Changes to the Facebook Algorithm Decreased News Visibility Between 2021-2024

2025-07-25

AI Summary

This study investigates the impact of Facebook’s algorithmic adjustments (2021–2024) on news visibility, addressing controversies surrounding Meta’s “News Ban” policy and alleged news de-prioritization. Leveraging longitudinal engagement data—over 5 million news and non-news posts from 2016 to 2025—and integrating policy-relevant temporal breakpoints, we apply causal inference methods to isolate algorithmic effects. Results show a 78% decline in news engagement during the period, attributable primarily to systematic algorithmic downranking of low-quality news sources—not diminished user interest or reduced news supply. Following a 2025 policy reversal, news engagement rebounded significantly, with disproportionate recovery among previously demoted low-quality sources, further confirming algorithmic causality. Methodologically, the study innovatively combines large-scale social media tracking, quasi-experimental comparative analysis, and time-series trend modeling. It delivers the first rigorous empirical evidence of platform algorithm-driven visibility shifts in news dissemination, elucidating the dynamic causal pathway linking platform policy → algorithmic implementation → downstream engagement outcomes, thereby informing evidence-based digital content governance.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Application Domains: Misinformation & Fake News
• Search and Optimization: Algorithm Configuration

Application Category

• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Web Mining and Content Analysis: Web data quality in the era of algorithmically-generated content
• Social Networks and Social Media: Fairness and bias in social network and social media analysis

Abstract

Platforms, especially Facebook, are primary news sources in the US. In its widely criticized "War on News," Meta algorithmically deprioritized news and political content. We use data from 40 news organizations (5,243,302 Facebook posts, 7,875,372,958 user reactions) and 21 non-news pages (396,468 posts; 1,909,088,308 reactions) between January 1, 2016 and February 13, 2025 to examine how these changes influenced news visibility on the platform. Reactions to news declined by 78% between 2021 and 2024 while reactions to non-news pages increased, indicating targeted suppression of news visibility. Low-quality sources were especially suppressed, yet the 2025 end to "War on News" increased user reactions to news, especially low-quality ones. These changes do not reflect decreased news supply, Facebook user base, or interest in news over this period.

Problem

• Examining Facebook algorithm changes reducing news visibility
• Analyzing targeted suppression of news content reactions
• Assessing impact of Meta's policy shifts on news engagement

Innovation

• Analyzed 40 news organizations' Facebook data
• Measured 78% decline in news reactions
• Detected targeted suppression of news visibility