2016 Republicans, March 15, Poll Watch, Uncategorized

Last Minute Polling Shifts (GOP)

March 14, 2016

We’ve taken a look at the latest and greatest data on the Democratic side. Time for a look at the GOP. It’s the last stand for Rubio and Kasich, a chance for Trump to pull well ahead of Cruz. What are the numbers telling us?

In order from smallest to largest variability:


A mere 36 hours ago, Marco Rubio was just a huge underdog. Now he’s a Bernie Sanders in Michigan underdog. Of course, Bernie won, but it took the biggest primary upset in 32 years. That’s where Rubio is now.

Trump leads by 18.6% in the Real Clear Politics Average. Among polls taken entirely in the last 6 days, his worst result is 44%. If Trump gets anything in the 40s tomorrow, Rubio can’t catch him with two other candidates in the race.

We now have four pollsters who have measured the race twice in March. On the first round, Trump led by an average of 17 points. Now the average gap is 20. Not a huge difference, but the wrong direction.

The one bright spot for Rubio is it appears he’s pulled ahead of Cruz again. He’s 6 points up in the RCP average. His advantage is two points greater in the second round of surveys taken by the repeat pollsters than the first.

Kasich is hanging tough in the 8 to 10 percent range. He’s registered there in 17 of the last 18 surveys, the outlier being 11%.

North Carolina

There’s way less data than Florida, and Trump’s lead is smaller, so there’s a small chance Ted Cruz can pull off a surprise. The stakes are relatively low though. North Carolina is completely proportional for delegate allocation, without a number to hit to qualify.

While Marco Rubio is wondering where these rules were when he was just falling short of delegates in Vermont, Alabama, Tennessee, Idaho, etc., it makes the difference between Trump winning 44/34 and Cruz winning 40/38 mostly symbolic.

The first of those is what the data is showing right now. The two newest surveys have Trump at 44 and 48, Cruz at 33 and 28. Trump has won every state bordering North Carolina. He’s spent plenty of time in the state over the past week.

It’s not a bad Cruz state, and he virtually always over-performs on election day. Kasich and Rubio are stuck at about 20% combined, with Kasich now getting the larger slice of that smaller pie.

Older surveys regularly had Trump under 40%. At that point, Cruz was in range. It seemed if he could take most of Rubio’s support and most of the undecided voters, exactly what happened in Louisiana, Kentucky and Mississippi, he could catch up.

He started the final stretch closer in North Carolina than any of those states. Trump has moved forward in the last week though.

One final caveat. There are very few national pollsters covering this race. CBS/YouGov is consistently showing strong results for Cruz. They have him within 4 in Illinois for example. If we had a poll of theirs for North Carolina, we could see where his upper limit is.

Enough rambling. Trump is the favorite, Cruz will likely wind up in the 30s. Don’t be shocked if he wins. It won’t matter much either way in the delegate count.


One poll. Taken over a full week from 3/3 to 3/10. Nothing since. No national pollsters. Just Fort Hays University, which contributed in Kansas too. To make this a bit more interesting, they do not push respondents to pick a candidate.

Trump led Cruz 36/29, with Kasich and Rubio combining for 17% and undecided claiming 18%. With the lesser two already marginalized here, it’s probably not wise to assume they will lose even more support, but the undecided voters will likely flock to a favorite.

I’m treating this as a likely Cruz win, based on collateral information. In Kansas, Trump fell a little short of his poll number, while Cruz more than doubled his. That won’t happen here, but it’s not necessary to create a Cruz win.

Kansas was a closed caucus, this is an open primary. That’s definitely favorable to Trump. Don’t expect his final result to wind up lower than the (single) poll number. Cruz isn’t going to win by a huge margin like he did there.

But he won Iowa, almost won Arkansas, even with Rubio taking a big share, and is apparently running strong in Downstate Illinois. With all those surrounding points of strength, short of multiple polls showing Trump at or over 40%, the majority of evidence favors Cruz.

However, it doesn’t matter as much as it seems. This is a winner-take-all by congressional district state for the majority of delegates. Some go to the statewide winner, but more on the district level.

We know Trump has an advantage over Cruz in most large metropolitan areas. Plenty of Missourians live in the Kansas City or St. Louis area. This means even if Cruz wins, odds are very good Trump takes a district or three. Same in reverse if Trump takes the state.

