How The States Voted On Semi-Super Saturday

It’s been a real bad night for Rubio, who has just 15 percent of the vote in Kansas and 9 percent in Maine based on results reported to date. Cruz, in my view and that of prediction markets, is now the second-most likely candidate to win the GOP nomination after Trump. The only silver lining for Rubio is that Trump is having a really bad night also. At the very least, Trump’s results aren’t consistent with those of an “inevitable” nominee. It’s also possible that Trump’s numbers are in decline, but we’ll know more about that after Kentucky and Louisiana results are in and after Michigan and Mississippi vote on Tuesday. If so, that opens up some possibilities for other candidates down the road, including an outcome where Cruz and Trump each wind up with something like 40 percent of delegates, with the other 20 percent belonging to Rubio, Kasich or going into the convention unbound. One almost wonders whether Rubio and Kasich, either of whom probably require a contested convention to win the nomination, might find ways to start working together.

We’re starting to get the vote count in Maine (36 percent of precincts reporting), and Cruz is holding on to his lead. According to our friends over at the Decision Desk HQ, Cruz has 43 percent to Trump’s 33 percent, with Kasich at 15 percent and Rubio at just 9 percent. That would be a great result for Cruz and a bad one for Trump (who was most favored in Maine today) and for Rubio, who would fall short of the 10 percent threshold to receive any delegates.

I’m seeing a lot of people get excited about an American Research Group poll that shows Kasich (!) pulling ahead of Trump in Michigan, 33 percent to 31 percent. It’s a newsy result, but here’s why you should be a little skeptical. First, American Research Group is not a highly-rated pollster. In fact, they rank toward the bottom of our pollster ratings. Much to their credit, they’ll sometimes post results that are way different from the pack, so you can’t accuse them of herding. But their results are interesting more often than they’re accurate. Second, ARG has (for whatever reason) a large, pro-Kasich “house effect.” On average, their polls have been about 6 percentage points more favorable to Kasich than other polls conducted of the same states at about the same time. Third, there are a lot of other recent polls of Michigan and they don’t show anything like that result. True, not quite as recent as ARG’s poll, which was conducted yesterday and today. But pretty recent nonetheless. Our polling average in Michigan, which adjusts for house effects, accounts for pollster quality and tries to balance recency against other factors, has Kasich at 19.4 percent in Michigan. That’s good for second place, but it’s still well behind Trump’s 35.9 percent. Of course, it’s definitely possible that Trump’s numbers are declining based on the debate and other events of recent days. That would give Kasich, Cruz and Rubio a better chance to pick up wins down the road. But the better indication of that is the poor results Trump has gotten so far tonight, and not an ARG poll.

Want more up-to-date results in Maine? Check out Decision Desk HQ. They have Cruz leading Trump 45 percent to 36 percent, with 27 percent of the state reporting.

We obviously still have some time to go, but with 73 percent of precincts reporting in Kansas, turnout is about 46,000. In 2012, it was just under 30,000. In other words, turnout is way up, and yet Trump is going down to a big defeat. That’s why I don’t much buy into these arguments that Trump is driving big turnout. Many people in Kansas, it seems, came out to vote against him.

As an election watcher, days like this are among the best. You get results pretty much all day, and the results haven’t been as expected. Yes, I thought Cruz had a good shot in Kansas, but I didn’t expect him to win by this margin. We also are waiting on official results from Maine, where Cruz wasn’t thought to be competitive, and yet he is ahead in the early results.

The Republican race is quite challenging to model demographically, and also isn’t all that well-explained by ideology. So I expect that personality really might have something to do with it. Is it a coincidence that some of Trump’s worst performances so far are in “nice” states like Minnesota and Kansas, and that his best is in neurotic, loud Massachusetts? I’m not saying this is the most important factor, but it’s something worth thinking about.

Regarding the Midwestern wall, what is it about Trump that Kansans and others dislike? His extremely non-Midwestern personality and outsize ego? Or his unorthodox ideology?

There’s so much going on in the race — including the differences between open primaries, closed primaries and caucuses — that everything is a little overdetermined. But it’s not that Cruz beat Trump in Kansas that’s a shock. It’s the margin that’s shocking: 27 percentage points (!) as I type this. Plus, Cruz possibly winning Maine, which would be a surprise by any margin. There’s the risk of reading too much into all of this this, but there’s the risk of reading too LITTLE into it also. We’ll know more after we see the results in Kentucky and Louisiana.

