You are viewing:


Information released online from January 20, 2009 to January 20, 2017.
Note: Content in this archive site is not updated, and links may not function. External links to other Internet sites should not be construed as an endorsement of the views contained therein.

printable banner

U.S. Department of State - Great Seal

U.S. Department of State

Diplomacy in Action

U.S. Presidential Election: The Latest Polling

Kyle Kondik, Director of Communications, University of Virginia Center for Politics; and Geoffrey Skelley, Media Relations Coordinator, University of Virginia Center for Politics
Philadelphia, PA
July 27, 2016

Date: 07/27/2016 Location: Philadelphia, PA Description: Kyle Kondik and Geoffrey Skelley, both of the University of Virginia's Center for Politics, brief journalists on the latest polling on the U.S. presidential race. - State Dept Image

1:00 P.M. EDT


MODERATOR: Okay, everyone. Hello, and welcome again to the FPC briefing space at the Democratic National Convention. Please take your seats at this time for our next event. I want to start with a few housekeeping events. First, I want to – housekeeping notes. First, I want to say that all of our briefings are streamed live on Transcripts will be posted there, and our video will be posted on All of our briefings are on the record, and the briefers’ views are their own and do not reflect U.S. Government policy. Lastly, for the Q&A portion of the event, please state your name and publication for the transcript, and be looking for our microphone, which could be coming from either your right or the left.

So without further ado, I’d like to introduce Kyle Kondik and Geoff Skelley, pollsters and political commentators at the UVA Center for Politics and editors of Sabato’s Crystal Ball – of the Sabato’s Crystal Ball newsletter. Without further ado, here are Kyle and Geoff.

MR SKELLEY: Hey, guys. I’m Geoff Skelley; that’s Kyle Kondik. I’m just going to start out just – we’re going to have some brief comments, and I think the first place to start because of where we are at the Democratic Convention is probably to point out – or to talk a little bit about the poll bounce, the convention bounce, however you want to call it, that Donald Trump appears to have gotten based on the few polls that we have between – or that came out following the Republican National Convention.

On average, going – well, I should say since 1988 it seems that the median poll bounce has been about 5 percentage points. And it looks like Trump, at least in the – we don’t have a lot of data, and there’s probably not going to be a lot of data, because it looks like a lot of polls are perhaps waiting until after the Democratic Convention to poll, which is annoying for us trying to figure out what’s going on. But nonetheless, it appears that Trump roughly had about a 3-point median bounce. But that actually – I think a good way to look at that is that that actually is sort of in line with recent cycles. Conventions have shown – have had really less of a bounce in the last few presidential cycles, really going back to 2004 and on, and there are other complications with that, in terms of the timing of the vice presidential announcement, because that can also produce a little bit of a bounce depending on the timing.

But generally speaking, Trump did get a bounce. In the RealClearPolitics average, he actually now has a slight lead, by about a point. And this is something that we very much anticipated just because this is how – this is how conventions usually work. After a convention, a party is usually better unified. And now the great question will be, well, will Hillary Clinton get a bounce, and we’ll just have to see on that point. You would expect one, based on historical precedent, but she may not. We just don’t know.

So anyway, I think that’s – I don’t know if Kyle has anything to add to that.

MR KONDIK: So I think it’s dangerous to sort of dramatically reevaluate what you think about a race in the middle of the conventions. And so you can look at the polls and say, oh, well, Trump’s winning now, there must have been a dramatic change. And maybe that’s true, maybe that’s the case, but I think that probably the better way to do it is to look at where the polls are at in August – maybe a week or two from now, after the – both Trump and Clinton presumably have gotten their polling bounce that you typically get out of a convention. And the polls historically become a lot more predictive after the conventions are over and about 100 days before the election. Well, we’re at about 100 days before the election. So let’s see where things are at in two weeks. Our Crystal Ball newsletter, again, which you can sign up for – it’s It’s free. It generally comes out every Thursday, although we had a special update this morning, talking about the polling bounce and our impressions of the Democratic National Convention.

