Within Popularity
Why Trending Lists Can Distort Judgment
Recommendation platforms can make already-popular items more visible, turning exposure into a feedback loop that feels like proof.
On this page
- How popularity bias changes what users see
- Why visibility is not the same as quality
- How to read rankings, likes, and recommendations cautiously
Page outline Jump by section
Introduction
Popularity bias in recommender systems is a modern form of appeal to popularity. Instead of a person explicitly arguing that something is true because many people believe it, a digital platform can make popular items appear more frequently, creating the impression that widespread attention is evidence of value, quality, relevance, or correctness. Trending lists, recommendation feeds, bestseller rankings, view counts, likes, and engagement metrics can all contribute to this effect.
The key issue is not that popular content is necessarily bad. Many popular items are genuinely useful, entertaining, or accurate. The problem arises when visibility itself becomes a source of further visibility. Once an item gains an early advantage, recommendation systems may repeatedly expose it to more users, generating a feedback loop in which popularity grows partly because the item was already popular. Researchers describe this as popularity bias, a common tendency of recommendation algorithms to favour already well-known items over less visible alternatives.[Springer Link]link.springer.comSpringer LinkA survey on popularity bias in recommender systemsby A Klimashevskaia · 2024 · Cited by 156 — In this paper, we discuss the…
How Popularity Bias Changes What Users See
Most recommendation systems learn from user behaviour. They observe clicks, views, purchases, listening habits, watch time, ratings, and other signals. Because popular items generate more interactions, they produce more data. Algorithms often treat this larger amount of data as a stronger indication that the item is worth recommending.[Milvus]milvus.ioWhat is popularity bias and how can it be mitigated in…Popularity bias occurs when recommendation systems disproportionately sug…
This creates a sequence that is deceptively simple:
- An item gains initial attention.
- The platform records many interactions with it.
- The recommendation system interprets those interactions as evidence of relevance.
- The item is shown to more users.
- Additional users interact with it.
- The item becomes even more prominent.
Researchers studying recommendation systems have repeatedly identified this cycle as a feedback loop that can amplify popularity over time. Simulations of user–algorithm interactions show that repeated recommendation cycles can increase popularity concentration while reducing diversity in what users encounter.[arXiv]arxiv.orgarXiv Feedback Loop and Bias Amplification in RecommenderFeedback Loop and Bias Amplification in Recommender…July 25, 2020 — by M Mansoury · 2020 · Cited by 398 — In this paper, we propo…[ACM]dl.acm.orgACM Digital LibraryFeedback Loop and Bias Amplification in Recommender…19 Oct 2020 — In this paper, we propose a method for simulating…
From the user’s perspective, the resulting feed may look like a neutral reflection of collective judgement. In reality, it is partly the product of a system that continuously converts existing attention into future attention.
Why Early Advantages Matter
Popularity bias means that small initial differences can become large outcomes.[youtube.com]youtube.comPopularity Bias In Recommender SystemsOvercoming Biases for a Better Recommender System: How Tech-Titans Combat…
A song that receives slightly more early engagement, a video that gains momentum during its first hours online, or a product that appears in a featured position can acquire an advantage that recommendation algorithms repeatedly reinforce. Over time, this process can create a highly unequal distribution of visibility, even when many alternatives are similarly relevant to users.[Springer Link]link.springer.comSpringer LinkA survey on popularity bias in recommender systemsby A Klimashevskaia · 2024 · Cited by 156 — In this paper, we discuss the…
The resulting impression is often that the most visible items naturally rose to the top because they were the best. Yet visibility and quality are not identical. Visibility itself can become part of the cause of success.
Why Visibility Is Not the Same as Quality
Popularity signals are useful shortcuts. If thousands of people buy a product or watch a video, that information may be worth considering. The logical problem emerges when popularity is treated as proof rather than as one clue among many.
Recommendation systems can unintentionally blur this distinction. Users often encounter content after it has already been filtered and ranked by algorithms. Because the highest-ranked items appear first, people may infer that those items deserve their position.
