- [Narrator] This is a production of PBS Charlotte.

- This week on Unspun, the Politics of Polling.

You know, they say the numbers don't lie, but the truth is not all polls are created equal.

We'll talk with political polling experts about polling accuracy and polling bias, plus the top five, what we think we know about polling, but really don't.

And remembering the most admired Charlotte native ever, by all polls in the world, our country, and right here in North Carolina.

Unspun is next on PBS Charlotte.

In today's America, welcome to the Spin Game.

Believe me, I know, I'm Pat McCrory.

When I was governor and mayor, I played the spin game.

I was played by the spin game.

But aren't we all done being spun?

Let's take the spin out of the whirlwind here on Unspun.

Good evening, I'm Pat McCrory and welcome to Unspun, a show that tells you what politicians are thinking, but not saying.

Every politician running for office is trying to get inside your head, trying to figure out what it takes to get your vote.

And that's why polls have become so important to so many campaigns.

And I'm not just talking about the polls that show which candidate is ahead or behind.

Those are the so-called horse race polls that we all see in the headlines, the public polls that all the major networks and newspapers pay big money for with exclusive poll results that wind up being this week's top story.

But what happens when today's poll numbers turn out to be wrong on election day?

Well, that's when we start hearing about the margin of error.

The polls we don't see are the internal polls that the campaigns are doing on their own, asking leading questions about a lot of issues to see which one push your buttons and asking about other candidates too to see where their weaknesses are.

Many campaigns, in fact, will then leak those private polling results to reporters to make their candidate look better.

By definition, a poll makes it sound like you're just out there asking the people their honest opinions.

But remember, in many cases, polls are bought and paid for by the media or special interest groups or by the candidates themselves.

They help decide what questions are actually asked and who is asked.

That's why more and more of us don't trust those poll numbers.

Maybe someone should do a poll about that.

This week on Unspun, we talk with two political polling experts about polling accuracy and polling bias.

Tom Jensen is director of Public Policy Polling, which is based in Raleigh and does political polling nationwide.

And Brock McClary is VP of polling of Signal, a Washington DC firm that also does polls for the Carolina Journal.

Thanks to both of you for joining us.

First of all, Tom and Brock.

Tom, your polling is known from a more liberal standpoint.

Brock, your polling is known more for a conservative viewpoint.

That's what the media says and that's what politicians think.

Is that true and why?

- Well, I'll jump in.

Brock and I are actually longstanding friends and during the 2016 election, we went to a baseball game together in York, Pennsylvania, and we each tried to convince each other that their side was doing better than we thought it was.

So I, as the Democrat was like, Brock, I think that in Pennsylvania and Ohio, you're doing better and he'd be like, oh, I don't think Republicans are doing quite that well, that sort of thing.

So I think that a little bit puts the lie to necessarily just because we work for Democrats or work for Republicans, our polls being biased, we're all just trying to get it right and we really wouldn't be able to stay in business if we didn't get it right.

Now, the one thing where that does come into play is if we do a poll and Joe Biden's down by 10 points, that's probably not something that's gonna end up getting publicly released.

So that may have an impact sort of on what polls our companies do and do not let out into public, but at least when we're working for our clients, we're really just trying to get it right.

- Real quick, Tom, so why wouldn't that be released?

It's because they're on the payroll or you're their customer?

- Well, usually private polling companies like Brock's and mine, the clients own the polls.

So they get to decide if it's something that they wanna put out in public or not and usually you're not gonna put something out into public if it's a really bad sign for your campaign.

- Is that the same for you, Brock?

- Yeah, I mean, it's a client choice, but we have our bias toward accuracy and that's whether you're a Republican pollster, a democratic pollster, you're judged and rehired by how well you do your job and accuracy is a big part of that.

- So is it accurate that you're known as a conservative polling?

Do you embrace that or do you try to push that away and go no, our polls are just the same as the liberal polls?

- Yeah, I think the results are hopefully the same if they're both accurate.

We work for Republican candidates from those that are more centrist to those that are further on the right and some of the best thing that Tom and I get to do is work for policy organizations that are kind of down the middle.

And so that makes the work really interesting.

- Now, as a former candidate in both the senate races, governor's races and mayor's races, I do notice the writeup for the polling can be different from a liberal group versus a conservative group because you release these polls to the media when you like the polls for your client and therefore your writeup or your headlines, you want the AP Wire or the Raleigh News and Observer, whatever to take as much of your interpretation of the poll that benefits your client.

Was that a correct assumption that I had during my tenure in politics?

- Well, I'll give you an example just from what we're seeing in our polling right now is that, and most of the key battleground states, President Biden is down by two or three points.

So if the election was today, Donald Trump would probably win.

But it's also a situation where we see that the voters who are undecided almost exclusively voted for Joe Biden last time around.