But our current Electoral College ratings do still show Hillary Clinton as a favorite. We have 347 electoral votes safe, likely or leaning to her; 191 safe, likely or leaning to Trump. You need 270 to win, and so officially our ratings still show Clinton as a favorite. However, if the polls look today – or look in two or three weeks like they do today, we very well may have to reevaluate that. And so while we still see Clinton as a favorite, Trump certainly also has a chance to win, and so we’re just going to try to be monitoring this the best we can. But I would – I guess the one piece of advice, just to reiterate, is I would not look at the polls right now and say, oh, the race has permanently changed in a way positive to Trump. Maybe that’s true, but it’s also possible that Clinton will get her bump and then she’ll be back in the lead as she was a few weeks ago.

With that, happy to open it up to questions about the conventions or if you – also if you have any questions about down-ticket races in U.S. politics – Senate, House, governors – happy to talk about that too. So thanks.

MR SKELLEY: And I don’t know if someone has a mike. There we go.

QUESTION: Hi. Priscilla Imboden from Swiss National Radio. I would like to know in which demographics Hillary Clinton is doing the worst so far.

MR KONDIK: So there was a great chart that New York Times’ upshot site just put out basically showing that Clinton was outperforming Obama amongst – this was amongst white voters – outperforming Barack Obama amongst basically college-educated white men and white women, but underperforming Obama amongst non-college-educated white men particularly, and then white women by a little bit. And so there’s definitely this kind of education gap in this election, and so it’s possible that the results will look significantly different than recent years, because Trump may have a special kind of appeal to kind of – to non-college-educated people, particularly non-college-educated whites, but he may have a special lack of appeal to people with a college education.

So it’s possible that if Clinton has a winning coalition, it will be doing better than Obama did amongst the college-educated whites and also retaining Obama’s very strong majorities among non-white voters. In 2012, Obama won non-whites 80 percent to 20 percent, and the best polling that we have suggests to us that Clinton is poised to do better than Obama did amongst non-whites. So if Trump is to win, he’s going to need huge majorities amongst non-college-educated whites.

MR SKELLEY: Yeah, I’m not sure if I have anything to add to that. I think that probably about cut it – covered it.

MR KONDIK: One other point I want to make that might be kind of more a long-range thing is that I see change in American politics as being kind of incremental. I’m sure it’s this way in other countries too, that you sort of see trends over time and then they sort of – like the non-white vote kind of trending toward the Democrats, or working-class whites trending toward the Republicans. This is something we’ve seen in previous elections. It’s not necessarily new, but the particulars of this election may widen that gap out in sort of unusual ways, and those gains or those changes may not necessarily be static – or, I’m sorry, may not necessarily be permanent, in that maybe Donald Trump’s candidacy is something special, and then after his candidacy in 2020, assuming he’s not on the ballot, maybe we go back to more patterns that were in line with 2008 and 2012.

So again, if it – just to think – when we do get the election results in November, I would caution that this election may be kind of different from previous ones, not only because Trump’s a different kind of candidate, but also that we have the first female presidential nominee now, and we don’t necessarily know how the public is going to react to that. We – I think we have a – there’s a pretty good sense from gubernatorial and Senate elections where we’ve had women candidates that they don’t really pay any sort of electoral penalty for being a woman. You might think that they might, but that’s not really what’s been suggested in the past. But maybe Clinton does, and maybe that has some sort of effect on the election, and maybe part of Trump’s strength amongst non-college-educated whites could be a little bit of sexism. And that’s not to say that Hillary Clinton is a perfect candidate or anything like that, but I think that’s – that there’s that possibility.

QUESTION: Thank you. I wanted to know about the electorate that Hillary Clinton is trying to reach now after the first part of this election. It seems like she is moving away from the progressive base of the Democratic Party by choosing a more – a middle-way vice presidential candidate. How important is the – where is the potential bigger for the Democratic vote, in the more progressive part that she seems to be moving away from, or in-the-middle voters, the swing voters that she seems to be going for?

MR SKELLEY: Well, I think the Clinton campaign’s calculation – political calculation at least was that the Tim Kaine pick wasn’t going to completely ruin her with progressive voters. I mean, they already have their doubts about her, but I think they decided that it was better to pick somebody politically speaking who looked ready to govern, so she – she’s talked about that, I guess, as her like number one priority, whether that’s true or not. And Kaine does fit that bill.

