New Evidence Shows Blaming Social Media for Political Polarization Is Misguided

Critics blame social media for a host of societal ills, from mental illness among teens to the decline of local journalism, often without concrete evidence. A perennial example of this, particularly in the context of election campaign cycles, is the tendency to blame social media for political polarization. But while algorithms may make convenient scapegoats, there is little evidence that they are to blame for the trend toward partisan polarization—and attempting to regulate them based on that faulty premise would lead to ineffective solutions, at best. In fact, doing so would more likely harm consumers.

It is true that Americans are more politically polarized now than in previous decades. From 1994 to 2014, the overlap in political values between Republicans and Democrats shrank and attitudes toward the opposing party worsened, according to Pew Research Center. During the 2020 presidential campaign, only about one in five Biden supporters believed they shared the same values and goals as Trump supporters, and vice-versa. Congress reflects this polarization: Republican and Democrat lawmakers are farther apart ideologically than they have been any time in the past 50 years.

Critics argue that social media exacerbates political polarization by creating “filter bubbles.” This term describes a situation in which individual users tend to consume and interact with political content they agree with, so algorithms show them more content in the same vein and less they would disagree with. This feedback loop eventually reaches a point where users almost exclusively see content they agree with and rarely see content they would disagree with, which reinforces their existing political beliefs and limits their exposure to diverse opinions, thereby reducing the likelihood they will change their own views. Some critics even claim that social media platforms create these algorithmic filter bubbles and fill them with extremist far-right or far-left content intentionally to increase engagement.

Yet existing research does not support the claim that social media is a primary factor driving political polarization. Four new papers published in the journals Science and Nature on July 27, 2023 studied the impact of Facebook and Instagram on key political attitudes and behaviors during the 2020 election cycle. The papers found little evidence that key features of these platforms led to polarization. For example, one study found that replacing Facebook and Instagram’s default algorithmic feed with a reverse-chronological feed did not significantly impact political polarization or political knowledge, contradicting the argument that algorithms create filter bubbles rather than, at most, replicating individuals’ existing selection bias. This finding is important since many lawmakers have proposed legislation to regulate social media algorithms, especially when they deviate from chronological feeds.

Older studies support these findings. From 1996 to 2012, political polarization increased the most in demographic groups least likely to use the Internet and social media, adults older than 75. Between 2016 and 2019, only 4 percent of Americans were isolated into partisan filter bubbles online, compared to 17 percent of Americans who consumed TV news from only partisan sources. Furthermore, the narrative about filter bubbles may not even hold true, as exposure to opposing views on social media can increase political polarization rather than reduce it.

Overall, the relationship between social media, or the Internet more generally, and political polarization is inconsistent. It may be that only individuals who frequently consume news online and are highly interested in politics experience polarization. Additionally, news audience polarization, both offline and online, is the highest in the United States, and the United States is polarizing faster than other democracies, suggesting that certain unique aspects of American politics, traditional media, and culture drive polarization.

Furthermore, the existence of filter bubbles on social media is not a unique or unprecedented phenomenon. Just as individuals may follow social media accounts that share their beliefs and unfollow or block those that do not, Pew found in 2014 that in the physical world 66 percent of consistent conservatives and 52 percent of consistent liberals mostly befriended those with similar views, while 24 percent of consistent liberals and 16 percent of consistent conservatives said they stopped talking to or being friends with someone over politics.

There is likely no single factor that is primarily responsible for political polarization in the United States. Relying on the argument that social media causes political polarization distracts from other potential factors, leads to bad policies that would negatively impact users, and is unlikely to produce effective solutions to the problem.