February 23, 2016
Polls are my friend so far. Trying to outsmart them leads to embarrassment, or at a minimum being wrong. Before the Democrats voted, poll evidence, scant as it was, gave Hillary a slight advantage.
Yet I picked Bernie to win by a couple points. Hillary was planning on leaving town soon after the vote. She’d pushed most of her resources and talking points towards South Carolina. Team Sanders was feeling pretty confident.
There’s always some interpretation. The final polling averages are a useful framework, not an exact view of what will occur. FiveThirtyEight has a great estimator, but it’s a range of outcomes. When the data is thin, they have limitations too.
I made the mistake of moving beyond what I could prove and into what I felt. Combined with wanting the more interesting narrative of a Bernie win, I goofed. So far, I’ve missed on the Iowa GOP winner and the Nevada Democrats.
Time to get back on the horse. With only 2 polls in 2016, none after South Carolina voted, we need to infer more than I’d prefer. Here are my assumptions:
The top 3 candidates will get at least 90% of the vote
Trump’s post-South Carolina bump should offset at least some of his caucus disadvantage
Cruz always outperforms his pre-vote position
Rubio is likely ahead of Cruz in Nevada
Put them in the blender and we get the following:
Donald Trump 39.8%
He was sitting around 40% in the polls. He reached 50% in a new Massachusetts poll, close to 55% on surveys taken on Sunday. No, Nevada is nothing like the Bay State. It shows two things, a surge for Trump after South Carolina, and a place where he’s way past the 35% ceiling.
If you adjust all that to Nevada, it would leave The Donald in the mid 40s. Until he proves otherwise, I’m charging him a caucus tax for inferior targeted ground game. Especially in a place where voters aren’t used to the process, it’s a thing.
Even if you take 5 to 7 points away, he’s still in the very high 30s. We know Trump has a motivated core of support that normally equals 70% of what his polls are showing. He doesn’t need that many late deciders to reach the upper 30s, and he’s tracking well at the moment.
A result in the 40s wouldn’t be a shock, but as you’ll see below, Marco and Ted have some advantages too. I’m expecting both to take advantage of the thinned out field, though Trump will get his share of transplanted voters too.
Marco Rubio 29.1%
Georgia is the place to look for a Rubio progress measurement. A post-South Carolina poll has him in second at 23%. The same pollster took a peek on 2/4 and found him at 18%. That was before he got Christied in the New Hampshire debate.
We know he took a hit for that. If the same poll was taken a week later, he would have wound up around 13% (+/-). Now he’s 10 up from that imagined floor. Nevada polls taken at a similar time have him in the 18-19% range. Add 10 and you get the above total.
He may pick up a few Jeb votes, but Bush didn’t have much Nevada support to give. That’s part of why he’s gone. A few Kasich fans are likely to opt for someone who seems more viable. That matches Sunday data from Massachusetts.
He will also grab a few voters from Cruz. There are a number of strong conservatives who are somewhat divided between the two. Ted gets more than his share of them, but when you hear about conservatives who desperately want to stop Trump, it’s these voters.
If you’re neutral between Ted and Marco, Rubio is looking like a better bet today. Expect a few of these voters to choose accordingly. If caucuses are total Kryptonite for Trump, Marco could win. If caucus turnout is as low as in 2008 and 2012, Cruz could push him to third.
This is the most likely landing spot though.
Ted Cruz 23.2%
So far, Ted has struggled to escape the Santorum Zone. In Iowa, Santorum (2012) +3. In New Hampshire, Santorum +3. South Carolina was a little better, Santorum +5. Unfortunately, the baseline in Nevada is 10%.
None of the polls are indicating a 13-15% result is in the cards. Back in 2012, Santorum didn’t have the time or money to build a strong caucus organization in the couple weeks after his Iowa win.
Cruz has outperformed his final polling average in each of the first three states. His organizational advantage is more important here than anywhere else, even Iowa, if you account for the newness of the caucus.
If you ding him a couple points for negative momentum from South Carolina, add a few points for game day turnout plus another couple/few extra caucus bonus, you wind up in the low-mid 20s.
John Kasich 3.9%
Limited organization. Temporary halt to momentum. Rubio seems more viable than last week. Candidate isn’t visiting Nevada. No debates or candidate forums between South Carolina and Nevada.
Caucus isn’t open to Independents. Change of registration was due 10 days ago. Nevada Republicans are fairly conservative. Trump, Cruz, and Rubio have dominated the narrative over the past couple of days.
There are a few voters who don’t find much to like in any of the alternatives. A couple Jeb supporters might object to the less tested Rubio. A few voters really like Kasich’s quirky charm. Hard to find much else in his favor.
Ben Carson 3.8%
I’m basically calling this a tie between Carson and Kasich. They weren’t separated by much in South Carolina either. This isn’t one of the Doctor’s better states. It’s about halfway between places like Iowa and South Carolina that were good to him, and places like New Hampshire where he struggles.
That would put him at around 5%, similar to his poll numbers. I’m betting he slightly underperforms them. Some of his voters have Trump as a backup, others Cruz. With Carson looking very limited right now, I’m expecting a few to tip the scales in that fight.