And then at the same time, from a political standpoint, he can help in Virginia, a key swing state. And at the end of the day, I think they’ve decided that there may be more voters toward the middle or even some recalcitrant Republicans who don’t like Trump who might see Kaine as a friendly face and might be more willing to vote for her with him on the ticket, whereas if she had picked someone like Elizabeth Warren, maybe that sort of – that could have potentially alienated those voters, and I guess that they’ve decided that the left is going to come home to her by November because – are they going to vote for Trump? I mean, the odds are no.

But the one risk in that calculation, of course, is lower turnout among those groups. And I think that’s particularly notable among young voters, potentially, because if you look at the primary exit polls during the Democratic primary, Bernie Sanders won over 80 percent of 18 to 29-year-old voters. And that’s kind of unbelievable in terms of just one age group to be so starkly going in one direction in an intraparty contest is just very unusual. I mean, Barack Obama didn’t even do that well against Hillary Clinton in 2008, when he was definitely the candidate who young voters preferred on the Democratic side. And then you have this 74-year-old kind of ornery white guy who’s – (laughter) – who doesn’t seem to be – doesn’t – isn’t at all like Barack Obama but is actually doing better than Obama among young voters.

So I do think that there is a potential risk in the – just generally speaking, in the Clinton campaign’s strategy of maybe not pulling those voters in and maybe their turnout is down. And if their turnout is down, that is bad for Clinton, because that was the group – that is the most Democratic age group. If you look at the 2012 exit poll, Barack Obama won 60 percent of 18 to 29-year-olds, and that the best he did among any of the sort of age cohorts that we break them up in. So I – there are sort of potential risks and rewards here, and I can – you can see, I guess, both sides of why they picked Kaine but also the potential risks, if it doesn’t play out the way that they’re hoping it’s going to play out.

And I’m going to use this moment to add something in, because I think an important thing to remember about the timing of these conventions is that they were unusually early. We haven’t had two party conventions this early in a long time. Usually – at least traditionally, until the last couple of election cycles, you usually had one sometime in July and one sometime in August, and they were roughly a month apart. But basically since really ’08 they’ve been like right on the backs of each other every time, which is insane for all the reporters and those of us covering it, having them back-to-back weeks.

But it also makes it more difficult to really figure out just what the impact of the conventions is going to be. Because usually – I mean, for instance, last time it was at the end of August, beginning of September. And so we were really getting into the final stretch of the election. Well, this time around we’re going to have the entire month of August and the Olympics and all this other stuff going on. And that’s – there’s time for other events to happen and I think it just makes it even more difficult to really get a good read on what’s going on, just because there’s just so much time after the conventions.

So anyway, I don't know if Kyle has something to add.

MR KONDIK: Just one quick point on the youth voting. And Geoff may have actually mentioned this, but not only were young voters in the Democratic primary very pro-Bernie Sanders, young voters in general are the most unreliable voters, in terms of actually showing up to vote. And so that would seem to be a potentially troubling combination for Clinton. Not only do these voters really like Bernie Sanders, but they also are not reliable voters to begin with, generally speaking. So – and if youth turnout is down, that’s good for Trump. I would – I think that’s pretty clear at this point from what we know.

QUESTION: Thank you. You mentioned that working-class, older, white people may generally go for Mr. Trump. Do we know why that is the case? Because at least economically speaking, he does not represent them.

MR SKELLEY: Well, there have been a lot of explanations, I think, put forward for why that’s the case. And I think there – they all have elements of truth. I think it’s – you can’t really nail it down to one thing. I think you – there’s partially something going on here where you have a large group of voters – of Americans in general – who do feel a lot of economic dislocation. The economy – traditional sort of working-class jobs in this country have shifted from manufacturing to services. And that’s been going on for a long time; it’s just that it seems like it’s even – I think it’s one of those things where you have a situation where change has been happening for a long time, but maybe the full impact of that change hasn’t quite struck home yet.