However, recommendation systems are usually optimised for objectives such as engagement, retention, click-through rates, or viewing time rather than objective quality. A highly engaging item may receive greater exposure regardless of whether it is more informative, more accurate, more creative, or more useful than alternatives.[Knight First Amendment Institute]knightcolumbia.orgKnight First Amendment InstituteA Public Service Media Perspective on the Algorithmic…24 Jul 2024 — Recommender systems play an import…
This matters because appeal-to-popularity reasoning becomes embedded in the environment itself. Rather than hearing someone say, “Everyone likes this, therefore it must be good,” users repeatedly see evidence that many others have already interacted with the same item. The social signal is presented before independent evaluation occurs.
The Hidden Assumption
Popularity bias encourages a subtle but important assumption:[youtube.com]youtube.comPopularity Bias In Recommender SystemsOvercoming Biases for a Better Recommender System: How Tech-Titans Combat…
Frequently recommended = widely chosen = probably best.
Each step may contain some truth, but the chain is not logically guaranteed. A recommendation system can increase exposure independently of quality. Once exposure changes, user behaviour changes as well. Researchers studying digital platforms note that recommendation systems do not merely observe preferences; they can also influence what users encounter and ultimately choose.[Nature]nature.comAlgorithmic Influence on Social Media Content and User…Social media platforms increasingly rely on algorithmic systems to curate…[Medium]medium.comPersonalization Algorithms and the Hidden Feedback Loop…Empirical research suggests that recommender systems can influence user…
What Happens to Less Popular Content?
One consequence of popularity bias is the underrepresentation of “long-tail” content—the vast number of items that receive relatively little attention individually.
Research in music, film, and other recommendation domains has consistently found that less popular items appear less frequently in recommendations than highly popular ones. As a result, niche creators, specialised products, and minority interests can struggle to gain visibility even when they may be highly relevant to particular users.[PMC]pmc.ncbi.nlm.nih.govThe Unfairness of Popularity Bias in Music Recommendationby D Kowald · 2020 · Cited by 234 — Research has shown that recommender syste…[arXiv]arxiv.orgarXiv Feedback Loop and Bias Amplification in RecommenderFeedback Loop and Bias Amplification in Recommender…July 25, 2020 — by M Mansoury · 2020 · Cited by 398 — In this paper, we propo…
This creates several effects:
- Users encounter fewer unexpected discoveries.
- New creators face greater barriers to attention.
- Markets become more concentrated around already-successful items.
- Recommendations become more homogeneous across users.
Studies of recommendation feedback loops have found that popularity amplification can reduce aggregate diversity and make user experiences more similar over time.[arXiv]arxiv.orgarXiv Feedback Loop and Bias Amplification in RecommenderFeedback Loop and Bias Amplification in Recommender…July 25, 2020 — by M Mansoury · 2020 · Cited by 398 — In this paper, we propo…
From the perspective of logical fallacies, this matters because the system increasingly presents consensus as evidence. The more visible an item becomes, the more users interpret its visibility as confirmation that it deserves attention.
When Popularity Signals Shape Belief
Popularity bias becomes especially important when recommendation systems distribute information rather than entertainment alone.
News feeds, social platforms, search suggestions, and video recommendations often display engagement indicators such as views, shares, likes, or trending status. These signals can influence how people evaluate credibility before examining the underlying evidence. Researchers examining algorithmic recommendation systems note that engagement-based amplification can strengthen feedback loops that shape attention and information exposure. ScienceDirect[2ojs.weizenbaum-institut.de]ojs.weizenbaum-institut.deReaders click on news articles selected by a recommender system and, in doing so…Read more…
In such environments, popularity may begin to function as a substitute for independent verification. A claim appears persuasive because it is repeatedly encountered, heavily shared, or prominently recommended.
The logical danger mirrors the traditional appeal to popularity fallacy:
- Many people appear interested in a claim.
- The platform promotes the claim because of that interest.
- Users interpret the promotion as evidence of merit.
- The claim gains further attention.
At no point does widespread exposure itself establish truth.
How to Read Rankings, Likes, and Recommendations Cautiously
Popularity signals are not worthless. They often contain useful information about what other people have found interesting or valuable. The challenge is interpreting them appropriately.
Several habits can reduce the risk of mistaking popularity for evidence:
Treat rankings as indicators of attention, not proof.
A top-ranked item shows what has attracted engagement. It does not automatically show what is most accurate, highest quality, or best suited to your needs.
Separate exposure from merit.
