An inside look at how the PBS News/NPR/Marist Poll is conducted (2024)

The presidential campaign has kicked into high gear as President Biden and Donald Trump prepare to meet for their first 2024 debate on Thursday. The two men have been locked in a tight race for months and no other candidates reached the required polling threshold to be on the stage. Lisa Desjardins takes a closer look at how polls work with our partners at the Marist Poll.

Read the Full Transcript

Notice: Transcripts are machine and human generated and lightly edited for accuracy. They may contain errors.

• Amna Nawaz:

  The presidential campaign has kicked into high gear, as Joe Biden and Donald Trump prepare to meet for their first 2024 debate on Thursday. The two men have been locked in a tight race for months, and no other candidates reached the required polling threshold to be on that stage.

  Lisa Desjardins takes a closer look at how polls work with our partners at the Marist poll.

• Lisa Desjardins:

  Inside the offices of the Marist Institute for Public Opinion…

  Jacob Gresens, Survey Assistant, Marist Institute for Public Opinion: My name is Jacob. I'm a student calling from Marist College.

• Lisa Desjardins:

  … there's a buzz in the room. Over several hours and days, nearly two dozen student workers make hundreds of phone calls and send thousands of text messages.

  Overseeing it all…

  Lee Miringoff, Director, Marist Institute for Public Opinion: It's a neat process.

• Lisa Desjardins:

  … Lee Miringoff, the director of the institute, and Barbara Carvalho, the director of the Marist Poll.

• Barbara Carvalho, Director, Marist Poll:

  They are the ones that are talking one-on-one with Americans.

• Lisa Desjardins:

  Talking about issues or…

• Woman:

  Hello?

• Lisa Desjardins:

  … in many, many, many cases…

• Woman:

  Is there a better time I could call you back?

• Lisa Desjardins:

  … not talking at all.

• Woman:

  No worries. Have a great rest of your day.

• Lisa Desjardins:

  How would you describe the ratio of calls you get all the way through to dials you have to make to get that one good call?

• Jacob Gresens:

  Yes, on a normal night, it's about 100 to 1.

• Lisa Desjardins:

  One hundred calls to get one survey?

• Jacob Gresens:

  See Also

  News Wrap: Ukraine and Russia carry out deadly strikes on each other's territoryHow to watch Trump, Biden debate on CNN | CNN PoliticsPBS NewsHour; September 13, 2013 3:00pm-4:01pm PDTPBS NewsHour; September 17, 2013 3:00pm-4:01pm PDT

  Yes.

• Lisa Desjardins:

  But, on this night, a volunteer.

  We are talking to people in your community.

  As I hopped on the phones to experience the rejection firsthand.

  That's the end, I think. He hung up.

• Woman:

  Start by being engaged and enthusiastic.

• Lisa Desjardins:

  Me and my fellow callers all were trained for this.

• Woman:

  How about we try a few questions and see how it goes?

• Lisa Desjardins:

  Both online.

• Woman:

  The survey should take about 12 minutes.

• Lisa Desjardins:

  And in person, training required before any calls are made. Every step of the Marist Poll is carefully considered.

  Now is the time for that. I think that it's going to be hot.

  Starting about two weeks before with our team at "News Hour," where we discuss, sometimes disagree politely…

• Man:

  The lead-up to that, it's like…

• Lisa Desjardins:

  It's like — no, it's — I know. It's a nightmare.

  … and decide what issues we'd like to raise.

  How satisfied are you with the two candidates for president?

  Then pollsters at Marist with decades of experience finesse the language and order of the questions.

• Barbara Carvalho:

  Make sure that the question is understandable no matter who that person is that may be answering it.

• Lisa Desjardins:

  Once that wording is set…

  How satisfied are you with the two major party candidates for president?

  … all the callers, including me, ask the exact same questions the exact same way.

  Alexandra Newton, Survey Operations Manager, Marist Institute for Public Opinion: The reason why everyone is saying the same thing is to not introduce any bias into the survey. And so that…

• Lisa Desjardins:

  Every single word the same every single time.

• Alexandra Newton:

  Every single word the same every single time, no inflection on any word.

• Lisa Desjardins:

  Breaking up the repetition on the phones, a prize wheel spins in the back of the room, signaling a streak of completed calls.

  You may be getting the point here. In-person calls are not easy and are labor-intensive. To get a nationally representative and random survey of respondents, the poll relies on companies that aggregate all the telephone numbers in the country. As technology has evolved, so has the way Marist does this. Over time, landline calls grew to include cell phones. And now there are text messages and online responses too.

• Lee Miringoff:

  We just couldn't do telephone surveys because people aren't necessarily going to be responding that way, because we want to use the method that's going to give everybody a known or equal chance of getting into the survey.

  Stephanie Calvano, Director of Data Science and Technology, Marist Institute for Public Opinion: This is the data.

