Misinformation posing as news has been a problem that only got worse with the ease of publishing on the Internet. But the COVID-19 pandemic seems to have raised it to new levels, driving lots of attention to rumors, errors, and outright falsehoods. Given the magnitude of the threat, there would seem to be a premium placed on ensuring the accuracy of any pandemic information. But that doesn’t seem to be the case.
It’s unlikely there will be a single explanation for why that was the case. But researchers based in Paris have looked into the dynamics of pandemic news and found a potential contributor: Unreliable news sources were better at producing content that matched what readers were looking for.
Supply and demand
The researchers behind the new work treated the news ecosystem as a function of supply and demand. The audience—in this case, the Italian public—is interested in obtaining answers to specific questions or details on a topic. News sources attempt to satisfy that demand. Complicating this relationship, the news ecosystem includes organizations that don’t produce quality information. Poor reporting can be due to carelessness or to satisfy an agenda separate from providing news.
It’s one thing to describe all of that; it’s a separate thing to figure out how to get numbers that allow you to analyze how the ecosystem functions.
In this case, the researchers had access to several tools to help them get at these issues. For starters, there is a database of all the articles published by Italian media, allowing the researchers to analyze the supply half of the equation. Separately, there’s a database of Italian media sources that fact-checkers have identified as unreliable due to publishing errors or misinformation, allowing them to be analyzed independently from general media. The database isn’t a complete list of unreliable sources, so this shouldn’t be considered a comprehensive analysis of the misinformation landscape.
For the demand side, they turned to Google Trends, which tracks the search terms people are interested in and the information they’re looking to discover. Google Trends is also an imperfect method since it doesn’t cover the whole populace (some won’t use the Internet or rely on other search providers), but it still gives a general picture of what many people are interested in.
A matter of semantics
To study the behavior of the news marketplace during the pandemic, the researchers chose several popular search terms. These included “coronavirus” to track its pandemic coverage and five control terms, like “eurovision” and “papa francesco” (the Pope), which were also popular during this time. These are tracked starting from when the virus was first identified in China, through several waves of infections to August 2020.
One clear thing was that the public’s interest (as reflected in search terms) would appear, on average, the day before reliable media reported on the topic. But news from unreliable sources was notably quicker than the average news source. The faster speed wasn’t simply because the less reliable sources tended to be online publications, thus able to react more quickly. Even when compared to reliable online news outlets, the unreliable ones tended to be quicker to place stories that reflected search terms.
The researchers also looked at other search terms that appeared simultaneously as “coronavirus”—things like “coronavirus N95 mask.” When this context was considered, it turned out that articles from questionable sources fit the public’s interest better than the general news media. This held true daily throughout the entire study period, indicating this is a persistent pattern.
Ultimately, unreliable sources appear to produce relevant articles more quickly and better match the content of those articles with the things the public is most interested in. The study isn’t able to address why these outlets are better able to meet the public’s interests, but the finding hints at why misinformation spreads widely during the pandemic.
To turn it into more than a hint, we’d need to look beyond Italy and see whether this behavior happens anywhere else. To help with that, the researchers developed an index based on the mathematic relationships identified here between search terms and news output. This index should also allow us to determine if the relationship applies to other topics where misinformation proliferates.