might turn out in a particular district, especially when there isn’t an incumbent running.
The presidential results by district are also an invaluable metric for gauging the advantage a map can give a particular party, whether it be from partisan gerrymandering or the state’s underlying political geography. One way we can measure this advantage is by analyzing the partisan leaning of the median district.
What do we mean by median district, and why is it important? We’ll explain it. The way we determine the median district is by calculating the presidential election margin in every district and ranking them from least to most Democratic. Or in other words, we’ll sort them from the district where Donald Trump won by the greatest amount to the one where Hillary Clinton prevailed by the biggest spread. The median district is the one in the middle, where half the districts are more Democratic and half are less.
Take a look at the chart below of Virginia’s presidential results by its 11 congressional districts to see what this looks like in action. Here we’ve ranked every district from Trump’s best to Clinton’s best. The highlighted row indicates the median district, which was Virginia’s 2nd. Trump carried this Virginia Beach district, which is represented by freshman GOP Rep. Scott Taylor, by 3.4 percent, meaning he won six seats by that margin or greater, while he lost the other five. Trump even managed to capture most districts in a state he lost.
We can then compare these district margins with the state as a whole to measure the median seat advantage. Since Clinton won Virginia itself by 5.4 percent, Trump’s 3.4-point edge in the median seat gives him an overall advantage of 8.8 percent.
That advantage is a useful number because it gives us an idea of how big of a margin the disfavored party would need to win by statewide to carry a majority of districts. Assuming all seats swung Democratic by the same amount, Clinton would need a statewide victory of nearly 9 points just to carry half the seats. Seats don’t usually all swing the same way in real life, but it’s still a helpful thought experiment.
This median seat advantage is particularly useful because it shows the built-in edge a party has. Democratic candidates running for House in 2012 won more votes than Republicans did nationally, but mainly thanks to gerrymandering, Republicans won a majority of seats. This was largely because Mitt Romney also won a majority of America’s congressional districts that year despite losing the national popular vote.
It just so happens that not only is Virginia’s 2nd the state’s median congressional district, but it’s also the country’s median, meaning Trump also carried a majority of districts in spite of the national popular vote. Trump had a national median seat advantage of 5.5 percent, which is a major reason why it will be so difficult for Democrats to win control of the House even if they win more votes.
You can think of that advantage as a 5.5 percent cushion Republicans have just from the maps. It might take a wave election similar to 2006 or 2008 for Democrats to regain the House thanks to these districts. If we want to ensure our electoral system rewards the party that wins the most votes with the most seats, reducing or eliminating the Republican median seat advantage would go a long way to doing so.