Brookings: “Using a combination of analytical methods – including natural language processing, machine learning, and manual classifying – this research provided the first overarching assessment of the role that popular political podcasting plays in spreading unsubstantiated and false claims. These podcasters, who span the political spectrum, make up the mainstream of the medium and regularly boast audiences in the millions. Drawing on data collected from 36,603 episodes produced by 79 prominent political podcasters, 17,061 evaluations, and 184 key terms and phrases, this analysis found:
- The spread of unsubstantiated and false claims across the popular political podcasting ecosystem was common: Nearly 70% of podcasters in the dataset shared one or more such claim, and at least one out of every 20 episodes (1,863 episodes) recorded by prominent podcasters featured an unsubstantiated or false claim. Due to the conservative nature of the research design, this is likely an undercount of the proliferation of this type of content across political podcasting.
- Unsubstantiated and false claims tied to the 2020 US presidential election spiked dramatically after the election and did not abate in the following months, despite multiple failed legal challenges: In the lead up to the January 6 attack on the US Capitol, political podcasters played a central and highly partisan role in sowing doubt about the security of the 2020 election. After Election Day, claims of election fraud rose by nearly 600%, to more than 28% of all episodes airing between the election and January 6, 2020, with the vast majority coming from conservative series that were among the most popular during this period.
- During the first two years of the coronavirus pandemic, unsubstantiated and false claims circulated widely, but they tended to be more nuanced and less partisan than election-related claims: Unsubstantiated and false claims tied to the pandemic featured in one out of every eight episodes examined during this period. These claims, however, tended to be less overtly partisan and were often rooted in misunderstandings of science, confusion over evolving public health guidelines, and a heavy reliance on non-peer reviewed preprints.
- 10 prominent podcasters were responsible for sharing the majority of false and misleading content: The 10 podcast series that shared unsubstantiated or false claims at the highest overall rates accounted for more than 60% of all the dataset’s unsubstantiated and false claims. These series, whose hosts lean conservative politically, had a combined reach of more than 28 million followers on Facebook and Twitter as of September 2022.”