Podcast Diversity 2008 - 2018

Podcasting and Race: The State of Diversity in 2018

Recently, my colleagues Melissa Kiesche and Megan Lazovick gave a presentation at WerkIt – A Woman’s Podcast Festival. Their presentation, Closing the Listening Gender Gap, presented some new data on the composition of the podcasting audience and also some of the reasons why some women don’t listen to podcasts. You can download the full presentation here–it’s terrific–but it reminded us here at Edison that it’s a good time to revisit another aspect of this evolving medium: the growing diversity of podcast listeners.

First of all, let’s take a look at a comparison of the U.S. Population in 2018, and our most recent Infinite Dial data on monthly podcast listeners:

Podcast Listener Composition by Ethnicity

Podcast Listener Composition by Ethnicity

Here’s what is notable about the comparison of these two pie charts: absolutely nothing. And that’s wonderful news for the medium–the podcast audience today looks nearly identical to the population in general, and that means podcast producers have a wonderful opportunity to create an equally diverse portfolio of content. With the podcast audience essentially mirroring America, podcasters truly have an entrée into winning over The 52–the 52 million Americans who have heard of podcasting, but haven’t yet taken the podcast plunge.

Now, here is why two nearly identical pie charts have me so encouraged about the near term prospects of Podcasting: it didn’t used to look that way. In fact, it used to look really different:

Podcast Diversity 2008 - 2018

Just one short decade ago, the podcast audience was nearly three-quarters White, which was as much a reflection of the types of content being produced at that time as any other explanation. But over the last 10 years, podcasters themselves have become more diverse, and with that so has the universe of available content.

A lot has changed in just 10 years. Just think–back in 2008, we didn’t have Spotify, Snapchat, Uber, GPS on your phone, or Ed Sheeran. But for the ethnic diversity of podcast listeners to change so much over the last ten years is a tribute to the great democracy of podcasting: anyone with a story to tell, can tell that story. Increasingly, America–and the world–is listening.

Marketplace Edison Research Poll

Marketplace and Edison Research Reveal a Dramatic Shift in How Partisans Perceive Economic Data

The late Senator Daniel Patrick Moynihan was famously quoted as saying “Everyone is entitled to his own opinion, but not his own facts.” In the case of economic data, however, this may not be true. In the October 2018 edition of the Marketplace/Edison Research Economic Anxiety Index poll, we asked a sample of Americans how much they trusted data about the economy that is reported by the Federal Government. 60% said that they at least “somewhat” trust the data, up from 55% in October 2016. And the most extreme reaction, “Do not trust it at all,” declined from 25% to 14%.

However, if we dig a little deeper into this question, we find something remarkable. The October 2016 survey was, obviously, fielded right before the Presidential Election, in the waning days of Barack Obama’s administration. The competing worldviews in the race between Hillary Clinton and Donald Trump could not have been more different: while Clinton ran on a platform of continuing the policies of the Obama administration, Trump ran under the argument that America was not, as surveys frequently ask, “on the right track.”

At the time of the October 2016 survey, U.S. unemployment was at roughly 5%, a number which is certainly below average (i.e., more towards full employment) compared to historical trends (by comparison, five years prior to October 2016, the unemployment rate exceeded 9%). So, with the government reporting “good” news, those who were currently aligned with the administration were more likely to believe that news. However, economic prosperity is never evenly distributed, a fact that then-Candidate Trump used to his advantage when he campaigned in pockets of America that had seen significant declines in manufacturing and well-paying jobs. If you lived in Youngstown, Ohio, or Flint, Michigan, your local economy was not doing well, regardless of what the national statistics said.

As a result, there was a sharp disparity between how Clinton supporters felt about government statistics, and how Trump supporters felt, as we reported back in 2016. Then, the 55% trust/45% distrust by the total sample masked a significant partisan divide. While 86% of Clinton supporters trusted government economic statistics, only 31% of Trump supporters felt the same—indeed, 48% of Trump supporters indicated that they didn’t trust these stats at all, compared to 5% of Clinton supporters. With Republicans and Democrats overall, these differences were still highly significant: 78% of Democrats trusted government economic data, compared to 38% of Republicans.

How you interpreted this disparity likely depended on where you personally identified yourself politically. If you were a Democrat, you likely would have been inclined to cite Senator Moynihan’s quote, above. But if you were a Republican, you would likely have made the argument that the facts on the ground are different; that, despite what the national statistics say, there are significant pockets of America that are economically only getting worse. Both sides, in other words, were demanding to be entitled to their own facts.

