Two days ago Gallup released a poll of Trump approval rating that included polls in every state. Before accepting results, one wants to look for potential bias, but his overall rating was at 38% approve which is pretty close to what 538 has in their tracker, so that checks out. Looking at the state results, they also seem to match up pretty well to expectations. Other than Alaska, the dark green states on the map were all states that Trump won by at least 20 points and Alaska he still won by 15. With one exception (you’ve probably spotted it by now), the lightest colored states were ones won by Clinton. And while there are a few Clinton states that are in the middle category (Nevada, New Hampshire, Maine), they were all states where she got very close margins.
Three colors doesn’t allow for very much distinction, though, so I’m going to make a 7 color map using Trump net approval as follows:
Dark Blue: -15 or less
Medium Blue: -14 to -10
Light Blue: -9 to -5
Tan: -4 to 0
Light Red: 1 to 4
Medium Red: 5-9
Dark Red: 10 or more
Note that this is somewhat biased toward Trump, mainly because I feel he’d be a favorite where his rating is level. Here is that map:
The first takeaway is that the Democratic candidate in 2020 could literally win the electoral college by only winning states in which Trump has a net approval rating of -15 or worse. Next, time to finally mention that big outlier that you’ve probably been staring at this whole time: Texas. It could very well just be a bad poll on Texas. When you do 51 separate polls, you’d expect 2 or 3 polls to give results whose 95% confidence intervals don’t cover the true value. Then again, the decisions that Trump and the GOP have made with regards to immigration affect Texas more than most states. If 2018 is the year that Hispanics start showing up at the same rate as white voters, it could end Ted Cruz’ time in the senate. Even if Cruz holds on, there could be a surprising number of flips in Texas House districts.
This data also looks good for Donnely and McCaskill. Indiana and Missouri were interesting states in 2016 in that there were a lot of Trump-Kander and Trump-Bayh voters. In most states, the GOP senate candidate outperformed Trump. It looks not so good for Heitkamp, though. The Manchin-West Virginia dynamic is a unique one, so it’s not necessarily bad for him that Trump is +26 in WV. Anyway, this is just another way of looking at the data and could be insightful in seeing which areas are vulnerable in 2018 and possibly 2020.