And now with stagnant wages – to some degree, maybe the 1990s sort of covered up some holes that were beginning to develop as we shifted toward a service economy, because wages were still rising. But with wages becoming stagnant, I think a lot of people feel like the economy isn’t working for them, that there aren’t good opportunities to sort of move forward with their lives, that the jobs they have are not secure, that there’s just a lack of security. And you can talk about that economically, in terms of foreign policy, all that stuff. But I think economic insecurity is certainly a part of this. You can’t overlook that.

And you also can’t overlook race and racial – and this just – obviously Hillary Clinton is not – she’s not a minority. Well, she’s a white woman, so she’s not in Barack Obama’s circumstance necessarily, but there is something to be said for the fact that in the Democratic primary – in 2008, Hillary Clinton did better with voters who had higher levels of racial resentment. And this time around, because she had overwhelming support from non-white voters, she did a lot worse with voters with high levels of racial resentment. This is how you could have her going from winning the West Virginia primary by like 40 points in 2008 to losing it by like 20 this time around. And I think it just – she represents continuity with Barack Obama. And that factor I think has to play in, to some degree, with opposition to her among some voters.

So I think – I tend to look at – I don’t want to – I’m not trying to say that like racism is driving all of this, because it’s not. I think the economic insecurity is a huge part of it. I think you can’t really understate – like, if you’re traveling through, like, rural America right now, I mean, you hear about just heroin epidemic, opioid epidemic in some places. There’s just tons of people dying from overdosing on these things because they – I don’t know – there’s a lot of unhappiness out there. And I think you can’t – really can’t underestimate the impact of that.

MR. KONDIK: Also in Trump’s messaging – I mean, his slogan is “Make America Great Again.” There is this kind of, I think, sense of maybe losing ground, particularly amongst some whites, in that the country’s getting more diverse. They’re traditional, I guess, kind of – it gets at the racial question too. But there’s this sense maybe that certain groups of whites are sort of like losing the country and losing their values, and Trump is a candidate of kind of cultural conservatism, in terms of being very supportive – not supporting free trade, also being very much in favor of heavy restrictions on immigration, building a wall along the Mexican border, that sort of thing. And his – if he wins the White House, it will be almost exclusively based on white voter support. I mean, yes, there will be some non-white people who vote for Trump, but non-whites should overwhelmingly vote for Clinton in all likelihood. And so it would almost be a – kind of a statement by Trump’s white supporters that after the presidency of Barack Obama, we want to go back to something that we’re – a president that we’re more comfortable with, and race is a piece of that. So, again, as Geoff said, a lot of factors.

But another thing is that it seems like the electorate is pretty cynical right now. Trust in government is pretty low, large majorities of the public believe that the country is on the wrong track politically as opposed to the right track. Now, that number is usually negative. Usually, more people say this country’s on the wrong track than right track, but the number is particularly bad now.

And also, if we see this kind of as a – the Trump phenomenon as this kind of revolt by people with lower levels of education, it could almost be kind of a revolt of the heartland versus kind of cosmopolitan elites. And certainly, this convention, I would say, has mostly highlighted – there’s – so it’s highlighted the importance of the first woman candidate nominated by a party. I think it has highlighted injustices committed against non-white people, which are very – obviously very legitimate to highlight.

But I would also say that there may be white voters who look at the Democratic Party and say, “Well, that’s not my party anymore,” that that’s the non-white party, that’s the party of – basically of universities and of protesters and of kind of upper-class women. And so there might be this kind of cultural revolt going on against a Democratic Party that is more white collar and more cosmopolitan in its leadership than probably anytime before. That’s a classic critique of Democrats, but it’s probably – to some people probably truer now than it’s ever been.

QUESTION: This is Thomas Gorguissian with Tahrir, Egypt. The question – first question about what your forecast or expectation about the composition of the Senate and the House, and what are the factors that are going to shape that new composition, if there is even one?

And if you can take us in the coming three months – let’s say now we are in July and then there is August and there is September and then is October. From observer point of view, what are the main steps or changes or variations we have to follow in order to figure out what’s going on in November?

