The headlines that travel farthest on social media are not always the most reliable, and that pattern appears to persist across today’s fractured platforms. In a sweeping cross-platform analysis published in Proceedings of the National Academy of Sciences, researchers examined nearly 11 million posts with news links shared in January 2024 across X/Twitter, Bluesky, LinkedIn, Mastodon, Truth Social, Gab, and GETTR. Their central finding: when the same user shares links, their lower-quality news posts consistently draw more engagement than their higher-quality ones.
The team, led by Mohsen Mosleh with coauthors Jennifer Allen and David G. Rand, paired those engagement metrics with domain-level quality ratings built from multiple expert sources. They also estimated each outlet’s political lean. The result is a rare panoramic view of news sharing in an era when social media looks less like a single town square and more like a patchwork of neighborhoods, each with its own house rules and lawn signs.
One striking detail lands early: average news quality differs by platform, with more conservative-leaning platforms hosting a larger share of low-quality domains. Yet the authors undercut a popular trope about a persistent right-wing advantage online. Engagement for partisan content depends on the platform’s dominant audience. Conservative news does better where conservative users cluster; liberal news fares better on left-leaning sites. That is, the crowd matters.
“If your post is in line with the norm on the platform, people engage with it more,”
That observation, from senior author David Rand of Cornell University, helps frame the paper’s most unsettling throughline: regardless of political lean, lower-quality news links tend to earn more likes and reshares from the same account than higher-quality ones. On Mastodon, where posts appear chronologically without engagement-based ranking, the same tendency holds. Algorithms may amplify what people already prefer, but the appetite for spicier, lower-quality fare predates the boost.
Seven Platforms, One Persistent Pattern
Methodologically, the study leans into rigor. The dataset includes every post with an identifiable news-domain link for the month across the seven platforms, capturing the universe of link-sharing rather than a convenience sample. The authors use user fixed effects, a workhorse approach that compares each person to themselves, canceling out follower counts, posting cadence, and other individual traits. They log-transform skewed counts, probe robustness with alternative models, and examine both likes and reshares. Where available, they even look at views: on X/Twitter, lower-quality links also garnered more impressions, though the gap was smaller than for engagement.
Crucially, high-quality news still dominates in volume. Across platforms, users post many more links to reputable outlets, and those posts attract more total engagement in aggregate simply because they are more common. The paradox is local, not global: on a per-user basis, any dip into lower-quality territory tends to perform a little better.
Anyone who has watched a sensational headline rocket through a feed will recognize the dynamic. The picture the authors paint is not of a social web overrun by misinformation, but of an engagement economy where the tastiest morsels are often the ones with the least nutrition. Imagine a crowded plaza at dusk, screens glowing like fireflies. The items that shimmer brightest are not necessarily the most nourishing, but they are very clickable.
Echo Platforms, Not Just Echo Chambers
The paper also repositions a decade of single-platform conclusions. Past research often generalized from Twitter or Facebook; this work shows why that can mislead. Patterns diverge: political lean tracks the local audience, whereas quality’s engagement advantage remains consistent. That combination yields a layered message for platform designers, journalists, and readers. First, do not presume one side owns the engagement game. Second, user preferences, not only algorithms, help explain why flimsy links pop.
There is nuance in the domain-level analysis, too. The negative association between quality and engagement seems propelled less by fringe sites overperforming and more by stalwarts underperforming. Posts linking to major high-quality outlets like the New York Times or Reuters tend to draw relatively fewer interactions from the same users. Familiarity, paywalls, or audience fatigue may play roles, but the pattern survives controls for paywalls and language.
“Many had argued that people on the right were better at getting engagement on social media, but we find that it totally depends on the platform.”
Policy implications follow. If user taste leans toward the provocative, interventions that focus purely on ranking systems will have limits. Transparency about source quality, friction for resharing, or prompts that reward clarity over heat might help. Yet the authors are cautious: their data capture a moment in time, January 2024, and major platforms missing from the analysis, such as Facebook or TikTok, pose measurement hurdles. Still, across seven very different sites, the signal is hard to ignore.
For newsrooms, the practical takeaway is bracing. Quality remains abundant and collectively engaging, but the quick-hit metric scoreboard can punish sober reporting at the post level. For readers, the reminder is simpler. If something seems engineered for outrage, it probably is. And it will likely perform.
Proceedings of the National Academy of Sciences: 10.1073/pnas.2425739122
ScienceBlog.com has no paywalls, no sponsored content, and no agenda beyond getting the science right. Every story here is written to inform, not to impress an advertiser or push a point of view.
Good science journalism takes time — reading the papers, checking the claims, finding researchers who can put findings in context. We do that work because we think it matters.
If you find this site useful, consider supporting it with a donation. Even a few dollars a month helps keep the coverage independent and free for everyone.