They don't approve of the job he's doing right now 'cause they're unhappy with one thing or another, but they are people who landed with him last time.

So if we put out a poll in North Carolina and say Biden's down by three, we aren't gonna just headline it oh, Trump's gonna win North Carolina, it's over for Biden.

We will emphasize the fact that he has a lot of room to grow because the people on the fence voted for him last time probably are more likely to vote for him this time.

So I think you're right that we are gonna put forward some analysis that might give a more nuanced picture.

- Tom, how about, or Brock, how about your writeup in the Carolina Journal?

Will you kind of put a little spin to help your candidate not fall too much, especially if it's negative?

- Well, not in that circumstance.

You're working for the John Locke Foundation, the Carolina Journal.

We're giving them the data as it is.

We give them an analysis of the data that is fair and balanced, the good and the bad depending upon what your perspective is.

And so certainly to the degree that polling can be used to drive narratives and campaigns, that is true, but when you're trying to give clients a good and accurate analysis of what the data says about a public opinion, that's job one.

- Okay, Brock, a couple of quick questions for both of you.

One is, what's the biggest mistake pollster's made or one that you've made which maybe could have even impacted the dynamics of an election?

- Yeah, I mean, I'll say that there's a... Understanding the accuracy of surveys is really difficult.

You can look at them on their face and not be able to discern a whole lot about them besides kind of the sample size and who you talk to.

But that isn't really tell you a whole lot about whether your survey is accurate.

That's determined by whether you are able to fill demographic quotas throughout your sample.

You're trying to... Tom and I are trying to achieve representativeness in our survey so we know if we're gonna do a poll in North Carolina, we have to have X number of independent women in the Charlotte media market and probably 100 other examples-- - So when didn't you do that?

Did you miss it at one time and go, what did we just do?

Were you surprised maybe because the sample was wrong?

- I mean, I think that there is always possibility in campaigns that turnout ends up looking different than is anticipated and that can drive variance and results.

And so that's the kind of circumstance where I think you want to be as accurate in predicting turnout is possible, but I'm hard pressed to figure out where there is a mistake made.

The most common mistake that was made pre-Trump prior to 2016 was not looking at the education of voters and meaning that polls tend to get voters that have college degrees and higher, 2016, and Trump drove out a lot of folks who were working class voters that weren't represented in the surveys.

- Tom, I saw something similar in the Obama election where there was a greater turnout of the African American vote than previous polls.

So some of the polls might have been wrong.

Have you made some mistakes where you've gone, maybe your sample was wrong?

- Well I think almost the entire polling industry really got the 2020 election wrong.

Not in terms of who ended up winning and losing, but certainly the margins.

And that was an interesting situation where you would generally expect polling to get more and more accurate closer to the election.

And in 2020, what happened was that polling got less and less accurate closer to the election.

If you look at the polls from March, April, that sort of thing, they were pretty close to what had ended up happening.

But if you look at the polls from September, October, they were way too good for Democrats.

And I think an interesting dynamic that happened that none of us as pollsters were really able to account for very well was that by the time you got to October seven months into the pandemic, Democrat leaning voters were much more likely to still be staying at home because they were worried about COVID and much more likely to thus be answering polls 'cause they were at home whereas Republican voters were a lot more likely to be going back about their lives, not be sitting around at home and answering polls.

I think that led to the polls being too good for Democrats.

I also think that's why we shouldn't just assume that the polls will underestimate Trump again in 2024 'cause I think the dynamics are really different this time around.

- US politicians often give soundbites.

I'm gonna each give you a 20 second soundbite as we end this on why polls should be trusted.

Because a lot of people just don't trust the polls, especially if you're behind, people go I don't trust those polls.

20 second soundbite on why you should trust polls.

- Well, I think there's good sound science for why you can trust polls.

I don't think that you should, however, look at them and think that the head-to-head horse race number is the only number in the survey that really is judged in terms of accuracy.

Pollsters like Tom and I where the ballot test, where the head-to-head is is probably not anywhere near the most important pieces of data that we're looking at.

We're trying to look at the data and see that it is predictive of outcomes.

- Gotcha, Tom, ten second soundbite.

What do you got, why should we trust polls?

- Polls really are doing the best that they can to keep up with the times.

We now contact people in a variety of different manners instead of just one way so we can reach people in the way that they're most comfortable with and we're also being more careful about waiting for different variables to make sure we have as representative a picture as possible.

I think 2024 will be a good year for polling.

- You met the 20 second criteria.

I really appreciate you both being on this show.

Thank you very much.

- Thank you.

- Thank you.

- Next, PBS Charlotte's Jeff Sonier takes Unspun on the street to hear your thoughts on the politics of polling.

- Here in Charlotte's busy plaza Midwood neighborhood, our unscientific poll on polling shows a lot of people are paying attention, but not many are being persuaded by those political phone polls.

