A systematic review of echo chamber research: comparative analysis of conceptualizations, operationalizations, and varying outcomes

May combine two claims.

Bail et al. challenge the polarization assumption, showing that social media use can increase exposure to non-partisan content.

“Supporting this, Bail et al. [2] further challenge the polarization assumption, showing that social media use can increase exposure to non-partisan content.”

• cites [2] Bail, C. A., Argyle, L. P., Brown, T. W., Bumpus, J. P., Chen, H., Hunzaker, M. B. F., Lee, J., Mann, M., Merhout, F., & Volfovsky, A. (2018). Exposure to opposing views on social media can increase political polarization. Proceedings of the National Academy of Sciences of the United States of America, 115(37), 9216–9221. sourceDisputed

  The Research Design section says the intervention exposed Twitter users to messages from political accounts with opposing ideologies, not non-partisan content. The Abstract reports that Republicans became substantially more conservative after following a liberal bot; the paper does not show that ordinary social media use increases exposure to non-partisan content. Page numbers are unavailable in our copy. (LLM, conf=0.89)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The amount of media a person consumes is anti-proportional to their chance of being trapped in an echo chamber, which contradicts findings by CSS studies that engagement and echo chambers are positively correlated.

“They find that the amount of media a person consumes is anti-proportional to their chance of being trapped in an echo chamber, which contradicts findings by CSS studies [48,50,66,67] that find that engagement and echo chambers are positively correlated.”

• cites [48] Del Vicario, M., Bessi, A., Zollo, F., Petroni, F., Scala, A., Caldarelli, G., Stanley, H. E., & Quattrociocchi, W. (2016). The spreading of misinformation online. Proceedings of the National Academy of Sciences of the United States of America, 113(3), 554–559. sourcePartly supported

  The citation is attached to the claim that engagement and echo chambers are positively correlated, not to the preceding claim about an inverse relationship. Results and Discussion reports that sharing largely occurs within homogeneous clusters, supporting a qualitative link between engagement and echo chambers. It does not establish that people who consume more media are more likely to be in an echo chamber or report the stated correlation. (LLM, conf=0.78)

  Experimental — accuracy not yet measured
• cites [50] Zollo, F., Bessi, A., Del Vicario, M., Scala, A., Caldarelli, G., Shekhtman, L., Havlin, S., & Quattrociocchi, W. (2017). Debunking in a world of tribes. PLOS ONE 12(7). sourcePartly supported

  The Results and discussion section (‘Echo chambers’; page unavailable) finds that Facebook users are highly polarized and mainly interact with one type of content, supporting a link between engagement and echo chambers. It does not show that greater amounts of engagement are correlated with a greater likelihood of being in an echo chamber. The inverse relationship is attributed to a different study in the citing sentence, not to this source. (LLM, conf=0.82)

  Experimental — accuracy not yet measured
• cites [66] Schmidt, A. L., Peruzzi, A., Scala, A., Cinelli, M., Pomerantsev, P., Applebaum, A., Gaston, S., Fusi, N., Peterson, Z., Severgnini, G., Cesco, A. F., Casati, D., Novak, P. K., Stanley, H. E., Zollo, F., & Quattrociocchi, W. (2020). Measuring social response to different journalistic techniques on facebook. Humanities & Social Sciences Communications, 7(1), 17. sourcePartly supported

  The Introduction says users continue selecting belief-consistent content as they enter echo chambers, supporting a qualitative link between content selection and echo-chamber behavior. The Results examine news framing and commenting, not whether greater engagement or media consumption is positively correlated with being in an echo chamber. The inverse relationship is attributed to a different study in the citing sentence, not to this source. (LLM, conf=0.76)

  Experimental — accuracy not yet measured
• cites [67] Koivula, A., Kaakinen, M., Oksanen, A., & Rasanen, P. (2019). The role of political activity in the formation of online identity bubbles. Policy and Internet, 11(4), 396–417. sourceDisputed

  The Discussion (page not available in our copy) says frequent social media use has an effect on identity-bubble formation, and the Results report a positive association involving social media activity and identity-bubble reinforcement. The paper does not find that consuming more media lowers a person's chance of being in an echo chamber; it studies social media activity and online identity bubbles rather than media consumption generally. (LLM, conf=0.86)

  Checked by a reasoning model — may be wrong.

