c-span, your unfiltered view of politics. >> c-span now is a free mobile app featuring your unfiltered view of what is happening in washington and, live and on-demand keep up with the biggest events with live streams of floor proceedings from the u.s. congress, white house events, campaigns, and more from the world of politics all at your fingertips. stay current with the latest episodes of "washington journal" and find scheduling information for tv networks and radio plus a variety of compelling podcasts. c-span now is available in the apple store and google play. visit our website, c-span.org/c-spannow. your front row seat to washington, anytime, anywhere. host: joining us is david

sterrett, the principal research scientist for the associated press's center for public affairs research. i want to talk about what the vice president is seen. here is a headline, democratic pollsters have a warning about kamala harris' lead, urged bite numbers to rosie in 2016 and 2020, democrats holding their breath as harris surges against donald trump. your reaction? guest: think it is important to remember that polls are a snapshot in time. i think there has obviously been caution because in 2016 and 2020, the polls did overestimate the support for the democratic candidate, whether that be

hillary clinton or joe biden. it is important that people don't play -- pay to close attention to what the numbers are now and that is what the democratic pollsters are trying to warn people to not get too caught up in what number satan out a few months before the election as there is a lot of time before and things can change. host: less than 70 days until the november election rate is there a way to tell that the polls may be off? guest: i don't think there is ever a way to know that they are off. if there were, they would be making sure they worked off. i think hole -- polling is rapidly changing. pollsters are looking hard to try to produce data that is the

most reliable and accurate. it is important to realize that polls, what is happening now could be different 70 days from now and also people who have a rigorous pulling methods -- polling methods is not as easy to just say, are they accurate or not and it is something that it is important to not put one particular poll or get too focused on the numbers, because they could fluctuate and all four -- polls have emergence of errors. from a statistical tan -- standpoint they have margins of

errors. people working recognizing they are and estimate and not designed to be exactly within a percentage point. host: the associated rest has come up with a new methodology. explain what this is and why the change. guest: it is something that the university of chicago worked with the associated press and fox news to develop it following the 2016 election. the goal of the survey is to try to reach voters where they are and try to get the most reliable estimates that can be used on election night and following the election. combines a probability base where we reached out to people and ask them to complete a survey in the week coming to the election and including online nonprobability panels.

it is very large and we often interview 120,000 registered voters and try to understand not just their choice but what is driving choice in terms of the issues in demographic groups of those supporting the candidates. host: is it an exit poll? guest: i would say it is an alternative to the traditional exit poll in terms of the traditional is just interviewing people traditionally as they leave the polling place. as more and more people are voting early and voting absentee, the goal with a p vote cast is we can catch them where they are and make it easy for them to complete the interview either online or via the phone so we can capture the most representative segment of registered voters and highly reliable data in all 50 states. host: this is done in the days

leading up to the election, but what about right now when we are weeks out from the 2024 election #how do you conduct polling? guest: most of it what we do with our partners is by panel which is a panel where we recruit people and actually send interviewers out to people's houses to help recruit them and reach out on the phone. they join the panel and participate in a wide variety of surveys and each time we want to conduct a survey, we reach out to them and they can complete the survey online or via the phone and this is a national probability based panel where we can get a really reliable segment of the population in terms of education, demographics, gender, age, those types of things. host: i want to show viewers and

have you react to this video to say that this is going to be a close race. [video clip] >> i am sure you are seeing the polls and feel excellent about them. this is a dead heat in the emerging of the error rates and we are winning some battleground states and losing some. we have not done enough to make sure that we have set ourselves on the path to 270, and it is the work of the less than 80 days in front of us. i think when i worry, i worry about a lot, i worry fundamentally that people could see what's happening and feel like we have had this great moment in now we are ok. the only way we are ok if we