Well, I’d say we expected Cruz to do well in Kansas, but the press he’s going to get off this win — plus what looks like a possible victory in Maine — can only help. I’d like to see more results before saying anything.

So Cruz has won Kansas, do we have any way to tell whether the GOP elites’ #NeverTrump campaign is working? Or maybe Cruz was always going to Kansas? After all, Rick Santorum carried Kansas in 2012, as did Mike Huckabee in 2008.

Our colleagues at ABC News have called Kansas for Cruz.

https://twitter.com/micahcohen/status/706248731881050112

Most news outlets are now calling Kansas for Cruz. Not a big surprise, as he holds a 24-percentage-point lead, and the 3rd congressional district is now reporting.

With Cruz set to win Kansas — the networks haven’t called it officially yet, but having a 24-percentage-point lead with half the vote counted is usually pretty safe — it will continue a pattern: Cruz tends to win geographically big states, like Alaska and Texas. Trump’s states? Well, they’re pretty small if we’re being honest. Nevada is the only Trump state in the top 20 in square mileage. Counting Kansas, Cruz has won states covering 1,142,430 square miles. Trump’s states cover just 422,054 square miles by comparison. Sad!

CANDIDATE	STATES WON	SQUARE MILEAGE
Cruz	5	1,142,430
Trump	10	422,054
Rubio	1	86,936

A: Good question Robert. I think the Associated Press is waiting for more votes from the 3rd congressional district. As soon as they get them, they’ll call it.

https://twitter.com/micahcohen/status/706243451600371713

Kentucky does seem more likely than Louisiana, yes. Especially given that Cruz is performing well in some areas Ron Paul won in Maine, which might suggest he could get some support among Rand Paul supporters in Kentucky. Louisiana, on the other hand … Trump might lose ground relative to the polls but that’s not the same thing as losing the state. I’m not quite sure what odds I’d give Cruz there, although they’re up from whatever they were this morning.

Nate, do you agree with Harry? Kentucky does seem more likely to reject Trump than Louisiana, right?

Well, Louisiana is a primary. It’s closed, which is bad for Trump, but it’s still a primary, which is good for Trump. That means the public polls are likely to be more reliable, and they have had Trump ahead. He also has done very well in the Deep South outside of Texas (if you consider Texas the Deep South). Still, Louisiana could be closer than the polls suggest. As for Kentucky, we’ve had such little polling there, and it is a caucus. I’m very interested to see the results there. I honestly don’t know what will happen.

Harry, we’re having a little bit of a lull in vote-counting in Kansas and Maine, so I’m going to ask you to engage in some irresponsible speculation. Given what we’ve seen form Kansas and Maine so far, do you think Trump’s also at risk of losing Kentucky or Louisiana? Or do those look pretty safe for him?

https://twitter.com/KerryPicket/status/706238364438175745

Another lesson from Kansas: caucuses and robopolls don’t mix. An automated poll from the Trafalgar Group earlier this week had Trump beating Cruz in Kansas by 6 percentage points. Right now, Cruz is winning by 25 points instead. Yes, a lot has changed in the last few days, but that’s a pretty bad miss.

We’re starting to get official results from Maine. With one of the state’s 22 precincts now reporting, Cruz is ahead 48 percent to Trump’s 35 percent, with Rubio (9 percent) and Kasich (7 percent) rounding out the field. We don’t know where these results are coming from, so caution is in order.

Is there any hope for Trump in Kansas? The state’s 3rd congressional district has yet to report any results so he’d have to make up a lot of ground there. But the 3rd district isn’t particularly Trumpian: It includes Kansas City, Kansas, along with wealthy, moderate, highly-educated suburbs like Overland Park and Shawnee. Their demographics aren’t particularly great for Cruz, but they aren’t good for Trump either. They look more like places where Rubio or Kasich might make up ground.

Assuming that Cruz goes on to win Kansas, where he leads by 23 percentage points with 18 percent of precincts reporting, then Republicans in Middle America will be on the verge of forming an anti-Trump wall. Minnesota (which went for Rubio), Iowa, Kansas, Oklahoma and Texas would all have voted against Trump. If either Missouri or Nebraska eventually does also, then a line of states stretching from the Canadian to the Mexican border will have rejected Trump for another candidate.

https://twitter.com/alexis_levinson/status/706228569450545152

Q: Are there any other notable historical instances of “the establishment” of a party rallying so quickly and early around a third-place performing candidate? — commenter Wes Cowan A: I’m not sure there is any historical precedent that close to what’s happening in the Republican race this year, Wes. I guess you could argue that the “establishment” rallied around Bill Clinton in 1992 despite Clinton having no wins and then losing many of the early contests. Then again, many members of the GOP “establishment” have NOT rallied around Rubio. He has more support than his competitors, but plenty Republican bigwigs have decided not to endorse anyone.