  Experimental — accuracy not yet measured

May combine two claims.

There is conflicting evidence on the existence and effects of echo chambers.

“However, there is conflicting evidence on the existence and effects of echo chambers.”

• cites [15] Terren, L., & Borge, R. (2021). Echo chambers on social media: A systematic review of the literature. Review of Communication Research, 9, 99–118. https://doi.org/10.12840/ISSN.2255-4165.028 sourcePartly supported

  The Highlights (p. 100) report studies finding clear evidence of social-media echo chambers and others finding none, supporting conflicting evidence about their existence. The Introduction (p. 100) describes competing views of their potential consequences, but the passages do not establish conflicting empirical evidence about their effects. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites [21] Arguedas, A. R., Robertson, C., Fletcher, R., & Nielsen, R. K. (2022). Echo chambers, filter bubbles, and polarisation: A literature review. Reuter Institute for the Study of Journalism. Accessed 10 March 2022.Not checked yet

  Not checked in the free run.

May combine two claims.

Users often consume partisan content, reinforcing existing beliefs.

“Users often consume partisan content, reinforcing existing beliefs [24,54].”

• cites [24] Kitchens, B., Johnson, S. L., & Gray, P. (2020). Understanding echo chambers and filter bubbles: The impact of social media on diversification and partisan shifts in news consumption. MIS Quarterly, 44(4), 1619–1649. sourceNot checked — no full text of the source
• cites [54] Nikolov, D., Lalmas, M., Flammini, A., & Menczer, F. (2019). Quantifying biases in online information exposure. Journal of the Association for Information Science and Technology, 70(3), 218–229. https://doi.org/10.1002/asi.24121 sourcePartly supported

  The Background (p. 5) reports that online users are somewhat more likely to encounter news aligned with their worldviews, supporting the selective-exposure part of the claim. But the Discussion (p. 15) says this study did not analyze content or the political bias of news sources, and it does not establish that consuming such content reinforces users’ beliefs. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

May combine two claims.

Attitude-based homophily is particularly pronounced in segregated networks, such as scientific vs. conspiracy communities, and among groups divided by political affiliation, such as Democrats and Republicans.

“This phenomenon is particularly pronounced in segregated networks, such as”

• cites [47] Bessi, A., Zollo, F., Del Vicario, M., Puliga, M., Scala, A., Caldarelli, G., Uzzi, B., & Quattrociocchi, W. (2016). Users polarization on facebook and youtube. PLOS ONE 11(8). sourcePartly supported

  The Results and Discussion section (“Polarized and Homogeneous Communities”; page unavailable) reports that users concentrated on either science or conspiracy content and formed well-separated echo chambers. The paper does not examine segregation by political affiliation or compare Democrats and Republicans, so it supports only the science-versus-conspiracy example. (LLM, conf=0.86)

  Experimental — accuracy not yet measured
• cites [48] Del Vicario, M., Bessi, A., Zollo, F., Petroni, F., Scala, A., Caldarelli, G., Stanley, H. E., & Quattrociocchi, W. (2016). The spreading of misinformation online. Proceedings of the National Academy of Sciences of the United States of America, 113(3), 554–559. sourcePartly supported

  The Results and Discussion section reports that sharing links are predominantly homogeneous and that science and conspiracy content circulate in distinct echo chambers. It does not report a comparison between Democrats and Republicans or establish the claim for groups divided by political affiliation. (LLM, conf=0.87)

  Experimental — accuracy not yet measured
• cites [49] Schmidt, A. L., Zollo, F., Scala, A., Betsch, C., & Quattrociocchi, W. (2018). Polarization of the vaccination debate on facebook. Vaccine, 36(25), 3606–3612. https://doi.org/10.1016/j.vaccine.2018.05.040 sourcePartly supported

  The Introduction (p. 2) describes users selecting belief-consistent information and forming polarized groups with opposing views that rarely interact. It does not establish that attitude-based homophily is particularly pronounced in such networks, or discuss scientific versus conspiracy communities or Democrats versus Republicans as the claimed examples. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites [50] Zollo, F., Bessi, A., Del Vicario, M., Scala, A., Caldarelli, G., Shekhtman, L., Havlin, S., & Quattrociocchi, W. (2017). Debunking in a world of tribes. PLOS ONE 12(7). sourcePartly supported

  The Conclusions (page not provided) describe highly segregated Facebook communities centered on scientific and conspiracy topics, and the Echo chambers section reports that users are polarized toward one or the other. The paper does not address homophily between Democrats and Republicans, so the political-affiliation example is not supported by this source. (LLM, conf=0.88)

  Experimental — accuracy not yet measured

May combine two claims.

