Gallup is the Warren Buffett of US Polling Firms. As old as time itself but still good at what it does. Gallup is most useful for presidential approval rating polls, as they update nearly daily a national presidential approval rating, and have since Harry Truman, which makes it useful to compare presidents historically, for history nerds like myself. But today, Gallup came out with a very different type of approval poll, one that polled Donald Trump’s approval in each individual state and published the results. Intrigued by it, I decided to do a quick breakdown of my thoughts on this treasure chest of data.
The Time Frame
Perhaps the most puzzling aspect of this poll is the time frame. The poll was conducted by sampling 81,155 US adults from January 20-June 30, 2017. That raises a few questions, but most notably, there’s a big chance that someone who viewed Trump favorably on January 21 didn’t approve of him on June 30. After all, FiveThirtyEight’s Trump approval rating tracker had him at 47.8% approval and 42.5% disapproval on January 26. Today he sits at 39.1% approval and 55.1% disapproval. He’s lost 8.7 approval points and gained 12.6 disapproval points since he entered office, and thus, it is absolutely feasible to believe that a number of people in this survey could have changed their minds since they were recorded. I know conducting a poll of this sample size takes a lot of effort and thus, time, but I just wish they hadn’t done the timeframe in this fashion. Therefore, take these results with a grain of salt, but given the way Trump’s numbers have trended (nint: down), the actual values today could be lower.
The Expected
Some of these numbers aren’t terribly surprising. Trump’s below 40% approval in every state that Hillary won except for three (Nevada, New Hampshire, and Maine). Trump’s above 50% in classically red states, but perhaps most interesting is that Trump’s approval in every state he won is lower than the percent of the vote he won except two (Alaska he’s at 51% after receiving 51% in 2016, but remember there’s a big 3rd party/Libertarian streak in Alaska. And Utah, which had the weird McMullin thing going on. Trump’s at 50/45 after getting 45 in 2016). That plays into a theory I have about how Trump’s biggest asset in the 2016 campaign was Hillary Clinton’s unpopularity, especially in red states and working class areas.
The Not So Expected
The single most stunning piece of data is Texas. What in the world? Trump did very poorly for a Republican candidate in Texas post-Southern realignment, getting 52% of the vote and winning by only 9 points (Mitt got 57% and won by 16, McCain got 55% and won by 13). This poll has Trump at 42/51!! Is that real? Who knows? Texas polling is few and far between, especially in a non-election year, which is frustrating for polling geeks like me who are really interested in what could happen if a legit Democratic candidate runs a statewide race in modern Texas, like Beto O’Rourke (translation: not Wendy Davis). UTexas Austin has done polling that shows Trump with these types of numbers and this only adds to the puzzle.
Another WTF piece of data is Mississippi. While Trump is strong in most of the South, he’s just a net +2 in in Mississippi. Yes, Mississippi!!!! I know Mississippi isn’t quite as red as other Deep South states due to its high black population (blackest state by percentage in America), but man is that a dreadful number for a GOP President.
Checking Up On Swing States
Another point of interest is Trump is not doing well at all in the Rust Belt states that fell hard for him in 2016. He’s a net -1 in Ohio and Indiana, -4 in Iowa, -9 in Pennsylvania and Wisconsin, and -10 in Michigan (note: he’s -18 in Minnesota, a classic blue state that came very close to falling to him). That also feeds my aforementioned theory about how the Rust Belt’s swing to Trump was more about a hatred for Hillary than a love for him (hopefully I’ll write a piece about this in the future).
As for warmer climate swing states, Trump is sucking in Virginia, which every VA Gov poll confirms, so no surprise there (if Northam doesn’t win that race… there are gonna be serious problems). He’s significantly underwater in North Carolina, Florida, and Arizona (more on that in a minute). Not as underwater in Nevada, but still underwater (net -6)
How About 2018?
Three of Trump’s four best states happen to be states with Democratic incumbents in 2018: West Virginia (Manchin), North Dakota (Heitkamp), and Montana (Tester). Good news for Dems is Manchin is individually very popular aside from the national brand and Heitkamp sorta is too (high approval ratings, won her election on the same night Obama got slaughtered in North Dakota). Meanwhile, Tester might escape 2018 solely because the GOP has no good candidates to run against him. As for the other Dem incumbents, the Rust Belt Sens (Brown, Stabenow, Klobuchar, Casey Jr., Baldwin) have to be very happy with these numbers, as does Bill Nelson in Florida. The most vulnerable Dem senators Claire McCaskill and Joe Donnelly have to like their state’s numbers, with Trump at just +3 in MO and -1 in Indiana. Essentially of the 5 Dems in very Trumpy states, the three in states with high Trump approval are the ones who are also best prepared to win in that climate, while the two in states with low Trump approval were the most vulnerable.
As for GOP incumbents, Jeff Flake and Dean Heller cannot like the numbers in their states of Arizona and Nevada, respectively. But those weren’t terribly unexpected given that Clinton won Nevada and Arizona swung hard against Trump relative to past GOP presidential candidates. The Texas one we already went over when it comes to Ted Cruz.
As for the House of Representatives, there are going to be a ton of competitive seats in California and New York, which REALLY dislike Trump. Now of course, we need a breakdown between the type of New York, with many of those seats being upstate/Staten Island/Long Island. Many other competitive seats will be in the Rust Belt, so those previous numbers are good, along with Florida and Texas and other swing states.
Good Use of Polling or Bad Use of Polling?
Playing a game that the FiveThirtyEight election podcast loves, I’ll say good use of polling. This poll is useful and many of its results are similar to what we’ve been seeing in more reputable state polls (like Virginia, Iowa, New Jersey for instance). As I said earlier, if anything, I’d say this poll overestimates Trump a tad given the downward trending direction of his numbers and the topline is a nationwide 40%, a bit higher than what Gallup had the daily approval at today (37%). But this is useful, and I’d love to see more organizations due a 50 state poll like this. Basically: the world is better with more polls and this is a real solid start.
Conclusion From the Data
Surprise: Trump isn’t very popular. Yes, he wasn’t on the campaign trail either, but having someone almost equally as unpopular to measure up against really helped. Midterm elections tend to be referendums on the President and for that reason, Democrats have to like this particular poll. And Trump’s vulnerability in key swing states highlights why Dems have to pick someone fresh and popular in 2020. But that’s a long time away. For now, so long, and thanks for reading! Feel free to comment!