• Lisa Desjardins:

  As the surveys are under way, the results are being carefully monitored by Stephanie Calvano, the director of data science.

• Stephanie Calvano:

  I am running prelim data. I'm looking at all the frequencies.

  See Also

  PBS NewsHour; September 18, 2013 7:00pm-8:00pm EDT

• Lisa Desjardins:

  To the untrained eye, including mine, it might look like endless columns and rows of numbers, but…

• Stephanie Calvano:

  Each of these are codes that are the responses to questions.

• Lisa Desjardins:

  What do you see when you look at this data? Do you instantly translate it?

• Stephanie Calvano:

  I can, yes.

• Barbara Carvalho:

  What do you think the most interesting thing is from these results?

• Lisa Desjardins:

  The raw data also need context, what pollsters call weighting, mathematical adjustments to make sure the final numbers actually reflect the larger population. For instance, if one survey happens to get 60 percent of its responses from women, when we know the U.S. population is actually just over 50 percent women, Marist would increase the weight of answers from the men who responded.

• Lee Miringoff:

  The goal of having a representative cross-section of America in this case is where you want to be. If you want to think of it as a soup recipe, well, what are the ingredients and how much of each ingredient do we want in it, well, that we can find out by literally checking with the census numbers.

• Lisa Desjardins:

  As the methods and the science have grown more complicated, polling itself has grown more prominent.

• Man:

  The sitting president has moved into the lead.

• Lisa Desjardins:

  Becoming a staple of cable news campaign coverage of who's up and who's down.

• Man:

  Should Trump be worried by those numbers?

• Narrator:

  Biden's unprecedented inflation.

• Lisa Desjardins:

  Shaping the candidates' messages.

• Narrator:

  … convicted criminal who's only out for himself.

• Narrator:

  The CNN presidential debate.

• Lisa Desjardins:

  And even helping decide who qualifies to be on the presidential debate stage this week.

• Narrator:

  The most anticipated moment of this election.

• Lisa Desjardins:

  That matchup will feature just two candidates, Joe Biden and Donald Trump, who have consistently pulled at least 15 percent support in national polls.

  Polls, including yours, are used as qualifications for debates. What do you think about that?

• Lee Miringoff:

  That's bad. We think it's a horrible use of public polls.

• Lisa Desjardins:

  Why?

• Lee Miringoff:

  Why? Polls have a scientific basis to them, but that doesn't mean there's not a range in the numbers.

• Lisa Desjardins:

  A little polling glossary here. He means the margin of error, the range above and below the poll result, that represents where mathematically the true feelings of the entire country or larger group could be.

  Joe Biden, President of the United States: So far, the polls haven't been right once.

• Lisa Desjardins:

  Margin of error, as well as some high-profile differences between polls and election results…

  Donald Trump, Former President of the United States (R) and Current U.S. Presidential Candidate: The polls are all rigged.

• Lisa Desjardins:

  … have helped some candidates fuel public doubts in polling at large.

  Courtney Kennedy, Vice President of Methods and Innovation, Pew Research Center: Polls are definitely better for some things than others.

• Lisa Desjardins:

  Courtney Kennedy is the vice president of methods and innovation at the Pew Research Center. She says it all comes down to how you use polls.

• Courtney Kennedy:

  Can we use polling to predict the winner in a very competitive election? The answer is no. I love polling, but it's just not a precise enough tool to do that.

  Polling is absolutely up to that task of giving us a high-level read of how the public feels about a major issue of the day.

• Lisa Desjardins:

  And not all polls are created equal, which is confusing. So how should people figure out if a poll is, in a word, good?

• Courtney Kennedy:

  Track record and transparency. Is this a polling organization that you have heard of that has a track record of doing high-quality, nonpartisan polls or not? Polls that are willing to disclose more detail about how they do their work, they tend to be more accurate.

• Lisa Desjardins:

  Where does Marist fit in all that?

• Courtney Kennedy:

  Their track record is one of trying really hard to be nonpartisan and doing rigorous polling.

• Lisa Desjardins:

  Despite the science and the work to be more transparent, how much do you trust public opinion polls?

  Polls themselves have an approval problem. Trust us, we asked. During this PBS News/NPR/Marist poll, six in 10 Americans told us they have little to no trust in public opinion polls.

  Why do you think people should trust polls?

• Barbara Carvalho:

  I don't know that it's about trust and it's about faith. It should really be about science.

• Lisa Desjardins:

  Regardless of what they said or whether the trust in polling is irrevocably gone, all those respondents still answered the call.

  We're getting close to the end here.

  For the PBS "NewsHour," I'm Lisa Desjardins on the phones at Marist college in Poughkeepsie, New York.

• Amna Nawaz:

  And be sure to tune in to PBS on Thursday at 9:00 p.m. Eastern for our simulcast of the CNN presidential debate with analysis to follow.

Listen to this Segment