Studies examining major political events, including the 2015 Polish elections, the impeachment of Brazilian President Dilma Rousseff, and the 2016 U.S. elections, support the echo chamber hypothesis, showing pronounced clustering in user interactions.

“Studies examining major political events, including the 2015 Polish elections [51], the impeachment of Brazilian President Dilma Rousseff [52], and the 2016 U.S. elections [53], also support the echo chamber hypothesis, showing pronounced clustering in user interactions.”

• cites [51] Batorski, D., & Grzywinska, I. (2018). Three dimensions of the public sphere on facebook. Information, Communication & Society, 21(3), 356–374. https://doi.org/10.1080/1369118X.2017.1281329 sourceNot checked — no full text of the source
• cites [52] Cota, W., Ferreira, S. C., Pastor-Satorras, R., & Starnini, M. (2019). Quantifying echo chamber effects in information spreading over political communication networks. EPJ Data Science 8(1). sourceSupported
• cites [53] Guo, L., Rohde, J. A., & Wu, H. D. (2020). Who is responsible for twitter’s echo chamber problem? Evidence from 2016 us election networks. Information, Communication & Society, 23(2), 234–251. https://doi.org/10.1080/1369118X.2018.1499793 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

The review follows the PRISMA 2020 guideline.

“We follow the PRISMA 2020 [26] guideline to present a transparent account of how we conducted our systematic review.”

• cites [26] Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., … Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. Systematic Reviews, 18(3), 1003583. sourceSupported

May combine two claims.

Differences in datasets used across studies have led to inconsistent results, with much of the evidence relying on data collected in 2020.

“Differences in datasets used across studies have led to inconsistent results, with much of the evidence relying on data collected in 2020.”

• cites [15] Terren, L., & Borge, R. (2021). Echo chambers on social media: A systematic review of the literature. Review of Communication Research, 9, 99–118. https://doi.org/10.12840/ISSN.2255-4165.028 sourcePartly supported

  The Highlights (before p. 100) link differing findings to the type of data used: studies finding clear evidence of echo chambers used digital trace data, while those finding none used self-reported data. They describe the review as covering the state of the literature at January 2020, not as relying largely on data collected in 2020. The data-collection-year claim is unsupported. (LLM, conf=0.87)

  Experimental — accuracy not yet measured

May combine two claims.

Gao et al. analyzed platforms such as Douyin, TikTok, and Bilibili, demonstrating how feed algorithms foster selective exposure and homophily, contributing to echo chamber formation, while cultural differences can correlate with the development of echo chambers.

“However, a study by Gao et al. [64] analyzed platforms such as Douyin, TikTok, and Bilibili, demonstrating how feed algorithms foster selective exposure and homophily, ultimately contributing to echo chamber formation, while also highlighting that cultural differences can correlate with the development of echo chambers and their influence is not equivalently likely for each community and culture.”

• cites [64] Gao, Y., Liu, F., & Gao, L. (2023). Echo chamber effects on short video platforms. Scientific Reports, 13, 6282. https://doi.org/10.1038/s41598-023-33370-1 sourcePartly supported

  Split into 2 atoms — 1 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Users of social media platforms can choose to be exposed to conversations with like-minded users and content that reflects their thinking, thus reinforcing their existing preferences or beliefs.

“Users of social media platforms can choose to be exposed to conversations with like-minded users and content that reflects their thinking, thus reinforcing their existing preferences or beliefs [11,12].”

