In the days following Hurricane Milton and Hurricane Helene, misinformation about FEMA’s response in Florida and North Carolina dominated the news cycle. Pundits, politicians, and residents alike alleged that the country’s foremost disaster management agency had prevented evacuations, blocked recovery and rescue operations, and even discriminated in its distribution of aid.
None of these rumors were found to be true, but the speed and intensity at which this misinformation spread signal a growing difficulty in discerning truth.
With the presidential election just around the corner, many observers are concerned about “the death of truth”—the idea that voters struggle to distinguish real news from fake news—and the concept of a parallel universe, in which voters only pay attention to and believe news that suits their partisan beliefs.
There has been substantial debate in past elections, going back to 2016, about what voters knew and believed, whether they truly understood the news, and if fake news has swung the vote. Becoming aware of what voters know and how it affects their voting decisions is crucial to understanding election dynamics.
To grasp voters’ knowledge and how it varies over the election cycle, Columbia Business School researchers worked with colleagues at MIT Sloan and Northwestern University to build a map of voters’ news knowledge, examining where it exists and how it fluctuates. The goal was to see if the “death of truth” phenomenon is real and, if so, to quantify it.
A related phenomenon is the parallel universe hypothesis—the idea that Republicans and Democrats live in almost parallel universes. As former President Obama once explained, what voters know about recent news often depends on their ideology. The researchers aimed to understand this more deeply.
To conduct the research, we recruited a panel representative of American voters, consisting of around 11,000 people. We assessed their knowledge of recent news stories, their choices in which stories to follow, and their ability to distinguish real news stories from fake ones.
This multi-step process began with a panel of journalists who identified what they considered the three most important stories about U.S. domestic politics. These stories served as the “real” ones.
We then examined participants’ ability to distinguish these from three widely circulated fake stories sourced from Snopes, a website that collects such misinformation. After lining up the real and fake stories, we asked our subjects to differentiate between them, providing only limited time and small incentives to ensure their engagement. We also asked the journalists to create a third set of stories that were completely false but as plausible as possible.
After running the experiment for eleven months, we found substantial heterogeneity among voters. Between 50 and 55 percent of Americans were proficient at distinguishing real from fake stories and could identify major news events. The rest of the population was more confused, with some struggling to identify even significant events, such as Biden’s election.
Further exploring the parallel universe hypothesis, we found that Republican voters were more likely to perceive stories favorable to the Republican Party as true, and Democratic voters were more inclined to believe stories favorable to Democrats. We relied on journalists to identify which stories favored which party. However, the partisan effect was relatively small, accounting for only two to three percent of the variation. The larger explanatory factors were socioeconomic variables like age, income, education, gender, and ethnicity. We found that wealthier, more educated, older, predominantly male voters were more likely to correctly identify real stories—reflecting a socioeconomic divide in the U.S.
Our conclusion in this recently released research is that standard socioeconomic variables have more explanatory power than polarization in understanding heterogeneity in voter knowledge in the U.S. While the parallel universe effect exists, it is relatively minor compared to the impact of socioeconomic factors.
Although many would like to see an end to polarization, fake news and the partisan divide may not be shaping public opinion as much as expected. Instead, demographic factors in this diverse nation play a larger role—something candidates may want to consider as they proceed with their campaigns.
