The occurrence of untrustworthy headlines receiving greater traction on social media than their reliable counterparts is well-established and continues to be evident across multiple platforms. A detailed cross-platform analysis published in the Proceedings of the National Academy of Sciences examines this matter by reviewing around 11 million posts with news links shared in January 2024 across X/Twitter, Bluesky, LinkedIn, Mastodon, Truth Social, Gab, and GETTR. The research team, led by Mohsen Mosleh alongside Jennifer Allen and David G. Rand, found that when users share links, posts featuring lower-quality news consistently garner more interaction than those with higher-quality content.
The analysis integrated engagement metrics with quality ratings assigned to domains from various expert sources and assessed the political bias of each outlet. This thorough investigation uncovers the intricacies of news sharing in a fragmented social media landscape that resembles a collection of separate neighborhoods, each with its own regulations and preferences.
Notably, the average quality of news varies by platform, with conservative-leaning platforms showing a trend towards a higher quantity of low-quality domains. Despite this, the researchers refute the idea of a persistent right-wing advantage online. The engagement for partisan content is shaped by the predominant audience on each platform. Conservative news thrives on platforms with mainly conservative users, whereas liberal news finds more success on platforms with a left-leaning demographic. Therefore, engagement pertains more to audience alignment than to intrinsic content bias.
The study employs a particularly rigorous methodology, utilizing a comprehensive dataset of posts linked to news domains across seven platforms throughout the month. Employing user fixed effects permits comparisons that negate individual differences like follower counts or posting frequency. They investigate engagement via likes, reshares, and, when feasible, views, uncovering that lower-quality links frequently attract more impressions on X/Twitter, albeit the difference is less pronounced compared to other forms of engagement.
Although lower-quality news links achieve more engagement per user, respected outlets dominate in terms of total posting volume and collective engagement. The contradiction lies in the individual propensity to interact more with sensational material. This illustrates an engagement economy where attention-grabbing but less reliable headlines are more appealing, even in the absence of substance.
The findings contest earlier research conclusions primarily derived from platforms such as Twitter or Facebook by shedding light on platform-specific tendencies in political affiliations and engagement behaviors. This indicates that strategies to combat misinformation should take into account platform-specific audience dynamics instead of viewing social media as a singular entity.
High-quality news outlets do not achieve as much effectiveness in engagement on a per-post basis, potentially due to reasons like audience familiarity, paywalls, or content saturation. The study cautions against relying solely on algorithmic solutions for misinformation, advocating for increased transparency regarding source quality and initiatives to foster thoughtful engagement.
In the end, the research highlights the essential role of maintaining high-quality journalism despite prevailing engagement metrics that favor more sensational content. For readers, it serves as a reminder of the necessity for discernment when interacting with news on social media, as content designed to provoke outrage frequently garners the highest levels of interaction.