keep doing the work because there are still more people than not that don't know enough about who the vice president is and what she stands for. there are still more people than i would like who are looking at donald trump and still are giving him support at the highest level he has had even since 2016. so fundamentally, this is going to be as close of a race as we have ever had and we are a polarized country. host: what do you make of her analysis? guest: i think she has a lot of points that are very accurate. i think a lot of the polling is a close and more importantly, people are focused on seven swing states being michigan, wisconsin, pennsylvania arizona and nevada, georgia in north carolina. we look at the polling there, it is a close race and usually most have it within the margin of

error. as she noted and we have talked about, there is still a lot of time for polls to change and events that lead to changes in the polls, especially in swing states where there could be more fluctuations potentially. host: we are talking with david sterrett about polling. want your opinion. republicans call (202) 748-8001, democrats (202) 748-8001, all others (202) 748-8002. you can also text. you -- we will get to your thoughts. the associated press-norc did a poll.

63% of democrats are excited about the idea of a harris administration, up sharply from 40% who felt that way in march about a biden victory. when you see those numbers, what does that mean or turn out, possibly of the democratic voters? guest: i think those numbers are indicating there is a much more growing enthusiasm and energy behind the harris campaign and there was previously for the biden campaign. that is important for turnout. it is hard, people being excited in august doesn't necessarily mean they will show up on a tuesday in november but you would much rather have people excited and enthusiastic. in general, that does tend to lead to higher turnout. it is very difficult from boko -- polls to know who is going to

show up. people often times will tell you they will or won't turn out that doesn't always play out that way. when you are the enthusiasm numbers, it is definitely something that is encouraging for those in the democratic party who are hoping for the turnout. you see growing enthusiasm with some traditional democratic support groups, such as women, people of color, young adults, and those are groups that previously there were concerns about when they show up to support biden and in the polling data coming out, you are seeing enthusiasm in those groups. host: who is enthusiastic? what voting blocs are enthusiastic when you look at another trump administration? guest: trump has a very strong support, especially with groups such as men, people without a

college education, evangelicals. he does very well with those groups and has strong support amongst those groups. host: we will go to to west palm beach, florida, republican. caller: good morning. i am calling, you have the perfect guest on right now. it is about the polling and how it works. what has led me to vote for donald trump became president. the biggest thing was right near the end watching the numbers, i better get there and i usually vote on the day of voting which

will be november 5 this year. when when i saw those polls and i saw hillary clinton not going to wisconsin, i said i better get up and make sure i get there to vote for trump, which i did, and he won. absolutely, i went by what's in the polling numbers. just about every poll that was out there at that time. when i saw she was not going to wisconsin, i said, ok, i better go vote. host: david stern, what happened in 2016 polls? guest: i think there was a lot of evidence that the polling in 2016 -- has been a lot of diagnostics trying to understand it. i think what organizations and researchers have found is that in 20 a lot of the national polls were not overly off and could be relatively accurate, but the polling in some of the swing states, like the caller

mentioned in wisconsin and places like pennsylvania, in florida -- a lot of the state polling in 20 16 was off pretty significantly. think what most people found was that a lot of the state polls in 2016 -- the polls tended to under represent a group of people without college degrees, who ended up showing up at high levels in voting for president trump. it was that underestimation, probably, of those voters, by some of these polls in some of these swing states -- inaccurate and less reliable than they had been. host: what led to the underestimation of those voters? guest: at the time, what was going on -- i just think those voters were not responding to surveys at the same rates as people with higher education. and at that time, most political pollsters were not including education in their waiting adjustments -- weighting

adjustments, so they were not getting enough of these people with lower education and were not accounting for that. what most posters have done since 2016 is try to make efforts to increase and reach more people without college degrees, recognizing they are a very important part of the electorate. also, if they don't reach enough of them, making sure they do the proper adjustment. it can be a hard group of people to meet -- to reach for pollsters often. now, they weight education, which means they make sure they have the right percentage of people in each education category in their survey. host: in minnesota, terry, republican caller. caller: except for one thing. in 2020, they were off by the same amount as they were in 2016. so let's be honest. the polling is wrong, way

outside the margin of error, because you guys can't get a hold of people. you say the numbers -- that implies that the numbers have some factual basis. you guys are not coming up with factual basis. it is changing constantly. the polls often use methodology to synthesize to a result they want. right now, if you go with the last two elections -- i have to say it upsets me to hear you pretend it is the uneducated people that really fooled us in 2016. we corrected it. wait a minute. in 2020, it went right down the craft for them -- crapper for them too. guest: good question, terry.