• cites [11] Van Aelst, P., Strömbäck, J., Aalberg, T., Esser, F., De Vreese, C., Matthes, J., Hopmann, D., Salgado, S., Hubé, N., & Stępińska, A. (2017). Political communication in a high-choice media environment: a challenge for democracy? Annals of the International Communication Association 41(1), 3–27. sourceNot checked — no full text of the source
• cites [12] Dubois, E., & Blank, G. (2018). The echo chamber is overstated: the moderating effect of political interest and diverse media. Information Communication & Society 21(5, SI), 729–745 https://doi.org/10.1080/1369118X.2018.1428656 sourceSupported

May combine two claims.

Offline political success correlates with detachment (isolation of political actors) in online networks, suggesting that online fragmentation is not solely a product of digital interactions but also reflects offline political dynamics.

“Bright [89] found that offline political success correlates with detachment (isolation of political actors) in online networks, suggesting that online fragmentation is not solely a product of digital interactions but also reflects offline political dynamics.”

• cites [89] Bright, J. (2018). Explaining the emergence of political fragmentation on social media: The role of ideology and extremism. Journal of Computer-Mediated Communication, 23(1), 17–33. sourceSupported

May combine two claims.

Younger and less educated users are more susceptible to personalized content, potentially limiting exposure to diverse viewpoints.

“Bodo et al. [82] highlighted that younger and less educated users are more susceptible to personalized content, potentially limiting exposure to diverse viewpoints.”

• cites [82] Bodo, B., Helberger, N., Eskens, S., & Moller, J. (2019). Interested in diversity: The role of user attitudes, algorithmic feedback loops, and policy in news personalization. Digital Jorunalism, 7(2), 206–229. https://doi.org/10.1080/21670811.2018.1521292 sourceSupported

May combine two claims.

Specific hub nodes play a pivotal role in spreading misinformation within networks and shaping echo chambers.

“Specific hub nodes play a pivotal role in spreading misinformation within networks and shaping echo chambers [99].”

• cites [99] Asatani, K., Yamano, H., Sakaki, T., & Sakata, I. (2021). Dense and influential core promotion of daily viral information spread in political echo chambers. Scientific Reports 11(1). sourcePartly supported

  The Results (pp. unavailable) show that top influencers spread large amounts of information within political echo chambers and suggest that their connected networks may contribute to opinion homogeneity. The paper does not establish that the information these hubs spread is misinformation; its Introduction distinguishes its study of daily information spread from earlier studies focused on fake news or rumours. (LLM, conf=0.87)

  Experimental — accuracy not yet measured

May combine two claims.

Nordbrandt and Nyhan et al. observed that while social media usage does not directly influence affective polarization, prior patterns of social media engagement can shape this relationship.

“For instance, Nordbrandt [76], Nyhan et al. [77] observed that while social media usage does not directly influence affective polarization-the individuals' sympathy or antipathy toward specific political parties-prior patterns of social media engagement can shape this relationship.”

• cites [76] Nordbrandt, M. (2019). Affective polarization in the digital age: Testing the direction of the relationship between social media and users’ feelings for out-group parties. New Media & Society. https://doi.org/10.1177/14614448211044393 sourcePartly supported

  The Conclusion and discussion reports that starting or increasing social media use did not affect affective polarization over time. It also says prior usage conditioned the reverse relationship—how affective polarization affected later social media use—not an effect of social media use on polarization. The citing sentence blurs that direction. (LLM, conf=0.88)

  Experimental — accuracy not yet measured
• cites [77] Nyhan, B., Settle, J., Thorson, E., et al. (2023). Like-minded sources on facebook are prevalent but not polarizing. Nature, 620, 137–144. https://doi.org/10.1038/s41586-023-06297-w sourcePartly supported

  The Treatment effects on attitudes section reports no measurable effect on affective polarization from reducing exposure to like-minded Facebook sources. That is narrower than saying social media usage does not directly influence affective polarization. The Discussion says the experiment cannot capture effects of prior Facebook use; it does not establish that prior engagement patterns shape the relationship. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

May combine two claims.

User attributes linked to echo chambers include, besides selective exposure, personality traits, fear of isolation and anger, openness, and reflectiveness.

“These user attributes that are linked to echo chambers include-besides selective exposurepersonality traits (e.g., [16]), fear of isolation and anger (e.g, [104]), openness (e.g., [84]), and reflectiveness (e.g., [93]).”

