Today, he will talk about analyzing the art of public opinion. Is a nationally recognized republican and political operative. Me in welcoming him to the stage. Mr. Baselice thank you. It is a pleasure to be here. I have been cast with sharing information about polling. I want to look at what happens in 2016. That is where we are going to start. One of the things i wanted to share with you i wanted to get quoted on this. Asked by a reporter on the eve of the president ial primary about Hillary Clinton and barack , what i thought was going to happen in the upcoming race between obama and clinton . I said, i think this race is going to hinge on who turns out. Really . He is typing away. What else . I think the race is going to hinge on how many people turn out. That is great q i put out that headline does come backt to polling, the race does hinge on who turns out, and which polls had a better handle on that and which did not. With the Trump Campaign goes back june of 2016. Kellyanne conway, you familiar with that name . Everyone is now. I wrote that she was expected to have a senior strategic role in this campaign. She helped donald trump when he was considering running for president in 2012, but she got involved with a super pac it was not available until june the. I got a call from michael depreciate no about a week ebrugio, who said, president ial races are like childbirth. Or how painful it is, you have to go out and do another one no matter how painful it is, you have to go out and do another one. We thought, now maybe the campaign will listen to wait datang to the polling now that we have a Good Campaign manager. People we were able to manage electoral votes to going from 160154. The race was pretty tight for most of the way. We see in the middle of September Clinton had a slight lead. Eight weeks from this point, you would find the race really ends up with the popular vote about the same place. A twopoint lead for Hillary Clinton on the New York Times paul in september, poll in september. If you look at the thirdparty candidate, or the Green Party Candidate polling, you would get a different read. Look at this scale. Is 15 30 8 . Ere dont think for a second that polling and data cannot be presented a different way to give a certain look. Time, the National Polls drop this campaign cycle, there were different statewide polls. The things we had to recognize is that there werent as many statewide polling as many statewide polls, as the were National Polls. S. As there were National Poll clinton and mccain were only a few votes away. Trump was at 170. We managed to get 10 electoral votes. We are on the way. Highlighting minnesota for a reason, minnesota was the toughest category, even in october. Real politics will put something in the tossup category if they points. Ve recent data the only public poll in 2012 was texas. They needed that their data point to actually move texas from tossup to likely republican. No one was looking at minnesota. They were looking at moving minnesota from a tossup state to a democrat state. Look, alle national the way up to the eve of the election, here is a poll by abc news and washington post, that adequately lead, 4743. There are other numbers and questions worth looking at. Which candidate is considered to be the most honest . Trump was seen as more trustworthy all the way through the eve of the election than Hillary Clinton. The enthusiasm going up to the favor ofwas really in donald trump, and then Hillary Clinton caught up at the end. When we look at methodology, and this is something we dont talk it is sout, but interesting horserace numbers. Who is winning . Is that all we need to know . We need to know more about these polls. You can go a little deeper, you can look at the methodology. Here is the methodology for the of news poll, dating in july 2015. They used the same sheet over and over to explain their methods. Deeper, you can see the county where i am from. This tells us something very important at the bottom. Over 60 of the interviews were conducted by cell phone, and 35 by landline. Is that a balanced representation . Some polls on the campaign were 15 50 cell and 50 landline. Cell phones are how you find younger people. Who here has a landline . The one guy with the better headline than me. Harairline than me. We wont find you students unless we have a cell phone. This makes a difference in the outcome of the ballot. We have to know exactly what that is. Exactlyl tell you within regions or nationwide. This is an article around the seventh of november, 2016, talking about how close the races. You see somethings in the data, that 47 for clinton and 43 for trump. You can see that trump is winning the independents. Several polls already show this later in the campaign. It happened late in october. The senate started breaking for trump. Ere is Something Else look at the demographics. Has 38 democrats, 31 republican. When you take this into consideration, it ties up to a bout a fourpoint differential. About ae the country is threepoint democratic advantage over republicans. Show you what i think the national map looks like on a regional basis. How do i get three points . What was the spread in the race of the actual outcome that occurred last year between trump and clinton. Two points. The spread between obama and romney in 2012 . Four points. Together, thetwo most recent National Election shows the country to have a threepoint democratic advantage. We going to do about a poll that shows a 10 point advantage . I cant go out and do anything about it because of the way that poll is published. Adont think there should be poll that has more republicans than democrats nationwide, because that is not the way the nation is. Look at the early votes by party registration. Pointstwoabout 1. 