you are correct to point out that a lot of polls in 2020 it also underestimate trump. i think a lot of people have looked at trying to understand that. it does seem like as you mentioned there were a lot of people who corrected for education in 2020, and that did not necessarily fix all the polls. it is a rapidly changing environment with pollsters. i think it is really trying to get the most representative group possible. and i think you are correct to point out in both those previous campaigns that a lot of polls have tended to under represent trump. that might be because of trump supporters being less likely to respond to polls. as you noted, most posters try to correct the demographic makeup of polls, so they make sure they have the right number of people. different ages, different races, things like that stuff. if people who support trump are

responding less, that could lead to an underestimation of trump. there are polls that have been off. obviously, there is a wide variety of polls that have a lot of different methods. usually polls are more accurate or less accurate. i think most posters -- i would say the goal is to get it as right as possible. the goal is to make the polls as accurate and reliable as possible. you are right that the past two presidential elections, posters have under predicted trump. people look and say, are they underrepresented him now? polling organizations are trying their best not to under represent him. are they? that is something we do not know at this point. we will see how it all shakes out after the election. host: let's go to john in

mississippi, independent. caller: how you doing? i really don't have a lot of faith in polls. i look at the last three elections and the democrats who supposedly are the party that is going to save democracy and manipulated the primary process in every election since obama -- i think we have probably the worst two candidates we have had in a long time. i'm voting democrat. i am in independent, i vote democrat because i promised my wife that i would not vote for anybody but joe biden or trump, and i said the only way i would vote democratic is if it was somebody other than biden. my faith in the polls is

nothing, so thank you very much. host: how do you get people to have faith in the polls? guest: i think the polling community is constantly trying to improve our methods, and recognizes that polls have not always been as accurate as we all hope. but i think it is also how you are interpreting polls and what you are using polls for. think a lot of the time polls have a margin of error. on a 50/50 rate, if it shows at 48-49, that makes a huge deal in an election and changes the result. it in terms of polling accuracy, polls are not designed to be picked up at the level of precision. i would say at the center we try to avoid doing horserace polls where we are asking who you are going to vote for this far out before an election. we think it is more beneficial to ask people their views about

the candidates and their positions on the candidates, to really try to understand the race so that the public can understand what things they are expecting out of the candidates and how they are viewing the candidates, so that you can understand things such as the last poll i was talking about. the polling showed there were many people who were not satisfied with an election between biden and trump, who were looking for someone else to look for. i think with polls the goal is not to get fixated on the top line number. obviously, we try hard to get the top line number as right as possible in the polling community. the richness of understanding stuff like what groups are supporting each candidate and what they view as the strengths and weaknesses of each candidate -- you can understand the dynamics of the race more than just what the number of people who say they are going to vote for him, which is something that

can change or fluctuate over time. host: the date of the poll, when it was conducted -- why do they matter? guest: i think the difference with the date is obviously there are people whose opinion changes over time. a lot of people, especially talking about now -- a lot of people -- not people who listen to this show and call -- many americans do not pay that close attention to politics during their regular lives. a lot of times, they don't start to form their final opinions and attitudes until he gets close to election day, the further out they are from election day, a lot of people are not paying attention to politics and there is time for things to change because of the debate performance or a new story or interview. people's opinions about the candidates and how they are going to vote is going to change. the further out you are from election day, and the more campaign strategists and stuff

will read that things can still change -- you put less into the top line numbers. host: speaking of interviews, the vice president promised to sit down for an interview before the end of the month, has scheduled that interview. she and her running mate, tim walz, will sit down with dana bash this afternoon. that will be at 9:00 p.m. eastern time. before the airing of that interview, the vice president will have a rally in savannah, georgia as well. the former president is campaigning in michigan and wisconsin this week. our campaign 2024 coverage can be found on demand, online, on c-span.org. in marietta, georgia, democratic caller. good morning. caller: i was just wondering, what is the methodology you are using, polling the people you