Ask whether an item became popular because of its qualities, because of marketing, because of timing, or because an algorithm repeatedly increased its visibility.
Look beyond the first recommendations.
Many systems heavily concentrate attention on a small set of already-successful items. Exploring beyond the top results can reveal alternatives that the popularity feedback loop suppresses.
Be cautious with social proof metrics.
Large numbers of likes, views, or downloads may indicate interest, but they do not independently verify factual claims.
Remember that recommendation systems optimise objectives.[researchgate.net]researchgate.netPDF) Exploring Popularity Bias in Music Recommendation…19 Aug 2022 — Research has shown that recommender systems are typically biased…
The platform may be maximising engagement, retention, advertising revenue, or user activity rather than truth, expertise, or quality.[Knight First Amendment Institute]knightcolumbia.orgKnight First Amendment InstituteA Public Service Media Perspective on the Algorithmic…24 Jul 2024 — Recommender systems play an import…
Why Popularity Bias Matters for Understanding Appeal to Popularity
Popularity bias in recommender systems demonstrates how a classic logical fallacy can become embedded in technological systems rather than expressed directly in argument. The system does not need to claim that popularity proves value. Instead, it repeatedly exposes users to signals of popularity and structures attention around them.
As a result, users may encounter a world in which what is already popular becomes increasingly visible, increasingly familiar, and increasingly persuasive. The central lesson remains the same as in traditional appeal-to-popularity reasoning: widespread attention can be informative, but it is not evidence that a claim is true, a product is best, or a piece of content is most deserving of belief. Visibility measures attention; it does not by itself measure merit.[Springer Link]link.springer.comSpringer LinkA survey on popularity bias in recommender systemsby A Klimashevskaia · 2024 · Cited by 156 — In this paper, we discuss the… 2arXiv
Amazon book picks
Further Reading
Books and field guides related to Why Trending Lists Can Distort Judgment. Use these as the next step if you want deeper reading beyond the article.
Weapons of Math Destruction
Explains unintended consequences of algorithmic decision systems.
Algorithms of Oppression
Examines how platforms and algorithms shape visibility and attention.
eBay marketplace picks
Marketplace Samples
Example marketplace items related to this page. Use the search link to explore similar finds on eBay.
Endnotes
1.
Source: link.springer.com
Link:https://link.springer.com/article/10.1007/s11257-024-09406-0
Source snippet
Springer LinkA survey on popularity bias in recommender systemsby A Klimashevskaia · 2024 · Cited by 156 — In this paper, we discuss the...
2.
Source: pmc.ncbi.nlm.nih.gov
Link:https://pmc.ncbi.nlm.nih.gov/articles/PMC7148048/
Source snippet
The Unfairness of Popularity Bias in Music Recommendationby D Kowald · 2020 · Cited by 234 — Research has shown that recommender syste...
3.
Source: milvus.io
Link:https://milvus.io/ai-quick-reference/what-is-popularity-bias-and-how-can-it-be-mitigated-in-recommendations
Source snippet
What is popularity bias and how can it be mitigated in...Popularity bias occurs when recommendation systems disproportionately sug...
4.
Source: arxiv.org
Title: arXiv Feedback Loop and Bias Amplification in Recommender
Link:https://arxiv.org/abs/2007.13019
Source snippet
Feedback Loop and Bias Amplification in Recommender...July 25, 2020 — by M Mansoury · 2020 · Cited by 398 — In this paper, we propo...
Published: July 25, 2020
5.
Source: dl.acm.org
Link:https://dl.acm.org/doi/10.1145/3340531.3412152
Source snippet
ACM Digital LibraryFeedback Loop and Bias Amplification in Recommender...19 Oct 2020 — In this paper, we propose a method for simulating...
6.
Source: nature.com
Link:https://www.nature.com/nature-index/topics/l4/algorithmic-influence-on-social-media-content-and-user-behavior
Source snippet
Algorithmic Influence on [Social Media]({{ 'social-media/' | relative_url }}) Content and User...Social media platforms increasingly rely on algorithmic systems to curate...
7.
Source: medium.com
Link:https://medium.com/brain-mind-and-intelligence/personalization-algorithms-and-the-hidden-feedback-loop-shaping-your-reality-c03ab687f9ca
Source snippet
Personalization Algorithms and the Hidden Feedback Loop...Empirical research suggests that recommender systems can influence user...