MR KONDIK: So again, historically, the conventions are a very important marker, and the polling we see after the conventions is more predictive than the polling that we saw prior to the conventions. I would also say that the first debate can be kind of important, although the first debate can also sort of skew things. A good example is in 2012, Mitt Romney took kind of a – took a lead, but that wasn’t really real. And what happens – and this may actually be happening now – is that we sometimes see in polling that when a party, like, has a convention or something good happens to them, that their voters are likelier to respond to polls. And so maybe what I think happened after the first debate in 2012 is that Democrats were disillusioned, their candidate had a crummy night, so they didn’t respond to polls as much and it might have created this kind of artificial lead for Romney.

Maybe that’s happening now. I mean, look, Hillary Clinton has had a rotten few weeks. She had James Comey, the FBI director, coming out and essentially accusing her of being negligent, even though she didn’t get indicted. And then we had the Republican convention, which was all about bashing Hillary Clinton – as conventions often are, and this convention is bashing Donald Trump – and this big party for the Republicans in Cleveland. And so if her numbers were ever going to get bad and if Democrats were ever going to get disillusioned about maybe even responding to polls, maybe this is the time and maybe that helps explain Trump’s lead – or not. Maybe there’s other things going on. But Democrats – probably a lot of them – will get excited by this convention and then maybe they start to be the ones who are more eager to talk to pollsters and who are more eager, more likely to say that they’re going to come out and vote.

At a certain point in the fall, and people who – we don’t actually – we study polls but we don’t actually conduct them – at some point, a lot of these polls are going to go from a registered model – registered voter model, which is you just ask – before you get – you ask the questions, you say, “Are you a registered voter or not?” If they say no, you end the call.

We’re going to transition into likely voters. So that’s a smaller universe of people. That’s people who are registered to vote, and based on the screen that the pollster uses, they determine that that person is actually likely to show up. Typically, the switch from registered to likely benefits Republicans, because their voters are basically older and whiter, so they’re more reliable voters. But there have been some polls this year that have shown Democrats doing better amongst likely voters because their voters may be more enthusiastic. So that’s another marker – that the numbers will change a little bit because the universe of people being surveyed will be different. Now, when that actually will happen? I would probably say after Labor Day, generally speaking, but most of the polls will go to that, and they’re not there yet. So that’s a – that’s not necessarily a campaign event, but it is – it’ll be a change in the numbers.

And then so we’ll have the first debate. It’s currently scheduled for September 26th, but it’s possible that there won’t be any debates. That’s just me – I don’t have any sort of inside information on that. It’s just that it’s possible that one of the two candidates – potentially Trump – will just say they don’t want to debate, or that they will – yeah, right – or that there will be fewer debates. Typically there are three presidential debates and one running mate/vice presidential debate.

The other thing is that the venue for the first debate was supposed to be Wright State University, which is in – outside of Dayton, Ohio. Wright State decided they couldn’t pay for it, they couldn’t raise the money, and so the first debate was moved from Wright State to Hofstra. One wonders if other institutions holding these debates – it costs a lot of money to host these things, millions of dollars and particularly for security – one wonders if other – some of these other places will decide they don’t want to have them, and also, will both campaigns want to participate?

One other factor – and if you were at our briefing last week, I talked about this a little bit – we have third-party candidates: Gary Johnson, Libertarian; Jill Stein, Green Party candidate. If you can get to 15 percent in a collection of national polls, you can participate in the debate. Johnson is not there at the moment, but maybe he will be, and maybe that will be a development that changes things, that we’ll actually have a credible third-party candidate.

So, again, I would look to what the polls say in August, what the polls say after we go to the likely voter model, and then whether we have debates, and if there’s some effect to the debates. And then you start looking at the campaign operations – who is spending much more money on campaign advertising and on ground game operations. As it stands now, the Clinton campaign has more money and is probably going to spend more on that stuff. Then it becomes how important is that; maybe it’s not important at all. Donald Trump got outspent during the primaries. It didn’t matter. So that’s another thing to watch. As we get closer to Election Day – I think the last debate is October 15th or 16th. So then there’s, like, a three-week period where the candidates are just doing campaign events and they’re activating their ground games. Early voting starts in some states in October, absentees. So a lot of people are going to vote before the – much – weeks before Election Day. So maybe they’ll cast their vote and then wish they had voted differently later on.