are going to poll? i'm in my mid 60's and have been active since the age of 18 and political stuff as well. just never, ever received any calls as far as polls, so that is what makes me doubt the accuracy of it. is there a different methodology that could possibly help pull more voters that would give you a more accurate counterpart? guest: thanks for calling. maybe we'll have to give me your number so we can make sure we get you in polls. pollsters are always looking to get as many people as they can. i would say with polling methodology, and has been changing rapidly in the last couple decades. traditionally, polls were done via telephone interviews.

the idea was that everyone who had a phone had an equal chance to be selected because they were randomly calling numbers over time. it seems like most people are not as excited to do a poll as you might be. we are seeing the response rates on phones have dropped dramatically. now, there is kind of two main methodologies pollsters are using. some are still calling people and trying to reach voters that way, in addition to testing. the other method that is becoming more and more common is using recruitment to a panel, where you recruit people into panels where they can do a whole host of different surveys over the course of months or years. i think both methodologies, whether you are interviewing people by phone or doing online panels -- both have strengths

and weaknesses, and i think you hit edit. both methods, the goal is to try to get a really representative sample of americans, so trying to get people in urban and rural and suburban areas, and different ages and races and genders. i think the biggest challenge for pollsters is more and more people are declining invitations to do surveys. especially when we get the phone calls, any people do not even know when pollsters are trying to call them because it might get blocked by a spam filter, by their carrier or by their cell phone. so trying to get people to do surveys is a growing challenge. we also try to outreach people via mail because at least they can see it and it gives them the option to call us or do the survey online. that is something we have done with the vote cast her -- caster. you can reach out to a nice representative sample that way and give people different options.

i think that is another thing we think is really important. we recognize some people want to do a survey on the phone. other people want to do it online. we could get a fully representative sample of americans. host: fairfax, california, republican. caller: i am a republican in name only because i was scared of george bush at the 9/11 time. i wanted to kind of say the same thing the last caller said. i wonder why you don't use the census workers, maybe, or just go out on the street. i have never been pulled. -- polled. i am older. a golden age woman, should we say. why not go out on the street and poll people? why is it so remote, on the phone? and quickly, i would like to say

-- about indigenous peoples rights, i have to squeeze it in, greta, and black people who should get money because they built the country with slave labor. i had to say it. how come you don't use the senses? -- census? guest: good question. by its nature, the census is designed to provide an accurate count of the people for funding in different congressional programs. it is an organization that -- they don't do anything related to politics. it would not ask you if you are a democrat or a republican. it is very cost prohibitive, usually. the census data is so widely used and needed that it can afford to do the in person interviewing, which is much more

expensive than doing interviewing other ways. also, it can be hard to get an accurate national sample with interviewers because obviously you would need to meet people in so many different communities. doing a survey of a thousand americans nationwide, interviewing people in different parts of the country -- to have interviewers in each of those areas knocking on peoples doors would be pretty expensive. there is also a decline in the number of people who answer their doors and agree to do those. i would also say that as far as your second question, there have been organizations that have done holding about issues -- reparations and that stuff. there was a partnership to do a survey about the views of evanson residents. i know there has been some other polling.

it was an issue in california, where there might be poles about what president views are like, the possibility of reparations. host: why should voters pay attention to national polls when we hear so much about swing states? and then it will come down to swing states deciding which candidate gets 270 electoral votes. guest: i think that is a really good question. if the focus is just on who is going to win the election, the national polls are not going to be as productive. if you are looking at the swing states, as we discussed. national polls have a lot of value in trying to understand the race. but it lets you see what groups of people are supporting each candidate and why. there is a reason to believe