Narrative wise, it’s potentially big. Missouri is Cruz’s best shot, and he isn’t going to want an 0-for-Tuesday result. On the delegate level, it’s not quite North Carolina, but the difference between a better outcome for one candidate or the other is closer to 20 delegates than 50.


This is very binary. Either John Kasich wins all 66 delegates, and continues on as the third wheel in a Trump-Cruz race, or Trump wins all 66 delegates and takes a commanding delegate lead.

No other outcomes. Rubio will finish an amazingly distant fourth. Cruz will finish a fairly to very distant third. All of the drama is for first. Not only is it completely winner-take-all, but we’re long past the point of caring who finishes third.

There were 4 polls taken between the middle of February and Michigan. Trump led all four, but not by much, anywhere between 3 and 6 points.

We have six surveys taken partially or entirely since Michigan. They are tied in two, Kasich leads in the other five. Quinnipiac is the one pollster to survey on both sides of the divide. They had Trump +6 from 3/2 to 3/7 and a tie from 3/8 to 3/13.

The two polls taken only from Friday forward have Kasich up 5 and 6 respectively. Things are moving in the right direction for him. It’s likely Quinnipiac (which didn’t break this out for public consumption) had him ahead for the latter part of their most recent survey.

So slight advantage overall for Kasich. Before you award him the delegates, keep a couple contrary points in mind. ARG just weighed in with a 44/38 Kasich advantage. It ties another poll for largest margin, and is his best overall number.

ARG loves John Kasich. They are always his best poll. Always. As such, he never matches their number. Their presence now allows us to determine his final result is something less than 44. If they have Trump within 6 points, it means it’s not completely over yet.

Neither candidate closes like Cruz. Kasich normally underperforms his final poll number by a couple points. Trump is usually very close to it when it’s a strong state for him, which Ohio is.

At the moment, Kasich is ahead by 3.7% in the RCP average. He’s probably going to win. The data indicates it, the momentum indicates it. His in-state approval rating indicates it. But it’s close.

The Democratic side is close too. It’s an open primary. Trump, Kasich, and Sanders are by far the three candidates who are most helped in open primaries. Most undecided Independent or potentially cross over Democrats or Republicans aren’t weighing all three, but might well find themselves deciding between two.

If we have any major surprise on either side, these ultimate swing voters are probably the culprit. Though we have two contests that are exactly or effectively two candidate choices, the open primary variable scrambles things.


Total wild card. We know Trump is around 35%. His RCP poll average is 35%. Each of the last 5 surveys, those taken after February 20, have him between 32% and 38%. Do not expect him to wind up under 30 or over 40.

So far, 35% is normally enough for a win. We’ve only seen a couple of states where that number was only good for second. In North Carolina and Missouri, Kasich and Rubio are combining for no more than 20%. In those places 35% won’t guarantee first place.

But in Illinois, Kasich is showing some strength. He’s averaging 18%, even with Rubio remaining in double digits in every survey. This one is a bit of a puzzle. Marco could retain his remaining 11% support, more than he wound up with in Michigan, more than he is showing in other March 15 states outside of Florida.

If he does, Trump is very likely to win. That would leave just over 50% for Cruz and Kasich to divide. Their appeal is to a completely different audience in Illinois. Ted is stronger Downstate, Kasich in the Chicago area.

If Rubio loses most of his stalwart, core support, at least some of it needs to gravitate to Cruz to give Ted a good chance of finishing ahead of The Donald overall. While Cruz did wind up at 34% in the CBS/YouGov poll, that’s a strong pollster for him normally, and he isn’t over 25% with anyone else.

If that support goes mostly to Kasich, it’s probably still not enough to get him a win. However, most delegates in Illinois are assigned winner-take-all by congressional district, just like Missouri. It would help Kasich get delegates in his best districts, which match Rubio’s.

We can expect Cruz will do a good job getting his voters out. If you give him the normal adjustment above his polling average, that would put him in the very high 20s, possibly very low 30s.

Depending on what happens with the Rubio vote, you could have a clean sweep for Trump (the state, most of the district delegates) if they stay with Marco, A win or at least plenty of delegates for Cruz if they go to him, or, if they go to Kasich, a situation where Trump wins the state, but Cruz and Kasich actually get more delegates.

There really isn’t enough polling to indicate which of those scenarios are most likely, though Trump comes out at least sort of ok in most.









Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s