But the other overall point I would just say is that historically, at least recently, most voters aren’t changing their minds. They’ve already made up their minds. And then there’s a small group of true independents, and they’re the voters who pay the least amount of attention, so they may only be tuning into this race right now. So there’s a lot of fluctuation in the polls and – like, CNN had a poll that went from Clinton +7 to a week later, Trump +3 before and after the debate. I really doubt that that much of the electorate changed. I think it was just their poll maybe caught Clinton a little bit too high and now maybe they have Trump a little bit too high. But the electorate is generally more static than that, so I would not expect dramatic changes to occur after we see the state of the race in mid-August.

MR SKELLEY: Yeah. And getting back – getting to your point of sort of like the impact of presidential election on everything and how we go from there, the main thing we talk about is the coattails effect. That’s a traditional phenomenon in presidential elections in the United States – well, really any election to some degree, but particularly in a presidential election where you have the highest number of people voting. And if one party does particularly well at the top of the ticket, that usually benefits other members of their party down-ballot with their coattails. And so looking ahead at this point with the polls where they are, and we can’t know where things are for sure, but based on what our current Electoral College forecast is, that would be favorable for Democrats with them looking at trying to take back the Senate.

At the moment, the Senate is 54-46 in favor of Republicans, but Republicans are defending most of the seats that are up actually in this cycle. In fact, there’s really only one Democratic-held seat that is really in play at this point, and that’s in Nevada, which is Harry Reid’s seat – he’s retiring, the Senate minority leader. But otherwise, Democrats are – I mean, Democrats are basically on the offensive everywhere else. And so Republicans, because they did very well in 2010, are now having to defend seven seats actually in states that Barack Obama won in 2012.

So if Hillary Clinton can win those same states that Barack Obama won at the top of the ticket in 2012, that could potentially carry Democrats in those states at the Senate level. And also, if you go down a little further to the House level – because we’ve talked a lot about where Trump’s appeal and maybe where he doesn’t appeal as well, and if you look at where – if you’re a Democratic strategist, at this point, the way the country has shifted to a very urban-rural divide politically, the battle is being fought in the suburbs.

And so for campaigns looking at this presidential election, you think about suburban Philly, a little outside of where we are right now – where we’re staying, unfortunately a little bit too far out, dealing with the traffic every morning. You’re talking about a couple seats in the Philadelphia area that are pretty 50-50 at the presidential level, but maybe Donald Trump does worse here in the suburbs because he doesn’t seem to – it just seems to be an area where you have higher education levels, more – higher wealth. Those are areas that Trump hasn’t performed as well in. And so maybe that’s – those are places where Democrats could do better, but if there are any rural Democrats left, which there aren’t really – there are hardly any. There are a few, but not very many, compared – especially compared to just a few years ago, those are areas where a Democrat might be in danger. So Democrats need to win 30 net seats to take back the House, and so almost certainly, if they were able to get even close to that, it would have to come through winning a lot of suburban seats.

And so that’s sort of what we see as a potential impact of the presidential election on the House race. I would say at this point Republicans have a pretty solid majority and it’s going to take – it’s going to take a pretty solid Hillary Clinton victory, I think, to potentially set up an opportunity for Democrats to take back the House. So they’ve definitely got a challenge if they’re going to try to do that.

QUESTION: Hi, thank you for doing this. Marcelo Ninio from Folha de Sao Paulo, Brazil. I would like to ask: Is there any group, any demographic or state that – in which Hillary is doing better than Obama did?

MR KONDIK: So in our Electoral College ratings, we actually just moved Virginia, our home state, from leans Democratic to likely Democratic, so we upgraded Democrats’ chances there. And a number of reasons – one is that a running mate is typically worth a point or two or three in his or her home state in the election. Vice presidential candidates are not typically that impactful, but again, they can be worth a little bit in the home state. And that’s based on a lot of political science research, although political scientists sort of disagree about the impact of vice presidents, but our thinking is that Kaine is worth a couple of points in Virginia.