that if african-americans in a national poll are showing greater support or less support for one candidate, it is likely that it might be the same thing in a swing state like georgia or north carolina. i think with national polls what you are really trying to see is the mood of different groups, especially key electoral groups. and if that is what is happening nationally, we assume that could happen in swing states. some emma graphics are more critical than others in swing states, given the demographic makeup like african-american voters are incredibly important in places like georgia and north carolina and pennsylvania, whereas hispanic voters might be incredibly critical in states like arizona and nevada. host: mentioning the swing state of georgia -- that is where the vice president was yesterday, and she will be there today as

well, holding a rally in savannah, georgia at 5:15 p.m. eastern time. we will have coverage of that here on c-span. if you are on the go or online, on demand at c-span.org -- and president trump will be in wisconsin at 7:00 p.m. eastern time and that will be live on c-span. that will be on the mobile app as well as online. jd vance, this afternoon, 12:30 eastern time, delivering remarks at a firefighters union annual convention in boston. our free video mobile app, or online at c-span.org. the vice president's running mate, tim walz, spoke to the firefighters union yesterday in

boston. we covered those remarks. you can find them along with our campaign 22 any for coverage online at c-span.org. let's hear from an independent. caller: david, i am surprised you are not more savvy than this. in the 2016 election, they were interviewing people regarding who they were going to vote for, and you had to have noticed that not many people were saying, i'm in line but i have not made up my mind yet. i recognized there was something extremely different in the coyness. people were not expressing, as they had before in the past -- i am voting for so-and-so. all of a sudden, it just turned

on us, and i thought that was odd. as i said, it just seemed like it was a very coy play. i think they had often ramped up on trump's side -- had already been programmed to not -- "i don't know. i don't know." i would just like you to respond to that if possible. being a pollster, you can really recognize the difference in the shift in people. guest: good question. i think what you are referencing is something a lot of people have tried to research. it is referred to as the shy trump folder in 2016. this idea that maybe posters were reaching people who said they did not know who they were going to vote for when they were planning to vote for trump. this is something a lot of polling organizations have gone back and tried to review mc. at this point, there was not a

lot of strong evidence that there were people not telling pollsters that they were voting for trump and that was the reason for some of the problems with the 2016 polling results. that does not mean maybe what you are alluding to -- that may people supporting trump have been less likely to take the surveys in the first place. those people who maybe were supportive of trump were less likely to do the survey, so there were less people voting for trump. crump supporters were being under representative. -- trump reporters are being underrepresented. this is the biggest challenges for pollsters. if one group of people who are more supportive of a certain candidate are less likely to do polls, refer to that as nonresponse bias. that can affect your poll because you are not having a representative sample. that is something pollsters are constantly trying to look at and

see if there is a difference, so responders of both parties are equally likely respond to the poll. if there are demographic factors or other things that are known about people that can allow us to adjust it to make sure we are getting a representative sample of americans. host: political observers of late have talked about the "turning the page" moment -- that there are certain elections where voters want to do that. can polls predict that this could be a turning the page election? guest: i think the question -- this gets to where it is important to understand maybe what people are prioritizing. i do think that often times you will see in polls that people will ask questions such as which candidate is more likely to bring about change, and trying to understand what features or parts of it candidate are driving the most support. i think people are looking for

potentially stability or a return to normalcy, whereas at other points in times, people are looking for change. looking at the most recent democratic presidents, i think obviously in 2008 obama ran much more on change and that seemed like something that polling picked up, because of the idea of change. i think in 2020, joe biden was focused on a return to normalcy was one of his themes. not whether he would make dramatic change, but whether he would bring things back to more stability. rather than just asking people which candidates they are supporting, trying to understand what they want out of a candidate -- someone who is going to be able to handle a crisis, who is disciplined or honest. something we focus on is trying to understand what factors people are most interested in when they are deciding who to support for president.

