Engagement, User Satisfaction, and the Amplification of Divisive Content on Social Media
2023/05/26 by Smitha Milli, Milli, Smitha, Micah Carroll +9 · 11 voices · 16 citations
#cs.SI #cs.CY
paper · pdf · doi:10.48550/arxiv.2305.16941
Abstract
In a pre-registered algorithmic audit, we found that, relative to a reverse-chronological baseline, Twitter's engagement-based ranking algorithm amplifies emotionally charged, out-group hostile content that users say makes them feel worse about their political out-group. Furthermore, we find that users do not prefer the political tweets selected by the algorithm, suggesting that the engagement-based algorithm underperforms in satisfying users' stated preferences. Finally, we explore the implications of an alternative approach that ranks content based on users' stated preferences and find a reduction in angry, partisan, and out-group hostile content, but also a potential reinforcement of pro-attitudinal content. The evidence underscores the necessity for a more nuanced approach to content ranking that balances engagement and users' stated preferences.
Cited by
Discussions
- Twitter's Algorithm: Amplifying Anger, Animosity, and Affective Polarization [hn, 106 points, 109 comments]
- Combating Misinformation Runs Deeper Than Swatting Away ‘Fake News’ [lemmy, 82 points, 7 comments]
- Great question. Difficult to test, but here's some work showing algorithmic feeds amplify hostile content relative to reverse chronological baseline. arxiv.org/abs/2305.16941 [bsky, 52 points, 1 comments]
- アルゴリズム生成「おすすめ」フィードは時系列フィードよりも怒りを増幅し党派性を強調、人々の分断を招いている https://arxiv.org/abs/2305.16941 7カ国中6カ国で、右派アカウントの方が左派より増幅されていた。アルゴリズムの影響 https://blog.twitter.com/en_us/topics/company/2021/rml-politicalcontent [bsky, 7 points, 1 comments]
- Incentive structure of both: attention and engagement (advertising $$$). Lots of great work on this in Communication, HCI, and adjacent disciplines. See e.g. - arxiv.org/abs/2305.16941 [bsky, 7 points, 0 comments]
- Twitter/X algorithms, massively promote polarisation and anger. "Engagement, User Satisfaction, and the Amplification of Divisive Content on Social Media" arxiv.org/pdf/2305.169... [bsky, 6 points, 0 comments]
- More broadly, algorithmic ranking systems that prioritise engagement tend to amplify emotionally charged, out-group hostile content, which polarises users. 8/14 arxiv.org/abs/2305.16941 [bsky, 3 points, 1 comments]
- Sí, pero también: arxiv.org/pdf/2305.16941 "Twitter’s engagement-based ranking algorithm amplifies emotionally charged, out-group hostile content that users say makes them feel worse" [bsky, 1 points, 0 comments]
- Danke. Hab inzwischen mehrere Studien gefunden, welche sich mit dem Thema befasst haben. Unter anderem: arxiv.org/abs/2305.16941 epjdatascience.springeropen.com/articles/10.... [bsky, 1 points, 0 comments]
- Ce qu'on a échoué à faire en revanche, c'est fournir des règlementations sur la création et mise à disposition de ces outils. Instagram est devenu by design un endroit parfait pour attraper des comple [bsky, 0 points, 1 comments]
- Här är tex en studie som visar hur algoritmer i sociala medier späder på politiska utgruppsaversioner arxiv.org/abs/2305.16941 [bsky, 0 points, 0 comments]
Related