In the dynamic landscape of modern political campaigns, social media platforms such as Facebook and Instagram have emerged as critical arenas for political advertising. Their unparalleled reach and sophisticated targeting capabilities have reshaped the way political messages are disseminated and consumed. Yet, despite widespread concern over the influence of these platforms on voter behavior and democratic processes, empirical evidence quantifying the actual impact of social media political ads remains scarce. Now, a groundbreaking study published in Nature Human Behaviour offers compelling insights by leveraging one of the largest randomized controlled trials ever conducted on this topic.
This exhaustive research involved randomizing tens of thousands of users—36,906 on Facebook and 25,925 on Instagram—to a treatment group where all political advertisements were removed from their news feeds for a substantial six-week period leading up to the high-stakes 2020 US presidential election. By creating a controlled environment simulating a political-ad-free social media experience, the authors sought to directly measure the causal effects of exposure to such ads on various political and electoral outcomes, ranging from knowledge and polarization to participation and perceived legitimacy.
One of the most striking findings from the study is the revelation that the bulk of political advertisements on both Facebook and Instagram were predominantly targeted toward a candidate’s own supporter base rather than swing voters or political opponents. This highly strategic focus on mobilization over persuasion suggests that political advertisers prioritize consolidating existing support rather than attempting to change minds—a nuance that challenges common public assumptions about the persuasive power of political ads on social media.
Moreover, the data uncovered that fundraising ads constituted the majority of the political ad inventory on these platforms. These ads aim to solicit donations rather than sway opinions or encourage voter turnout directly. The dominance of fundraising efforts underscores a critical economic dimension within political advertising strategies, as campaigns leverage social media not only for communication but also for resource generation.
Against this backdrop, the researchers meticulously analyzed a variety of outcome measures, including political knowledge, polarization indices, perceived election legitimacy, political participation metrics (such as campaign contributions), candidate favorability ratings, and actual voter turnout. Intriguingly, their results revealed no statistically significant or meaningful effects of removing political ads on any of these dimensions, either on Facebook or Instagram. This absence of measurable impact held consistently across the general user population and when stratified by political affiliation—Democrats and Republicans alike showed no detectable reaction to the removal of political advertising.
The robustness of these findings was reinforced by the large sample sizes and the proscribed six-week intervention window, spanning the crucial pre-election period characterized by intense political messaging. The null effects challenge the dominant narrative that social media political advertisements wield considerable influence in shaping election outcomes or in deepening political divisions among the electorate.
One plausible interpretation for these null results is that political attitudes and behaviors are largely entrenched and influenced by multiple factors beyond short-term exposure to advertising. Voters may rely on a broad array of informational inputs, social networks, personal experiences, and traditional media, diluting the potential influence of any single source—no matter how pervasive it appears.
Additionally, the fact that parties’ own supporters received the lion’s share of ads might mean these advertisements mainly served to reinforce pre-existing political preferences without substantially altering voter knowledge or engagement. In this light, political advertising on social media appears more as a tool for maintaining base enthusiasm rather than a mechanism for persuasion or voter mobilization.
The study also sheds light on the complex interplay between technology platforms and democratic processes. While social media’s role in politics has heatedly debated—amid fears about misinformation, foreign interference, and echo chambers—this rigorous randomized field experiment provides a tempered perspective, suggesting that fears about the outsized influence of paid political advertisements may be overstated.
Furthermore, the investigators’ decision to cover both Facebook and Instagram enhances the generalizability of the findings, given Instagram’s younger demographic profile compared to Facebook’s broader, often older user base. The consistent null effects across both platforms underscore that the impact—or the lack thereof—transcends user age and platform-specific dynamics.
It is also significant to consider the temporal context: the 2020 US presidential election was a uniquely polarized and high-profile political event with unprecedented media coverage and public engagement. The findings indicate that even amid such historic electoral circumstances, political ads on social media did not materially sway voter behavior or perceptions in any meaningful way.
This extensive research contributes to academic and policy debates by providing a rare empirical baseline against which claims about social media political advertising can be evaluated. It also presents important implications for regulators, political strategists, and platform policymakers who grapple with how to manage political content to safeguard democratic integrity without unduly censoring speech.
The authors prudently emphasize that while their study focuses on the role of paid political advertisements, it does not address other forms of political communication on social media such as organic posts, misinformation campaigns, or coordinated inauthentic behavior, which may have distinct and potentially more powerful effects.
Overall, the findings encourage a nuanced understanding of social media’s political ecosystem, highlighting that paid political ads, despite their visibility, might not be the potent causal force in shaping electoral outcomes or polarization that many presumed. It signals a shift toward re-examining the tools and channels that truly drive political influence in the digital age.
Future research might build on this foundation by exploring the effects of different types of political content, varying user demographics, and longitudinal impacts beyond a single election cycle. By further dissecting the multifaceted information environment of social media, scholars can deepen understanding of political communication’s evolving landscape.
In essence, this landmark study affirms the importance of rigorous, evidence-based inquiry into the digital infrastructures of democracy, cautioning against simplistic assumptions about the influence of social media political advertising. It also provides a beacon for transparent, data-driven policy discourse aiming to uphold democratic ideals in an increasingly interconnected world.
By leveraging randomized controlled trial methodologies at unprecedented scale, the research sets a new standard for political communication scholarship and represents a vital step forward in deciphering the complex relationship between technology, politics, and society in the 21st century.
Subject of Research: The causal effects of political advertising on social media platforms (Facebook and Instagram) prior to the 2020 US presidential election.
Article Title: The effects of political advertising on Facebook and Instagram before the 2020 US election.
Article References:
Allcott, H., Gentzkow, M., Levy, R. et al. The effects of political advertising on Facebook and Instagram before the 2020 US election. Nat Hum Behav (2026). https://doi.org/10.1038/s41562-025-02328-w
Image Credits: AI Generated
