The reason for that is simple. House districts change shape every 10 years, often in unexpected and unnatural ways. They cross county lines and incorporate weird pockets of residents in a way that demands analysis at the precinct level to figure out how constituents of a representative voted in other contests.
Why does this matter? It matters in part because it's helpful to consider voting patterns in a broader context. For example: It's interesting to consider how voters felt about their options in the House races vs. at the top of the ticket. Did Donald Trump persuade a lot of Democrats to split their tickets, voting for him but then their local incumbent? Without district-level vote tallies, that's impossible to know.
Happily, the Daily Kos produces just such a tally. It's not yet complete, in part because vote totals haven't been certified in a number of states yet. But with 88 House districts complete — in Alaska, Arkansas, Connecticut, Delaware, Georgia, Hawaii, Idaho, Kentucky, Louisiana, Minnesota, North Dakota, New Hampshire, New Mexico, Nevada, Oklahoma, South Carolina, Utah, Virginia and Wyoming — we can see some interesting trends.
But before we do, we haven't yet marveled at the ridiculousness that is New Hampshire's 1st Congressional District. In 2006, during a Democratic wave election, Carol Shea-Porter won the seat. She held it in the Democratic-friendly 2008 election, too. In the 2010 Republican wave, she was ousted by Frank Guinta. In the 2012 reelection of President Obama, though, Shea-Porter took the seat back. But in 2014, Giunta retook it, as a result of the wave election that year. This year? Meet Carol Shea-Porter, freshman congresswoman from the great state of New Hampshire.
Hers was one of seven districts that voted for a presidential candidate of one party and a House candidate from the other. Meaning that NH-1 voted Trump, one of five places to back a Democrat for the House and the Republican businessman for president. Therefore, two districts (of the full 88) voted for Democratic presidential nominee Hillary Clinton and a Republican member of the House.
Let's look at a graph!
The first thing to notice is that most of the blue dots are under that diagonal line and most of the red dots are over. That's often the power of incumbency: It shows that winning House Democrats got more support than Clinton and winning House Republicans got more support than Trump. Which makes sense.
Note, too, that four of the seven districts that split their tickets were in Minnesota. (The seventh, not indicated because it's clustered right at the middle, is Nev-3.) In three cases, Daily Kos's data suggest that Trump won places that elected a Democrat to the House. In one case, Clinton won a district that elected a House Republican. In the case of Rep. Collin C. Peterson (D-Minn.), he won reelection by five points in a district that backed Trump by 30 points. That's pretty remarkable.
It will be interesting to watch what patterns emerge as more states are added to this list. Minnesota is one of the Midwest/Rust Belt states that helped make the difference for Trump. Will the same pattern of ticket-splitting hold in Michigan, Pennsylvania and Wisconsin? It seems likely.
At the state level, the picture is different. As we noted last month, no state split its vote between parties in the presidential and Senate contests — the first time this has happened in a century.
But compared with the results in each House district, that little factoid was relatively easy to figure out.