Q: Any winner-take-all thresholds to be aware of? — commenter Raymond N Desiree Eddings A: The only winner-take-all scenario possible today is in Maine IF one candidate gets over 50 percent statewide. Otherwise, all the delegate allocation is proportional.

Micah, I think Carson dropping out could definitely be a factor helping Cruz. I’ve been spending some time looking at how the Republican vote has gone congressional district by congressional district. A lot of the results are what you might expect — Rubio’s vote has a strong inverse correlation with Trump’s, for example. But there has also been a reasonably strong positive correlation between Cruz’s vote and Carson’s. Here are those results; they exclude Cruz’s home state of Texas, where he’s done far better than anywhere else and which would otherwise skew them pretty badly.

Carson’s vote could be going to Cruz

CANDIDATE	CORRELATION WITH CARSON
Cruz	+0.61
Trump	-0.17
Rubio	-0.20
Kasich	-0.67

Nate, with Cruz looking like he could have a good day, is it safe to say he may be getting a bump from former Ben Carson supporters? As Harry chronicled this week, Carson dropped out: https://abcnews.go.com/video/55783002

We now have 12 percent of precincts reporting in Kansas, and it looks like Cruz land. He leads 50 percent to Trump’s 25 percent and Rubio’s 14 percent. Not only that, but Cruz holds a significant lead in all three of the congressional districts currently reporting results. Unless something changes significantly, Cruz is going to win Kansas with ease.

Here’s where things stand, from The New York Times:

And to follow up on Harry’s point, the only primary today, Louisiana, is closed. To date, there’s been only one other closed primary, Oklahoma, and it went for Cruz, defying polling that had shown Trump with a lead there. There are many important closed primaries down the road, however, including Arizona, California, Florida, New York and Pennsylvania.

When watching the results from Kansas and Maine, keep in mind that both are closed caucuses. That means you have to be a registered member of the party to participate, while many Trump supporters come from outside the Republican Party. If the early results hold and Trump disappoints in both states, the closed nature of the caucuses could be one reason why. Also, organization matters a lot in caucuses. Cruz is generally regarded as the best organized candidate.

It’s very early, but it’s looking like a potentially good day for Ted Cruz. As Harry said, Cruz leads with 45 percent of the vote so far in the Kansas caucuses, although only 2 percent of precincts have reported and they’re in Central and Western Kansas, which should be relatively favorable for Cruz. A Kansas win wouldn’t be hugely surprising for Cruz given that polls there showed a toss-up and that prediction markets had Cruz favored. A win in Maine would be a bigger upset, however, and it looks as though Cruz could get one there also. The state has not officially reported results yet, but reports from caucus-goers on Twitter suggest that Cruz has won big in Waldo, Hancock and Kennebec counties. All three voted for Ron Paul in 2012, so perhaps Cruz is inheriting some of that vote. Cruz is also performing strongly in Google searches in Kentucky and other states. In the past, this has sometimes been a good indicator of a candidate who is drawing a lot of last-minute interest from voters. All of this is a bit fragmentary, of course. But that’s what we have to work with for now.

We’re starting to get official results for the Republicans in Kansas, and so far they look pretty good for Cruz. He leads the very early returns 45 percent to 29 percent over Trump. Keep in mind, it’s just 2 percent of the precincts. Perhaps more shocking is that Kasich is running about even with Rubio in the early results, which would be awful for Rubio.

Republicans go to the polls in four states today: Maine, Kansas, Kentucky and Louisiana. Democrats vote in Louisiana, Kansas and Nebraska. What’s going to happen? Well, we don’t really know; there have been very few polls released for any of the contests. But we can pretty much guarantee the results will be interesting. Republican Party elites, including the party’s last two nominees, have come out in full force against Donald Trump in recent days, but how will Republican voters respond? Perhaps Marco Rubio and Ted Cruz’s attacks on Trump in Thursday’s GOP debate will have some effect. Or, perhaps some Republican voters will start to rally behind John Kasich’s more unifying message. And on the Democratic side, can Hillary Clinton solidify her lead over Bernie Sanders? Nate Silver and Harry Enten will be following all the results and posting occasional updates here. So check back with us throughout the evening and night for the latest results and analysis.

Tweet us a question or comment

Leave a comment, and send us questions @FiveThirtyEight.