• cites [16] Boulianne, S., Koc-Michalska, K., & Bimber, B. (2020). Right-wing populism, social media and echo chambers in western democracies. New Media & Society, 22(4), 683–699. https://doi.org/10.1177/1461444819893983 sourceNot checked — no full text of the source
• cites [104] Wollebaek, D., Karlsen, R., Steen-Johnsen, K., & Enjolras, B. (2019). Anger, fear, and echo chambers: The emotional basis for online behavior. Social Media + Society 5(2). https://doi.org/10.1177/2056305119829859 sourcePartly supported

  The Discussion and Conclusion says anger can reinforce echo chambers through searches for confirming information. It says fear counteracts that dynamic and reports no relationship between fear and participation in online debates. The cited paper therefore supports the anger portion of [104], but not the claimed link between fear of isolation and echo chambers; the other attributes have separate citations. (LLM, conf=0.88)

  Experimental — accuracy not yet measured
• cites [84] Matz, S. C. (2021). Personal echo chambers: Openness-to-experience is linked to higher levels of psychological interest diversity in large-scale behavioral data. Journal of Personality and Social Psychology, 121(6), 1284–1300. sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Homophily is a key driver of content dissemination, and polarization within these communities exacerbates the spread of misinformation.

“The authors argue that homophily is a key driver of content dissemination and that polarization within these communities exacerbates the spread of misinformation.”

• cites [48] Del Vicario, M., Bessi, A., Zollo, F., Petroni, F., Scala, A., Caldarelli, G., Stanley, H. E., & Quattrociocchi, W. (2016). The spreading of misinformation online. Proceedings of the National Academy of Sciences of the United States of America, 113(3), 554–559. sourceSupported

May combine two claims.

Culture or demographics drive polarization and echo chambers.

“novel research [61,64,92] finds that culture or demographics drive polarization and echo chambers.”

• Also stated in: § 4.1.3.1 Polarization and fragmentation of the public sphere · plain

  “This finding aligns with findings by Monti et al. [61], Gao et al. [64], and Grusauskaite et al. [92] who find that culture or demographics drive polarization.”

• cites [61] Monti, C., D’Ignazi, J., Starnini, M., & Morales, G. D. F. (2023). Evidence of demographic rather than ideological segregation in news discussion on reddit. In: Proceedings of the ACM Web Conference 2023 (WWW ’23) (pp. 1–10). ACM, New York, NY, USA. https://doi.org/10.1145/3543507.3583468 sourcePartly supported

  The Introduction describes demographic factors as a possible explanation for affective polarization, and the Discussion says the findings support a link between a divided society and polarization. But the Discussion also reports a lack of ideological echo chambers in generalist Reddit news discussions; the paper does not find that culture or demographics drive echo chambers. (LLM, conf=0.87)

  Experimental — accuracy not yet measured
• cites [64] Gao, Y., Liu, F., & Gao, L. (2023). Echo chamber effects on short video platforms. Scientific Reports, 13, 6282. https://doi.org/10.1038/s41598-023-33370-1 sourcePartly supported

  The Introduction and Conclusion say cultural differences among platform users can hinder the development of echo chambers, so the source supports a cultural influence on echo chambers. It does not establish that culture drives polarization, and its Results describe members’ ages and other characteristics without showing that demographics drive either outcome. No page numbers are available in our copy. (LLM, conf=0.87)

  Experimental — accuracy not yet measured
• cites [92] Grusauskaite, K., Carbone, L., Harambam, J., & Aupers, S. (2024). Debating (in) echo chambers: How culture shapes communication in conspiracy theory networks on youtube. New Media & Society, 26(12), 7037–7057. https://doi.org/10.1177/14614448231162585 sourcePartly supported

  The Abstract and Conclusion and discussion say that conspiratorial subcultures help explain differing degrees of echo-chamberness across four sampled YouTube comment sections. They do not find that demographics drive echo chambers—the conclusion proposes that as a topic for future research—or that culture drives polarization. (LLM, conf=0.88)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

The first two hours of an information cascade are critical for developing opinion clusters.

“Notably, the first two hours of an information cascade are critical for developing opinion clusters.”