5 points more democratic than it is republican in the turnout. Looking at the polling in a one point five pointtwo point democratic advantage. In colorado, i was in charge of that state. We got to within three points, but you know better. Those thatbout voted were registered democrats, 35 republicans. Only about half the state had party registration. Texas does not, so we have to look at other measures. If eve of the election you are being and wake up at three or four in the morning eastern time, even earlier here, and you see clinton has a 90 chance of winning, what do you believe . At, thisyou looking public polling data . How we absorbing it and using it . They are talking about it in philadelphia. Michigan was too close to call. A surge of Republican Voters among blacks and democrats. It could pennsylvania and play. Ohio is looking like a trump state at about this time. This project data is looking more tossup. Michigan and pennsylvania were much closer. They showed that clinton enjoys a onepoint lead over florida. That is what the states looked back. It does matter who turns out. This was a look at the map as it finally ended up. 361 12 votes for trump, 232 for clinton. Some people dont like the look of that map. Lets make those states lets base this on the size of the state. That looks more better, more balanced, in terms of the blue and the red. You can go up shades you can negotiate of red and blue. There are a lot of different ways you can look at data. A lot of things were said after the election. A fundamental rewriting of the map. Off,if clinton pulls this i and so many others missed a lot. Believed insaid, i data for 30 years of politics, tonight, data died. It didnt die. Ive never been wrong on anything in my life. Someone said, i didnt know how deep the divisions were. Polling cannot keep up with changing demographics. The polls have demographics in them. The polls are representative of the demographics who actually voted. Here is a poll that came pretty close. Some of us in the polling field have a problem, even though we know it saves money in the media. Look at the likely voters below that. Are we getting the right number of interviews in each state . At least they tell you what they did, and there is not 940 likely voters, and the margin of error is 3. 2 . 36 of the respondents in that survey are democrats, 28 republican. They were limited to democrats by my standards for a nation wide polls. They had it as a onepoint race. Democrats, nine out of 10, voting for Hillary Clinton. Nine out of 10 republicans were voting for donald trump. Nine out of 10, showing trump won independence at the very end. Looking at the states, arizona was pretty close. I had arizona and colorado. Thought we were going to win in arizona, so we used our resources elsewhere. 2. 9 is the clear political average. I was one of the ones recommended we go elsewhere. Think we have the signs and formula behind the decision to move out of colorado and go to michigan and pennsylvania, it was kind of a necessity. The map requires the Trump Campaign to go somewhere else. Necessity, we started looking to miss constant michigan, wisconsin, and pennsylvania. Advantage trump here. There were concerns, but there ended up being about a five point differential. A number of states were right on. I was a little bit off the mark. Trump defeated clinton by 10 points here. The state of maine was in the right direction. Most people thought that michigan was going to lean for clinton, and there was a slight differential in favor of trump. Margin off these error. New hampshire, new mexico we can see it getting any better. North carolina was going to go for trump. Ohio was one that missed the mark by the polling area and why . Pollingin about some being too democratic, when i think they should be three points more democratic than republican. If you have too Many Democrats, you get it inflated for Hillary Clinton. Polls here showed that clinton would win by two, but lost by about one. Virginia was pretty close. Because of all the predictive models out there, the polling was off. Look at this right here. The public polling is getting 4. 8 in the fourway telik test. Jill stein . Getting 1. 1 . , you are looking at 2. 4 of the vote. Where did it go . President trump. Clinton falls to points short of the public polling numbers. Vote. Was the hidden trump here we see polling right in the final days of michigan. In only one poll, the last that all the polling had trump with a twopoint lead. Lead,n has the 6. 5 point but there was nothing inside the second of november. This is somewhat dated and goes all the way back to the 26th the 26th of october . My complaint is that there was not enough public polling in the states. You have to keep in mind that polling are the Balance Sheets of politics. Accounting, they teach you the Balance Sheet measures. Liabilities. T that is what polling is. Assets against liabilities at any given point in time. Thatncome statement measures that income or net loss are the election returns. Look at pennsylvania. Hillary clinton lost pennsylvania. Here is what is interesting, and but we didnt see until about two and a half or three weeks ago in other states like this. The margin of victorys in the suburbs was pretty similar in terms of what obama had done in 2008 and 2012. The average margin of victory in the suburban philadelphia counties, you get about a hundred and 7000 votes for Hillary Clinton. In philadelphia philadelphia county, it is going democrats getsback to 1988 and she margin 65,000 votes of a of error. It is not quite what obama had done, but at that with the margin of victory she had enjoyed in the philadelphia counties she is basically where obama was. Philadelphia,in how they expected in terms of turnout by different age groups, africanamericans, anglos. Philadelphia is in the southeast corner. Bush won in 88, and clinton wins in his two races. See that bush beat kerry there by 350,000, and in 2012, romney defeated clinton and obama by 300,000. 