HMN 2026: How Do political social media ads influence the outcome of elections?

Social media platforms, such as Facebook, Instagram and X, are accessed daily by millions of people worldwide. In the weeks or months leading up to elections, many political parties use social media platforms as part of their campaigns to promote candidates, raise funds, or disseminate their proposed policies.

While some of this content is diffused as regular posts via political channels or accounts, political parties can also pay to post adverts. The extent to which online political ads influence the opinions and voting behavior of social media users, however, has not yet been clearly determined.

Researchers at Stanford University, Meta and University of Pennsylvania recently carried out a study aimed at investigating the effects of political advertising on people’s views and electoral votes. Their paper, published in Nature Human Behaviour, specifically focused on the impact of political ads posted on Facebook and Instagram before the 2020 U.S. presidential election.

Exploring the effects of political ads on social media

As part of their study, the researchers randomly selected 36,906 Facebook users and 25,925 Instagram users. For the six weeks preceding the 2020 U.S. presidential election, they removed all political ads from the feeds of a random subset of these users, while the remaining users viewed ads normally.

In addition, the team tracked the behavior of all the users considered in their study using both surveys and data collected by Meta. This allowed them to investigate the effects of the ads, while also uncovering patterns about what users the ads targeted and what their objective was.

The results of the researchers’ analyses suggest that most political ads during the 2020 U.S. election targeted people who already supported the advertised political party. The goal of the ads, therefore, primarily appeared to be to mobilize supporters or remind them to vote for the party, rather than persuade people with different political inclinations.

“We show that most presidential ads were targeted towards parties’ own supporters and that fundraising ads were the most common,” wrote Hunt Allcott, Matthew Gentzkow and their colleagues in their paper. “On both Facebook and Instagram, we found no detectable effects of removing political ads on political knowledge, polarization, perceived legitimacy of the election, political participation (including campaign contributions), candidate favorability and turnout. This was true overall and for both Democrats and Republicans separately.”

Reinforcing existing beliefs

The findings gathered by Allcott, Gentzkow and their colleagues suggest that online political ads primarily influence loyal voters, convincing them to fund or support the party that their views are typically aligned with. In other words, they might not be as persuasive as some people believe them to be and might not actually convince social media users to change their political views or vote for a party that they don’t generally support.

The researchers observed no detectable changes in the political knowledge or participation of social media users, their trust in the election result, the candidate they supported or the donations they made that could be linked to ads on Facebook or Instagram. Notably, the study only focused on paid political ads and did not account for other non-paid political content (e.g., posts, videos or comments) that users could still encounter online.

Future studies could carry out additional analyses considering a broader range of political content and information that people might encounter while using social media or accessing other online platforms. This could help to better elucidate the possible impact of information disseminated online on people’s beliefs, political polarization and the outcome of elections.

The content is provided for information purposes only.