• cites [48] Del Vicario, M., Bessi, A., Zollo, F., Petroni, F., Scala, A., Caldarelli, G., Stanley, H. E., & Quattrociocchi, W. (2016). The spreading of misinformation online. Proceedings of the National Academy of Sciences of the United States of America, 113(3), 554–559. sourcePartly supported

  The Results and Discussion report that cascade lifetimes peak at about 1–2 hours, while the Abstract says selective exposure generates homogeneous clusters. Neither passage says the first two hours are critical for developing those clusters; the two-hour finding concerns cascade duration, not cluster formation. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

May combine two claims.

Dubois and Blank demonstrated that users with more diverse media consumption habits are less likely to become trapped in echo chambers, as varied exposure promotes openness to differing viewpoints, and politically engaged users are particularly resilient to echo chambers.

“Similarly, Dubois and Blank [12] demonstrated that users with more diverse media consumption habits are less likely to become trapped in echo chambers, as varied exposure to information promotes openness to differing viewpoints and, again, politically engaged users are particularly resilient to echo chambers, with their media choices further mitigating risks of ideological isolation.”

• cites [12] Dubois, E., & Blank, G. (2018). The echo chamber is overstated: the moderating effect of political interest and diverse media. Information Communication & Society 21(5, SI), 729–745 https://doi.org/10.1080/1369118X.2018.1428656 sourcePartly supported

  Split into 2 atoms — 1 confirmed, 1 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

A small number of highly connected users are responsible for a significant portion of misinformation spread, with 10% of hub communities generating 36% of retweets of false information.

“For example, Choi et al. [100] found that a small number of highly connected users are responsible for a significant portion of misinformation spread, with 10% of hub communities generating 36% of retweets of false information.”

• cites [100] Choi, D., Chun, S., Oh, H., Han, J., & Kwon, T. (2020). Rumor propagation is amplified by echo chambers in social media. Scientific Reports, 10(1), 310. sourcePartly supported

  The Introduction and Results report that the top 10% of highly connected rumor echo chambers contributed to 24% of rumor cascades and elicited more than 36% of retweets. The measured units were groups of users, not individual highly connected users, and the paper describes the content as rumors rather than exclusively verified false information. Page numbers are not provided in our copy. (LLM, conf=0.86)

  Experimental — accuracy not yet measured

May combine two claims.

Cross-cutting discussions can paradoxically heighten affective polarization through unfriending, with this effect exacerbated by exposure to incivility on social media.

“Similarly, Lin et al. [79] demonstrated that cross-cutting discussions can paradoxically heighten affective polarization through unfriending, with this effect exacerbated by exposure to incivility on social media.”

• cites [79] Lin, H., Wang, Y., Lee, J., & Kim, Y. (2023). The effects of disagreement and unfriending on political polarization: A moderated-mediation model of cross-cutting discussion on affective polarization via unfriending contingent upon exposure to incivility. Journal of Computer-Mediated Communication, 28(4), 022. https://doi.org/10.1093/jcmc/zmad022 sourceSupported

May combine two claims.

On Weibo, Wang et al. observed that retweeting mechanisms promote polarization, while commenting fosters consensus, highlighting the importance of interaction types in shaping polarization dynamics.

“On Weibo, Wang et al. [78] observed that retweeting mechanisms promote polarization, while commenting fosters consensus, highlighting the importance of interaction types in shaping polarization dynamics.”

• cites [78] Wang, X., Sirianni, A. D., Tang, S., Zheng, Z., & Fu, F. (2020). Public discourse and social network echo chambers driven by socio-cognitive biases. Physical Review X 10(4). https://doi.org/10.1103/PhysRevX.10.041042 sourcePartly supported

  The Results and model–data integration sections describe increasingly polarized retweet networks on Twitter during the 2016 U.S. election. A figure caption in Appendix E says that discussion-partner influence can reduce extreme clusters in simulations. The paper does not report observations on Weibo or an empirical comparison showing that commenting fosters consensus. (LLM, conf=0.85)

  Experimental — accuracy not yet measured

May combine two claims.

YouTube's recommendations amplify far-right content, fostering echo chambers.

“while YouTube's recommendations amplify far-right content, fostering echo chambers [60,86].”