8.
Source: pmc.ncbi.nlm.nih.gov
Link:https://pmc.ncbi.nlm.nih.gov/articles/PMC11373151/
Source snippet
Drivers and Algorithmic [Mechanisms]({{ 'mechanisms/' | relative_url }}) on Digital Mediaby H Metzler · 2023 · Cited by 240 — Algorithmic mechanisms on digital media are power...
9.
Source: arxiv.org
Link:https://arxiv.org/abs/1912.04696
Source snippet
The Unfairness of Popularity Bias in Music Recommendationby D Kowald · 2019 · Cited by 234 — Research has shown that recommender systems...
10.
Source: sciencedirect.com
Link:https://www.sciencedirect.com/science/article/pii/S0047272726000253
Source snippet
Ranking for engagement: How social media algorithms fuel...by F Germano · 2026 · Cited by 11 — This paper investigates the...
11.
Source: ojs.weizenbaum-institut.de
Link:https://ojs.weizenbaum-institut.de/index.php/wjds/article/view/4_2_4/132
Source snippet
Readers click on news articles selected by a recommender system and, in doing so...Read more...
12.
Source: sciencedirect.com
Link:https://www.sciencedirect.com/org/science/article/pii/S1548367326000062
Source snippet
Integrating user behavior, content...Read more...
13.
Source: medium.com
Link:https://medium.com/%40kamalmeet/popularity-bias-in-recommendation-engines-2542d1cdb353
Source snippet
Popularity Bias in Recommendation EnginesPopularity bias refers to a recommender system's tendency to over-recommend items that are alrea...
14.
Source: medium.com
Link:https://medium.com/%40emreeyukseel/challenges-in-recommender-systems-understanding-implicit-feedback-negative-sampling-and-00fc3c9c34e4
Source snippet
tes a popularity bias which is a common problem that must...Read more...
15.
Source: arxiv.org
Link:https://arxiv.org/pdf/1912.04696
Source snippet
Research has shown that recommender systems are typically biased towards popular items, which leads to less popular items being un...Rea...
16.
Source: arxiv.org
Link:https://arxiv.org/pdf/2208.09517
Source snippet
2208.09517v1 [cs.IR] 19 Aug 2022by DR Turnbull · 2022 · Cited by 14 — In this paper, we attempt to measure popularity bias both in...
17.
Source: arxiv.org
Link:https://arxiv.org/abs/2308.01118
Source snippet
A Survey on Popularity Bias in Recommender Systemsby A Klimashevskaia · 2023 · Cited by 156 — In this paper, we discuss the potential rea...
18.
Source: link.springer.com
Link:https://link.springer.com/article/10.1007/s10844-026-01025-y
Source snippet
fairness, popularity bias, and user group disparitiesby Y Zoralioglu · 2026 — In this study, we evaluate how users are differently affect...
19.
Source: link.springer.com
Link:https://link.springer.com/book/10.1007/978-3-642-13287-2
Source snippet
The Long Tail, Long Fail, and Long Play in the Digital Music Space... Order, [context]({{ 'context/' | relative_url }}) and popularity bias in next-song...Read more...
20.
Source: sciencedirect.com
Link:https://www.sciencedirect.com/science/article/abs/pii/S0306457323003436
Source snippet
Flexibly manipulating popularity bias for tackling trade-offs...by H Okamura · 2024 · Cited by 8 — We can effectively and flexibly manip...
21.
Source: sciencedirect.com
Link:https://www.sciencedirect.com/science/article/abs/pii/S0306457321001436
Source snippet
Investigating the impact of recommender systems on user-...by M Elahi · 2021 · Cited by 81 — In this paper, we study the impact of diffe...
22.
Source: youtube.com
Title: Popularity Bias In Recommender Systems
Link:https://www.youtube.com/watch?v=D4Us–bvFQo
Source snippet
Overcoming Biases for a Better Recommender System: How Tech-Titans Combat...
23.
Source: knightcolumbia.org
Link:https://knightcolumbia.org/content/a-public-service-media-perspective-on-the-algorithmic-amplification-of-cultural-content
Source snippet
Knight First Amendment InstituteA Public Service Media Perspective on the Algorithmic...24 Jul 2024 — Recommender systems play an import...