And we also don’t think that Trump is a very good fit for Virginia because Virginia has, particularly in northern Virginia, which is really the growing part of the state, a lot of white collar, highly educated wealthy people. Some of them may identify as Republicans, but they’re the kinds of Republicans who might not like Trump. Trump should do really well in the western part of the state outside of the three big urban areas – northern Virginia, greater Richmond, and Hampton Roads – that’s Virginia Beach, Norfolk. But the western part of the state doesn’t have that many votes, and so we actually are – think that Virginia is – could potentially be better for Clinton than for Obama.

The same is also true of Colorado. Colorado is also kind of a white collar state. It has a significant Hispanic population that is probably going to be fairly anti-Trump. Colorado is also a vote-by-mail state now, which generally increases turnout. And other states that have implemented vote-by-mail have become more Democratic over time – Washington and Oregon are two examples. So you put that all together and Colorado is another where Clinton might be able to do better than Obama. And then the flip side of that – the places where Clinton might underperform Obama – might be some of the whiter states of the Midwest that have a lot of those white men with lower levels of education, so maybe an Ohio, a Pennsylvania, Michigan, some of those states.

If Trump is going to win the White House, then probably the map that makes the most sense is that he would win all the states Romney won, so 206 electoral votes – the only state of those that was close last time was North Carolina – and then to win Ohio, Pennsylvania, and Florida, which are three states that Obama won twice. Those are three big states. Florida in particular is a very diverse state, but that seems to make the most sense in terms of a path to 270 for Trump. But again, to answer your question, Colorado, Virginia definitely seem like places that are trending Democratic in this particular election.

QUESTION: Hi, thanks for doing this. My name is Renzo Ruf. I’m here with some Swiss newspapers. I have a question regarding polls. The USC Dornsife poll asks people who would you vote for. They also ask people who do you think will win the election. And the answers are different. For example, this week – or no, today, it’s 49 say Clinton will win, but 47 say they will vote for Trump. Why does a Trump voter think that Clinton will win or the other way around? Which number is more important when I look at the poll?

MR SKELLEY: There’s actually some interesting research on this question, and I wish more pollsters asked it, because the research suggests that the who-you-think-will-win question actually might be more valuable than just the straight horserace question that you ask voters, because – and I think the reasoning behind that comes from the fact that when you’re asking someone who do you think will win, you’re not asking them necessarily for their opinion. You’re asking them for the opinion – for sort of what they know. In a way, you’re almost expanding the sample size that you are getting, because that person is taking in not only their opinion but all the opinions of all the people they know.

Now, of course, famously, you might have someone say like, how did that person, when everyone I know is voting against him or her, and it just turns out you just have friends who have the exact same political beliefs as you. But nonetheless, you still – I think most people still encounter enough people who have different opinions that they may be taking in all that information when they’re answering that question. So the research does suggest that that question actually is more accurate.

So at the – but I will say – cautionary – the – when that question was asked in the Brexit vote, people thought remain was going to win. Well, remain lost. So it’s still about who shows up, obviously. So it’s – I think that question has a lot of value, but I do caution – if Clinton fans are, like, trying to find solace in the recent movement of the numbers, you should probably not put it all on that question still showing Clinton ahead. But it’s certainly a question that has value, and I think the fact that it is like a different way of looking at it and a way of taking in more information than just the one respondent’s information, almost giving you a multiplying factor in terms of the sample size you’re getting, I think that is a very interesting, like – an interesting, and the research has shown, perhaps more accurate way of looking at things.

MR KONDIK: Has anyone here seen the movie he Grand Illusion? It was a 1940s movie. It was about French POWs in a German prisoner of war camp in – during World War I. And the movie’s really about social classes and kind of about how the German – or the French POWs get along with the German officers and the enlisted folks get along together. And the only reason I bring that up – I think it’s called the Grand Illusion; I saw it several years ago. The only reason I bring that up is that this may be an election where different – there’s not a lot of crossover amongst the Trump people and the Clinton people, in that the Clinton vote may very well be, again, the cosmopolitan, high society kinds of voters who live on the coasts, and that the Trump voters are a lot of rural voters who – and live in small towns and maybe just don’t have much interaction with the media elite.

There’s no question that the media elite in the United States, if you polled them and if you give them truth serum and asked them who they’d be voting for, they’d overwhelmingly be voting for Clinton. And I would say that generally speaking about the – a lot of the elite media types in the country and celebrities and that sort of group. And sometimes those people don’t connect well with people who don’t come from their same social circles. And so to the point about the who-do-you-think-will-win question, it may be that there’s not much interaction amongst these different groups, and so people are not getting – Clinton people probably don’t know a lot of Trump people and Trump people may not know a lot of Clinton people, and that may skew people’s understanding of the election.

The other interesting thing is – and this is something that Sean Trende of RealClearPolitics has talked about; he’s a very good election analyst here in the United States – is that typically in an election – and, like, if people are on Twitter or on social media or on television sets or whatever, you’ll have a Democratic strategist who’ll sort of lay out the Democratic case, and you’ll have Republican-leaning person – strategist or a media person – laying out the Republican case. In this election, most of the elite-level Republican commentators don’t like Trump at all. And so you don’t have this balance between one side calling the other on certain facts, and so they might be creating this kind of information bubble in which all of us pointy-headed elites think that Clinton’s going to win, and then she doesn’t, and we said oh my God, what happened.

And that actually is probably pretty reminiscent of the – of what happened with Brexit, too, because Brexit was certainly a – kind of an upstairs versus downstairs issue in that you had elites supportive and the countryside maybe not so supportive. And this might be kind of an upstairs/downstairs kind of election too, although let me also add we have every expectation that the vast majority of self-identified Democrats will vote for Clinton and the vast majority of self-identified Republicans will vote for Trump. And that’s borne out in polls. Now, there’s this bubble of people who call themselves independents, most of whom actually have a leaned party preference. And that’s where some of those crossover voters may reside, and so we – it’s harder to sort of track them.

But I do think there’s this cultural difference, and I mean, we – there’s a cultural divide between the parties that has existed for a long time, but it may be particularly stark in this one to the point where it might be hard for a lot of analysts to sort of really figure out what’s going on. So that’s what I kind of worry about in terms of analyzing this election.

QUESTION: Thank you very much. My last question is: What did the polls say about positive as opposed to negative messaging? We heard a lot of negative messaging last week at the RNC. It seems like the Democrats are trying to change it towards positive messaging. How does that work with voters?

MR SKELLEY: I think there’s a lot of – I mean, in terms of just overall messaging, I think there’s some – there’s mixed evidence on the impact of negative and positive ads – thinking about like TV ads, since that’s sort of the main way people have thought about messaging for a long time. But I will say that the reason why it doesn’t surprise me that you see a ton of negativity is not only just the two candidates, but – and the fact that they are so strongly disliked by large groups of people, including people who say they belong to the same party – you have this development in American politics called negative partisanship, where you actually find more people who may be voting for one party because they really, really don’t like the other one than just because they like the party they’re voting for.

And this was borne out in the 2012 election. The American National Election Study showed that a plurality of voters actually voted more because they disliked the other side than they liked their own, and we’re actually seeing that in polls so far. Plenty of polls have shown a majority of Hillary Clinton voters saying that they are voting more against Donald Trump than for Hillary Clinton and more – and a majority of Republican voters voting for – against Clinton rather than for Trump. And this – I’m sure that there are many think pieces we could write about what does this say about American democracy, but it’s a trend that has been going for some time, and to some degree it’s exacerbated by having two particularly dislikable individuals as the two major party nominees. So this was already happening even when you had, like, Barack Obama and Mitt Romney, and now you have two people who are viewed even more unfavorably. And that just – I think probably just makes it even worse.

So this is something we’ve been seeing for a time now, and so the stark negativity at the Republican convention, and even here – I mean, they’re trying to put on a good face for Hillary Clinton, and I think they know that they need to try to raise her favorables a bit among the public. But there’s going to be – there’s a lot of Trump-bashing going on too, and we’re probably going to hear more and more of it here in the last couple days, I would assume. And that negativity just goes hand in hand with this development and what’s going on with the electorate.

I don’t know if you have anything to add to that.

# # #