• cites [60] Whittaker, J., Looney, S., Reed, A., & Votta, F. (2021). Recommender systems and the amplification of extremist content. Internet Policy Review 10(3). sourcePartly supported

  The Policy discussion reports that YouTube recommended more extreme videos after accounts interacted predominantly with far-right content, supporting the amplification claim under that condition. The Filter bubbles section says evidence for echo chambers is limited; the study does not establish that its recommendations fostered them. (LLM, conf=0.87)

  Experimental — accuracy not yet measured
• cites [86] Kaiser, J., & Rauchfleisch, A. (2020). Birds of a feather get recommended together: Algorithmic homophily in youtube’s channel recommendations in the United States and Germany. Social Media + Society 6(4). https://doi.org/10.1177/2056305120969914 sourceNot checked — no full text of the source

May combine two claims.

Reducing exposure to like-minded content during the 2020 US presidential election had no measurable effect on polarization in beliefs or attitudes, questioning the causal role of echo chambers in driving societal divides.

“Furthermore, Nyhan et al. [77] found that although exposure to like-minded content is prevalent on Facebook, reducing such exposure during the 2020 US presidential election had no measurable effect on polarization in beliefs or attitudes.”

• Also stated in: § 4.1.3.1 Polarization and fragmentation of the public sphere · hedged

  “However, the study by Nyhan et al. [77] demonstrates using a randomized controlled experiment that reducing exposure to like-minded sources may not significantly alter polarization, further questioning the causal role of echo chambers in driving societal divides.”

• cites [77] Nyhan, B., Settle, J., Thorson, E., et al. (2023). Like-minded sources on facebook are prevalent but not polarizing. Nature, 620, 137–144. https://doi.org/10.1038/s41586-023-06297-w sourceSupported

May combine two claims.

Reddit's customizable recommendation algorithms encourage exposure to diverse viewpoints, reducing echo chamber effects compared to Facebook.

“For example, Reddit's customizable recommendation algorithms encourage exposure to diverse viewpoints, reducing echo chamber effects compared to Facebook [17,61,63].”

• cites [17] Cinelli, M., Morales, G.D.F., Galeazzi, A., Quattrociocchi, W., & Starnini, M. (2021). The echo chamber effect on social media. Proceedings of the National Academy of Sciences of the United States of America 118(9). sourcePartly supported

  The Comparative Analysis section (‘News Consumption on Facebook and Reddit’; page unavailable) finds clearer separation by leaning on Facebook, while Reddit users, including those with more extreme leanings, tend to interact with the majority. It does not attribute that difference to Reddit’s customizable recommendation algorithms or establish that those algorithms encourage exposure to diverse viewpoints. (LLM, conf=0.88)

  Experimental — accuracy not yet measured
• cites [61] Monti, C., D’Ignazi, J., Starnini, M., & Morales, G. D. F. (2023). Evidence of demographic rather than ideological segregation in news discussion on reddit. In: Proceedings of the ACM Web Conference 2023 (WWW ’23) (pp. 1–10). ACM, New York, NY, USA. https://doi.org/10.1145/3543507.3583468 sourcePartly supported

  The Discussion (section 6; page not available in our copy) reports no ideological selective exposure in the Reddit news discussions studied and contrasts generalist Reddit communities with reported echo chambers on Facebook. Background and Related Work (section 2; page not available) also reports higher segregation of news consumption on Facebook than Reddit. The paper does not attribute this difference to customizable recommendation algorithms; it discusses a hypothesis about the platforms’ different structures instead. (LLM, conf=0.87)

  Experimental — accuracy not yet measured
• cites [63] Morales, G. D. F., Monti, C., & Starnini, M. (2021). No echo in the chambers of political interactions on reddit. Scientific Reports 11(1). sourcePartly supported

  The Discussion reports interactions between opposing political groups on Reddit during the 2016 election and suggests that its forum structure may be less conducive to echo chambers than social networks such as Facebook. It does not attribute that difference to Reddit’s customizable recommendation algorithms; its discussion of Facebook instead emphasizes friendship-network homophily over the news-feed algorithm. (LLM, conf=0.88)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

YouTube amplifies extreme and fringe content through algorithmic recommendations, while Reddit and Gab do not exhibit the same patterns.

“They found that YouTube amplifies extreme and fringe content through algorithmic recommendations, while Reddit and Gab do not exhibit the same patterns.”

• cites [60] