Back to News
NewsInflationInterest RatesFederal Reserve

Engaging in Arguments on X May Lead to More Unwanted Content, Study Finds

A study shows that engaging in arguments on X can lead to more unwanted content, especially for Democrats, due to algorithmic biases.

Mercation Admin
August 18, 2026 2 min read
Engaging in Arguments on X May Lead to More Unwanted Content, Study Finds

Study Findings Reveal Algorithmic Discrepancies

A recent study has found that engaging in arguments on X may lead to an increase in unwanted content for users, particularly those with specific political affiliations. Conducted by researchers from prestigious institutions including MIT and Stanford, the study analyzed the behavior of 715 active users in the U.S. over a two-month period.

The study, published in the Proceedings of the National Academy of Sciences, indicates that X's "For You" algorithm tends to promote posts that conflict with users' declared values. This effect is notably stronger among Democrats, who experienced a correlation score of -0.229, compared to Republicans, who had a score of -0.041. A lower score signifies a greater misalignment between the content promoted and users' stated beliefs.

Implications of Engagement Behavior

Researchers suggest that responding to posts opposing one's beliefs may inadvertently signal to X that a user desires more of such content. This finding highlights a potential flaw in how social media algorithms interpret engagement, often amplifying content that contradicts users' explicit values. On average, the content amplified by the algorithm had a score of -0.150, contrasting with posts from followed accounts that typically aligned more closely with users' values.

The study also revealed similar results in a repeat analysis, with Democrats showing a correlation of -0.199 and Republicans at 0.0372. The researchers concluded that this discrepancy illustrates the tension between users' values and the engagement-driven content that social media algorithms prioritize.

Future Research Directions

While the study sheds light on these algorithmic outcomes, it did not explore whether the accounts behind the posts that elicited misaligned responses were operated by humans or bots, leaving a gap in understanding the full extent of influence on content promotion.

This research comes on the heels of X’s decision to open-source its Grok-powered algorithm earlier this year, revealing its method of ranking posts based on predicted user interactions. Elon Musk acknowledged that the algorithm requires significant improvements, asserting the company's commitment to enhancing transparency in its algorithmic processes.