host: in pennsylvania, democratic caller. caller: polls don't take into consideration the electoral college. for example, hillary actually beat the people of trump and 2016 the popular vote. so the polls were probably right when it came to voting, if you will, in a normal weight. under the stupid electoral college system, the polls are skewed by that. that is what i wanted to say. that makes the polling inaccurate. thank you. host: do you agree? guest: this goes back to the point we were talking about earlier. the national polls are not designed to predict the electoral college and you could look at a national poll and see how that relates potentially after the election to the popular vote, as the caller was

referencing. if the goal is to try to understand who is going to try to get the electoral vote, that is where the state polling does become more important. often you do see a lot of news organizations and pollsters doing polls in the swing states. i would think you would continue to see more and more polling in those states. if you are just focused on the electoral college, obviously understanding those estates is more important than understanding the popular vote. that is what national polls are designed to pick up -- the national overall opinion. host: talk about the ap's work in the votecast methodology in swing states in the days leading up to the election. what will you do? guest: one of the big advantages of ap votecast is we have a peak people in all 50 states.

we pay a lot of attention to swing states. a lot of swing states in the week after the election -- we will interview between 2000 and 4000 people and get their opinions. we will interview people online or via the phone. we'll ask after the election a series of questions about how likely they are to vote and if they have already voted, who they are supporting, and try to ask questions about understanding the reasons for their vote, what issues are most important to them, so we can understand who voters are and what other top priorities, and what led those people to support either harris or trump. host: just a reminder, the electoral votes are allocated among the states based on the census. this is from archive.gov. every state is allocated a number of senators and representatives in its u.s. congressional delegation. you vote for senators in the u.s. senate.

the number of votes equal to the number of its congressional district. under the 23rd amendment of the constitution, the district of columbia was allocated three electors and treated like a state for purposes of the electoral college. each state decides how to appoint its electors. however, they must do so according to a law enacted before election day. that is a little education and a reminder. paul in pennsylvania. sorry, pat in seaport, new jersey. caller: i am interested in how accurate polls can be, going into the future, given that americans are self sorting. we used to have people of all walks of life, democrats, republicans, independent, all living in the same area, and now people are fleeing the north and moving south.

except on the most local level, how accurate can any poll be, when people with like-minded views are self sorting and moving to places where others share their view? thank you. guest: i think it is a good question and there is a lot of research to support the statement you were making about how americans are becoming more polarized, and there are starting to be more sorting amongst where people live and the communities they live. i think that is something that polling is constant and trying to adapt to and understand there is a need to then make sure that communities of both sides are being representative. whether it is rural or urban or suburban areas, that you are representing all those communities. good polls are working really hard to make sure they are accounting for factors whether it is demographic

characteristics, so all those points can be heard. as the country becomes more polarized, it could be more of a challenge. host: pat in seaport, new jersey, republican. caller: i was just on. host: you were just on. pat, i'm sorry. that was our last call for david stern, who is the principal research scientist for the associated press. >> hey look at our live campaign 2024 coverage coming up today. approximately 5:15, vice president kamala harris holds a really with supporters in savannah, georgia, a final stop of her tour of the peach state. later, former president donald trump meets with voters at a campaign town hall in la crosse,

wisconsin. you can watch these live on c-span now also come our free mobile video app, or online at c-span.org. >> all this week, we're showing several major rulings from the supreme court's most recent term, and we will talk to reporters about the backgrounds of each case and the impact of decisions. we are highlighting fisher versus united states, where the supreme court ruled to narrow the interpretation of federal statute used to charge january 6 defendants with obstruction, and it could impact special counsel jack smith's election to case against former president donald trump. the justices heard the argument in april. tune in for this and other rulings all this week at 9:00 p.m. eastern on c-span. you can find our supreme court coverage on our website, c-span.org/supremecourt. >> saturdays, american history

tv features historic convention speeches. watch notable remarks by presidential nominees another political figures. this saturday, texas governor george w. bush accepts the republican presidential nomination in 2000 and addresses leadership in america. >> our opportunities are too great, our life's too short to waste this moment. so tonight, we vowed to our nation we will seize this moment of american promise. we will use these good times to great goals. >> watch historic convention speeches saturdays at 7:00 p.m. eastern on american history tv on c-span2. >> next, phoenix mayor kate gallego leading a conversation on infrastructure development, workforce development, and the importance of labor unions at an event hosted by the democratic mayors association.