24.
Source: knightcolumbia.org
Link:https://knightcolumbia.org/content/understanding-social-media-recommendation-algorithms
Source snippet
Knight First Amendment InstituteUnderstanding Social Media Recommendation AlgorithmsMar 9, 2023 — The algorithms driving social media are...
25.
Source: Wikipedia
Title: Recommender system
Link:https://en.wikipedia.org/wiki/Recommender_system
Source snippet
Recommender systemA recommender system, also called a recommendation algorithm, recommendation engine, or recommendation platform, is...
26.
Source: iss.nus.edu.sg
Title: nus.edu.sg Recommender Systems
Link:https://www.iss.nus.edu.sg/executive-education/course/detail/recommender–systems/data-science
Source snippet
Systems - SingaporeLearn to design AI-driven recommender systems using machine learning and Python. This NUS-ISS course covers collaborat...
Additional References
27.
Source: researchgate.net
Link:https://www.researchgate.net/publication/362860235_Exploring_Popularity_Bias_in_Music_Recommendation_Models_and_Commercial_Steaming_Services
Source snippet
(PDF) Exploring Popularity Bias in Music Recommendation...19 Aug 2022 — Research has shown that recommender systems are typically biased...
28.
Source: blogs.biomedcentral.com
Link:https://blogs.biomedcentral.com/on-physicalsciences/2021/04/13/algorithm-generated-music-recommendations-low-accuracy-for-fans-of-beyond-mainstream-music/
Source snippet
Accuracy for Fans of Beyond-Mainstream Music13 Apr 2021 — However, it is a widely-known problem that recommender systems are prone to pop...
29.
Source: ftp.informatik.rwth-aachen.de
Link:https://ftp.informatik.rwth-aachen.de/Publications/CEUR-WS/Vol-633/wom2010_paper10.pdf
Source snippet
Recommendation and the Long Tailby M Levy · Cited by 85 — In this paper we present an empirical study of the rec- ommendations actually m...
30.
Source: GOV.UK
Link:https://www.gov.uk/government/publications/research-into-the-impact-of-streaming-services-algorithms-on-music-consumption/the-impact-of-algorithmically-driven-recommendation-systems-on-music-consumption-and-production-a-literature-review
Source snippet
impact of algorithmically driven recommendation...by D Hesmondhalgh · Cited by 58 — Their discussion of the sources of popularity bias i...
31.
Source: christophtrattner.com
Link:https://www.christophtrattner.com/pubs/PopBias2024.pdf
Source snippet
A Survey on Popularity Bias in Recommender Systemsby A Klimashevskaia · Cited by 179 — In this paper, we discuss the potential reasons fo...
32.
Source: researchgate.net
Title: 343253269 Feedback Loop and Bias Amplification in Recommender Systems
Link:https://www.researchgate.net/publication/343253269_Feedback_Loop_and_Bias_Amplification_in_Recommender_Systems
Source snippet
Feedback Loop and Bias Amplification in Recommender...31 Jul 2020 — In this paper, we propose a method for simulating the users interact...
33.
Source: orfonline.org
Title: from clicks to chaos how social media algorithms amplify extremism
Link:https://www.orfonline.org/expert-speak/from-clicks-to-chaos-how-social-media-algorithms-amplify-extremism
Source snippet
From clicks to chaos: How social media algorithms amplify...Feb 13, 2025 — This article explores the mechanisms behind algorithmic ampli...
34.
Source: ceur-ws.org
Link:https://ceur-ws.org/Vol-633/wom2010_paper10.pdf
Source snippet
ally made by the widely-used Last.fm music recommender system, in particular via its...Read more...
35.
Source: liner.com
Title: feedback loop and bias amplification in recommender systems
Link:https://liner.com/review/feedback-loop-and-bias-amplification-in-recommender-systems
Source snippet
Feedback Loop and Bias Amplification in Recommender...Jul 25, 2020 — Regarding this CIKM 2020 paper, this review summarizes how feedback...
36.
Source: scispace.com
Link:https://scispace.com/papers/feedback-loop-and-bias-amplification-in-recommender-systems-43dh6ke8pd
Source snippet
ecommenders in an offline setting and study the impact of feedback loop on the...Read more...
Topic Tree