July 6, 2016
Yes. Donald Trump is either catching up or rebounding from bottoming out. As of the time of this writing, he’s improved a bit in the FiveThirtyEight forecast, moving past a 20% chance of winning on their polls-only measure, getting near 30% with polls-plus, and reaching 25% if the election were today.
If you’re a Trump partisan, a one-in-four chance may seem low, but it’s better than the 1 in 5 from last week. Though they look at this from three different angles, the most important input for FiveThirtyEight is poll results, both at the state and national level. Trump has gained on Hillary Clinton in two important respects.
When matched head-to-head, the gap is narrower. The Real Clear Politics average has Clinton leading by 4.6%, down from almost 7% a week to ten days ago. Not a massive change, but heading in the right direction for him. The average is a bit of a lagging indicator, as RCP combines a whole bunch of surveys that were released in the past couple of weeks.
Some of the currently included polls began surveying as early as June 19, so it doesn’t pick up every shift in the wind. Often, the easiest way to measure momentum is to see how the candidates are doing compared to the last poll done by the same organization. It works even better when those pollsters are participating more than once a month.
When Trump began falling behind again, this approach caught the move more quickly. He was consistently losing ground to his previous result. Now he’s gaining a bit. Rasmussen Reports is the most consistently pro-Trump poll. They’ve shown a GOP house effect in previous elections, and right or wrong are shading the same way now.
If he’s trailing with them, you know he’s looking bad overall. When Trump had his bad few weeks, he dropped from +5 to -1, then -4, -5, and one more -5. Now he’s +4. That’s not enough to put him ahead overall, but is an indication he’s recovered significantly. PPP has him a point closer (-4/-5) than a couple weeks ago, Gravis two points closer (-2/-4.)
Perhaps Trump bottomed out, maybe Hillary is reminding voters why they don’t like her. Either way, the trend is slightly to somewhat positive for him (though not unanimous–Reuters/Ipsos is showing Trump down the same 10 as in their previous measurement) Overall, the surveys that tend to show him doing better are indicating the gap is closing faster than those which favor her.
That gives us a wider variance between poll results than we had a couple weeks ago. In the newest batch, there’s a 16 point gap between Rasmussen (Trump +4) and ABC/Washington Post (Clinton +12). Normally, you can fit most all of the surveys into an 8 to 10 point range. When there’s a disparity of this size, it’s harder to confidently judge where we’re at. The pollsters are considering different electorates. Somebody is wrong, but we won’t know who until the voters actually show up in November.
The other trend going Trump’s way is the third party candidates, who are at a minimum sustaining their support and taking more of it from Clinton. The RCP average has them just over 11% combined, in line with the range of polls over the past several weeks that have shown them as low as 6% and as high as 16%.
That’s a pretty big spread, partly due to how they’re mentioned in the surveys, but as usual, the truth as well as the majority of results are somewhere in between. Most pollsters who do include the additional candidates also measure a two-person race. This makes it easier to see who they’re taking votes from. At first, it was damn close to 50/50. If Clinton was leading by 4 without them, she’d have a 3 to 5 point edge with them.
Now, we’re finding several polls that cut Clinton’s lead by a few points when you include the others. It’s not a massive split, more like 7% from her, 4 or 5% from him, but a couple points makes a big difference if the election winds up close, just ask Al Gore.
Looking at the RCP average by itself (which has Clinton ahead by 4.7%) doesn’t show the impact. You have to compare the same pollster with and without.The most Clinton-friendly poll is included in the four candidate chart. The most Trump-friendly isn’t.
Let’s assume Trump doesn’t change his approach very much (not exactly going out on a limb.) Let’s also assume his favorability rating also stays well south of 50% (also a safe guess.) There is a limited voting pool that is actually excited to vote for Trump, or will automatically select any Republican candidate. It’s a number not in excess of 40% of the electorate.
Barring an apocalypse, he’s not reaching 50% of the national popular vote, even against Hillary. But if the alternative candidates can combine for a decent percentage of the vote and take more support from Clinton than Trump, it becomes possible for Trump to win with a number in the low 40s. He’s not there yet. The RCP average has him in the mid-30s in a four-way matchup. Even accounting for which pollsters were included, that’s not good.
There is still a group of undecided voters who can’t pick anyone yet, even given four options. If Trump can grab a few of these and the Berners who are still staying away from Clinton remain with alternative candidates, he has a path. For any Sanders voters who Trump plans to attract, switching to Trump from Johnson is easier than from Clinton.
Hillary Clinton is still ahead in the polls. Donald Trump is still toxically unpopular. But things are trending his way a bit.