700,000, more than the differential he was down in the southeastern part of the state. These bring up some points area pollsters pollsters do not want to admit this. Like 2012,i look typical president ial turnout we have seen more recently. Places up in the panhandle, tampa, like 2014 and was going to be good for republicans. Done higherlly turnout models with africanamerican voters and hispanic voters in texas to see with that would look like. For the first time, we had turnout models in one part of the state that may be needed to model,ied in a different which allows you to go more deeper into it. We saw a lot about these different programs, making these predictions, for one side of the other for one side or the other. On the first of october, it was 262 votes for clinton, and she is getting close. Ohio was still pretty tight by all the public polling, but it ends up really taking off for trump in the final weeks. We have to look at the data within the poll and look at what is in the poll. The polling which only took place in certain states, only 15 of those in the exit poll or Senior Citizens. A good fourth of them for more is going to be Senior Citizens in any president ial election, and 24 will not be under the age of 44. Agesdoes not reflect the on the ballot. They have the right africanamerican and latino percentages. Four points more democratic than republican was not far off either. Quick, these comments on for somel go questioning. Trump identified voters that wanted change in the country, but werent supporting him. Look at these ballot numbers, and look at the other data. Third party, looking at people that were voting. Map,d it to expand the look at other states. We had to capitalize on enthusiasm for trump. Regardless of what you may think about the two candidates, one of the things we noticed in our data was that people were not associating Hillary Clinton with change. She couldnt articulate what voters cannot articulate what Hillary Clinton would do for them. Trump was going to turn back traded, address immigration, cut taxes. Give a higher credit for working mothers or childcare. A lot of things he was talking about were issues we felt were resonating. I want to talk some more about what happened, and what is going on with the polling. Roe you should clap for that. That was good. [applause] i would like to answer some of the questions from the crowd. I typed these. Important swings in the white workingclass and africanamericans. How do you make this decision, when you construct your sample, to take account for that. Baselice we need to talk about with the turnout will be, based on recent elections. Example, i know there will be about 12 or 13 that will be africanamerican. Around 21 of the turnout should be latinohispanic. Texas is almost 40 hispanic. If you look at all populations, adult populations, you look at registered voters with a spanish surname. The turnout is about 21. Understand, what is the electorate . You have to look at other factors such as their enthusiasm , questions we ask to get people in on the very last days or weeks of the campaign. What is your likability, on the scale of 110 . Have you ever voted early . Whether you see whether it is africanamerican, anglos, different age groups, that can affect your model a little bit. That can make you wait the data differently. There was an africanamerican candidate in 2002 in texas running for campaign in the senate. Tony sanchez was running for governor against rick perry. There is an africanamerican on the democratic side and hispanic. The talk was that, this great Rainbow Coalition will turn out for the minority vote. They were talking 820 5 20 6 turnout of latinos and 50 of africanamericans. There is no Historical Data to back that up. That didnt mean that that was going to be the turnout. Rick perry was anglo. John cornyn was anglo. So more anglos would vote . No. You have to see that the interest is there. You started out talking about cell phones and land lines and the percentages of each. I saw three hands go up. One was my wife. The other was me. Two thirds of the three hands. How do you account for that . t take the entire survey walk the crowd through, how do you actually get people on the phone . Mr. Baselice when i started the , there were 336 phones down in houston, texas. We were together for a few years with gallup. Phyllis was the head of the phone bank. Concerned that at that time, when i joined the firm in the spring of 1989 that there were two refusals for every completed interview. How are we ever going to get a valid sample if two people refuse for every one that complete a survey . Refusals, soft refusals, no thanks. Hard refusals they have different four letter words. With the hangup. We have to code them, hard refusals. 15 to one. Including no answers . Will talk to 16 people. One out of the 16 will do the survey. 15 will refuse. Other people screen out. About half the dollars we make go to voicemail. It takes a lot more dials. It used to take about 1516 dials. Now, it takes about 100. Does any of this worth working . The National Polls had it pretty good. They were right on. Fortunately for polling, by peoplene, we are seeing that are old, young, different rural, suburban, answering surveys, at equal rates. We are getting representative samples. The first thing we do is look at polling to get a malefemale put in each part of the state. After that, we look at ages and partisanship and race at ethnicities. You can do interviews in the state and let it roll. People participated more readily. It is now harder to find young people. You call a landline on a senior citizen, there is a 94 chance youll get a senior citizen. Only a 34 chance i will get an 1844yearold on a landline. If we dont get that person, we go on to someone else in the household. Should we find you as a 42yearold and as someone else answers the phone and qualifies as a registered voter in . We talk to them in that county. We check our demographics as we go through the polling. We might say, stop calling republicans in this region of the state. We have enough and dont want to have to wait them down more than necessary. Do you find people answer fewer as a person that utilizes your services . I dont feel like the link of the service has been reduced. Does that mean that people will stay on the phone . We are doing m