This year we had a remarkable opportunity to revisit this phenomenon, once again just prior to a significant election, only now under a Republican administration. As noted above, the degree of trust in government statistics did tick up by five percentage points, and it should also be noted that unemployment today is even lower (currently 3.7%) than it was two years ago. While the percentage of those who trust government economic data did rise modestly from 55% to 60%, that rise once again masks a significant partisan divide.

Trust in Economic Data by Party

The October 2018 data show that the “trusters” have completely flipped positions from 2016. Today, 73% of Republicans trust government data (compared to 38% in 2016) and 51% of Democrats trust these data (compared to 78% in 2016.) This is truly a remarkable shift in just a two-year period. The percentage of Republicans who “do not trust [government economic data] at all” declined from 37% to 7%. And the “somewhat distrust” figures for Democrats rose from 12% to 32%, nearly tripling in two years.

The opportunity to revisit this question under a new administration has given us a profound insight into how Americans from either side of the aisle perceive government communications. While one might have been tempted in 2016 to proclaim that Republicans were willfully ignoring “good” economic news, the truth is that both sides are inclined to believe “facts” when they are presented by their party, and less likely when they are presented by the opposition party. Yet, the underlying data is the same: unemployment statistics tracked by the Bureau of Labor Statistics, just as they have been since 1948.

All of which brings us back to Senator Moynihan. Perhaps we are all entitled to our own facts, after all.

Election Day

Powering Democratic Gains: First-time Midterm Voters

A report from Edison Research Exit Polls for the National Election Pool

After any election, analysts and pundits debate the “narrative” created by who won and who lost.  For the 2018 midterm elections, there is a vital parallel story: The explosive increase in turnout over previous midterm elections.

The current count shows approximately 114 million people voted in this year’s midterms.  This compares to about 83 million people who voted in the 2014 elections.

In this year’s National Exit Polls we asked respondents if this was the “first time [they had] ever voted in a midterm election.”  Fully 16% of respondents said so.  This represents about 18 million people voting for the first time in a midterm election.  That’s about the same number of people who live in the state of New York – all voting for the first time in a midterm election.

This means that 60% of the boost in voting since 2014 came from first-time midterm voters (the remainder presumably having not voted in 2014 but in some previous midterm).

The first time midterm voters are extremely young – 47% of those who voted for the first time in a midterm are 18-29 years old; 36% are 18-24 – most of whom were not eligible to vote in 2014.

At the same time – fully 27% of these first-time midterm voters are 45 years old or older.

The first-time midterm voters tilted strongly to the Democrats.  While nationally 53% of the midterm electorate voted for the Democratic candidate in their House election, among the “first-timers” 62% chose the Democrat.  This means that nearly 20% of all of those who voted for Democrats around the country were first-time midterm voters.


Related to the race of this group – first-time midterm voters are significantly less white than the electorate as a whole – 46% of first-time midterm voters are non-white vs 28% of the electorate as a whole.

While the significance of the outcome of the elections can be debated, there can be little debate about their ‘success’ from a participation standpoint.  Many millions came to the polls for a midterm for the first time, and most anyone would interpret this level of engagement with a democracy as a good thing.


About Edison Research:
Edison Research (http://www.edisonresearch.com) conducts survey research and provides strategic information to a broad array of commercial clients, governments and NGOs, including AMC Theatres, The Brookings Institute, Disney, The Gates Foundation, Google, the U.S. International Broadcasting Bureau, Oracle, Pandora, The Pew Research Center, Samsung, Spotify, SiriusXM Radio, and Univision Communications. Edison Research works with many of the largest American radio ownership groups, including Bonneville, Emmis, Entercom, and Radio One. Another specialty for Edison is its work for media companies throughout the world, conducting research in North America, South America, Africa, Asia, and Europe.

Since 2004, Edison Research has been the sole provider of Election Day data to the National Election Pool, conducting exit polls and collecting precinct vote returns to project and analyze results for every major presidential primary and general election.  Edison conducts more than 100,000 interviews in a single day for this project. For the 2018 and 2020 U.S. elections, Edison will provide exit polls and tabulate the national vote across every county in the United States for ABC News, CBS News, CNN, and NBC News.