What's your response normally when you get that phone call?

- I generally don't talk to them.

- These days, you get polls about the lollipop that you picked up in the grocery store.

So it's kind of sometimes a waste of time.

- Depending on what time of day it is and whether I'm working and they call me on my work number, I will tell them not to do it.

Otherwise I will listen to what they say and just answer their questions.

- [Jeff] Why do you think there's an accuracy problem sometimes with poll?

- I think it depends on their sample size.

I mean, it's like, which group of people did they ask?

I mean, you're gonna get different answers depending on age group or even your income.

It's gonna be different.

- Don't make it executive, don't make it all white collar workers, don't make it all blue collar workers.

Just get a piece of everybody.

Just get a piece of everybody and try to make sure that you're doing that.

That way you get an honest report.

- Honestly, I think I feel like it's rigged to lean towards specific candidates for different reasons.

- And during the election year, do you pay more attention to the polls, why not?

- Because it doesn't affect my who I'm gonna be.

- Absolutely not, no.

- I think it's basically because of the makeup of the people that they get the polls from.

They don't really know the issues and they don't really know the candidates that well.

- You can tell in people's voice whether they're telling you the truth most of the time or not.

So I don't know necessarily whether they're telling you the right things on polls or not.

- Yeah, kind of interesting that nearly every person we spoke with says they want more transparency in political polls.

More information about who's asking the questions and who's answering the questions, Pat.

- I'll tell you, that was such honest feedback.

So what do you think about the issue?

Email us your thoughts on how you see the politics of polling to unspun@wtv.org.

Alright, we're gonna have some fun tonight.

On our top five countdown, we've got the top five reasons why politicians, the press and the public just don't understand polls.

In fact, they're suspicious of polls.

Let's go to number five right now.

Number five, voters versus likely voters.

What in the heck am I talking about?

I'll tell you what I'm talking about.

Some polls are very cheap and they interview everybody, even those who aren't registered to vote.

Well, what difference do they make if they aren't even registered to vote, we know they won't vote.

Then some of the voters are registered to vote, but they never do vote.

So some of the polls you're seeing on Biden and Trump are registered voters, but the most important poll is likely voters.

Likely voters is the most important because those are the ones we the candidates care about most, are those who are actually going to vote.

That's who we wanna reach, that's who we need.

And that's the good polls have likely voters and anticipate the voters that actually do vote.

Number four, job approval versus likability.

This is really interesting because we see this with Trump and Biden a lot.

Some of the polls about Trump and Biden.

One will say, do you like Trump or Biden?

Well, a lot of people don't like President Trump and they like Biden, but that doesn't necessarily mean they'll vote for either one.

And the second survey that often comes out is job approval.

Well, job approval is different than likability.

They might hate the individual, but they like the job they're doing.

So when you look at polls, whether it be announced by one of the major networks or the cable network C is the poll likability or job approval.

Job approval might be more important than likability, but they often mixed and some people vote because they like the people.

Some people vote because they like the job they're doing.

Number three, the sample size.

The sample sizes are usually pretty accurate with polls.

In fact, what typically the biggest surprise in my lifetime on bad sample sizes was when Obama ran and more African Americans voted than were anticipated.

Therefore, the sample size were wronged and therefore the polls were wrong.

And when Trump ran the first time, the sample sizes were wrong regarding the blue collar worker.

They came out at much larger numbers than previous elections.

So the sample size is very, very important.

If the sample size is correct, the polls are gonna be correct.

Number two, manipulation and bias.

Let me warn you, some polls are what we call imbalance.

The polls will ask questions like, would you vote for Pat McCrory if you knew he robbed three banks?

Because what they wanna find out then I can run a commercial against Pat McCrory saying he robbed three banks because the poll show, if we knew he robbed three banks, we won't vote for him.

But that doesn't necessarily mean anything, especially if I didn't rob three banks and second, if the candidate doesn't have the money to run those types of ads.

Be careful of that manipulation of polls.

And last and most important is I hate to tell everyone that polls are usually accurate.

The only times they're not accurate is when they don't do the sample.

That's the facts of polls.

We don't trust them, we don't like them, but I gotta tell you, they're usually accurate with some exceptions.

That's the politics of polling.

Next up, PBS Charlottes Jeff Sonier joins me for our Unspun one-on-one segment.

Jeff, I gotta first tell you, we're talking about polling.

The only people I dislike more than reporters like you are pollsters.

So I just want to clear the air before we start.

- We can be friendly enemies here for a little while.

You know how this works, right?

We take questions from the headlines or from tonight's topic.

I ask 'em, you answer 'em.

That's why we call it one-on-one, you ready?

- You got it.

- All right, do polls determine the issues for a candidate or are the issues the issues no matter what?

- Polls determine what candidates talk about 'cause their campaign managers say, why are you talking about those other things when the people voting don't care?

And I've had to be corrected many, many, many times because often I think there are certain issues that have more of an impact, but the voters, you can't waste the time on those.

- We heard during the interview with the two pollsters about how they put together questions.

Do the candidates have much say in which questions are asked by their pollsters when they're trying to find out about issues or about their own popularity?

- If the candidates are involved in writing the questions, they're wasting their time.

They should be out either raising money or trying to get votes.

Let the pollsters do their job.

And most pollsters know the basic questions asked on the demographics, the basic information and the issues at hand.

Stay out of it, candidates.

- So we know when you're a candidate, you're looking at the polls.

But what about after you're elected?

When you were governor, were you aware of what the polls were saying?

Your own popularity, your approval ratings?

Are elected politicians still looking at the polls after the election?

- The closer you get to the reelection, the campaign manager that you've hired for your reelection keeps reminding you, McCrory, why are you spending time on this issue?

It's not working to your advantage.

They elected and I'd go, but I'm governing.

They go, it's time for governing to step aside, you have to get reelected.

And often I listen to them too late.

- Change the subject a little bit.

There's a pollster that was quoted in Reuters talking about new voters in North Carolina, saying 41% of the new voters in the last four years are either coming from Charlotte or from Raleigh.

How does that affect the next election, all those new city living voters in Raleigh and in Charlotte?

- Well, in the Charlotte immediate area, in Raleigh immediate area inside the city limits, it's gonna help the Democrats 'cause Charlotte's gone totally blue and a lot of the people are moving from New York, primarily in Ohio and Pennsylvania.

They're primarily Democrats.

But you've gotta look at the whole metro area.

The suburbs are going more Republicans, so sometimes it's gonna be a wash. - Also, seeing our first ads on television in this election season, Josh Stein running for governor.

Is it too early to be running ads now and the election's still five, six months away?

- Not when you have the money and I think Josh has the money.

The dilemma is when running ads now and you stop running the ads, they'll forget those previous ads.

So once you start ads, it's good to keep going.

I think it's too early at this point in time.

But he might have plenty of money and you gotta ask who's running the ads?

I think it's a super PAC running the ads, not the Josh Stein campaign and that makes a big difference.

- Our last question on one-on-one is about polls again.

I saw the results from a poll in 2021 showed you had a 75% approval rating for your four years as governor.

This was six months after you lost the election.

How do you have a 75% approval rating and lose an election?

- I did not know that, a little too late, but those are the unlikely voters I was talking about.

Those are the people who probably didn't vote and that's mean I failed to get to the people who liked the job I was doing.

I didn't get them out to vote and that's why I probably lost the campaign by 10,000 votes out of 4.6 million.

- Maybe another difference, as you mentioned earlier, between likability and job approval.

- Could have been, could have been.

I'm not sure which side I fell on.

- Well, like it or not, that's the end of this week's Unspun segment.

- Thanks Jeff, well, to wrap up our show, each week we've got a segment we're calling Backspin, where we give thanks to good public servants for the things they've done.

Since we're talking about polls tonight, well, every year Gallup Poll has a most admired poll, the top 10 most admired men and women in the world.

And since 1955, nobody's been in that top 10 more than North Carolina's own Billy Graham.

And this month, Dr. Graham was honored with a statue in the US Capitol.

And I was honored as governor to push for and sign the legislation that made this happen.

I first met Billy Graham in my first term as mayor in 1996.

As the first ever event in Charlotte's brand new uptown football stadium, it was a three day revival at the stadium and a historic homecoming for Dr. Graham, who was born right here in Charlotte.

On that first night, I rode to the stage, I rode out from underneath the football stadium onto the stage with Dr. Graham and South Carolina Governor David Beasley, to kick off the event in front of 70,000 people who were there to hear Dr. Graham preach.

And as we came outta the stadium, all these flashbulbs went off.

This was before the iPhone.

And as everyone cheered, me and the other politicians started waving to everybody.

And I leaned down to Dr. Graham and I said, welcome to your hometown.

Isn't this a great, great reception that you're getting?

And he turned back to me and he said, mayor, it's not about me.

I immediately pulled down my hand.

I learned a lesson.

As a newly elected mayor back then, that was a valuable lesson from Billy Graham that I still remember today.

He was admired throughout the world.

He was a hero, he was modest.

And right here in his hometown, we got to celebrate his life.

But Billy Graham always knew it was never about him.

Well, that's the reality as I see it and remember it.

Thanks for joining us.

I hope you'll come back next week as we explore politics, protestors, and the press.

That's on our next Unspun, where we'll tell you what politicians are thinking, but not saying, goodnight folks.

(light music) - [Narrator] A production of PBS Charlotte.