Exit poll and vote count: Providing the data that explains how America votes

The polls closed less than 24 hours ago and much of the time in the coming days will be spent analyzing how Americans voted in the 2018 midterm elections.  At Edison Research we are proud that the exit polls we provided to the members of the National Election Pool (ABC, CBS, CNN and NBC) will be the principal information guiding these discussions.

Election day was the culmination of years of planning at Edison as we assembled and organized over 6,000 workers including exit poll interviewers, field supervisors, county vote reporters, data entry operators, and data analysts to collect crucial data in the 2018 midterms. Our thanks to thousands of staffers who stood in rain and wind to conduct exit polls, who stayed up until the wee hours of the morning to report vote counts, and who dedicated their election day to this massive undertaking that provides essential data in the election process.

Starting before the polls opened morning through the polls closing, thousands from Edison’s survey team were stationed at local voting precincts in every state to collect exit poll data from over 100,000 voters. The national exit poll is the only nationwide voter information on voting trends, differences in vote patterns by gender, age, region, and other demographic and geographic groups that is based on verified voters. Exit poll data is important because it is gathered after votes are cast, so the respondents are people who have actually voted.

Exit poll data is incredibly valuable, not only to predict the winners of political races, although that makes for compelling viewing on election night, but also to provide analysis about the issues that drive American voters to make their choices. Edison’s national exit poll data continues to be the only one conducted in the United States. Along with the national exit poll, Edison’s state surveys are providing more in-depth analysis of all the key U.S. Senate and Governor’s races.

After the polls closed yesterday and Edison’s exit polling had been concluded, our vote collectors then went to work, covering over 4,000 voting locations across all 50 states to provide data for the tabulation of the national vote count. These vote collectors worked into the early morning hours today, gathering data from every precinct reporting in America, helping Edison provide accurate and timely vote returns to viewers across the country.

“We are proud of how our polling and data collection guided our clients’ coverage of election night and how it continues to lead the analysis and conversation on how America votes,” said Joe Lenski, Edison Executive Vice President and leader of Edison’s election efforts.

National Election Pool And Edison Research To Once Again Conduct Exit Poll of Record for 2018 Midterm Election

The National Election Pool (ABC, CBS, CNN and NBC) and Edison Research will once again conduct the Exit Poll of record in the United States. Using methodology that has provided reliable estimates and served journalists and scholars for more than a generation, we will interview voters across the country after they cast their ballots in this year’s critical midterm elections in the largest single-day data collection effort in the U.S. after the Census.

Our Exit Poll team will interview voters at hundreds of polling locations in all 50 states. The in-person interviews that form the core of our data guarantee that our polls and our projections reflect the views of actual voters nationwide. No poll conducted solely by telephone or online can ever promise that level of accuracy. Americans who tune in to our broadcasts or look at the poll results on our websites can be sure that they are seeing an accurate result that includes the opinions of people just like them.

The Exit Poll’s methodology and question wording is always being refined and edited to maintain the highest methodological standards and ensure that our practices are in keeping with the best practices of the survey research industry. Here are some of the new methodological efforts we will be implementing for this year’s Exit Poll.

Americans are increasingly voting early and dropping off absentee ballots in advance of Election Day. The Exit Poll incorporates these voters in two ways: by interviewing early and absentee voters by telephone in the final days of the campaign as the NEP has done since 2004, and in a new way this year, by deploying in-person interviewers to early polling locations in the days before the election in key states.

The results of the Exit Poll in 2016 demonstrated that knowing a voter’s educational background is critical to understanding American politics today. To ensure that the Exit Poll continues to accurately reflect the views of voters by their educational background, we have modified the question wording to more closely mirror the wording used by the Census Bureau and make it easier for voters taking the polls to accurately categorize their educational background and to better facilitate comparisons to Census data for later scholarship.

Changes to question wording on the Exit Poll are common. In the entire history of modern exit polling conducted by members of the National Election Pool dating back to 1972, just one question has been asked with the same question wording over that entire time.

The NEP has also added the ability to apply weights to Election Day interviews for voter characteristics that are not easily observable by the interviewers conducting the polls. In the past, exit polls were weighted only for such characteristics, including gender, age and race. Now, the Exit Poll will be adjusted for non-response by education and by age using a parameter developed by comparing past Census estimates of turnout among these groups and past Exit Poll estimates of turnout among these groups. While this adjustment does not meaningfully change the results of the Exit Poll in terms of the voting behavior of various subgroups, it does serve to adjust the poll’s estimates of the age and education makeup of the day’s voters in a way that is a best practice for polling in today’s environment.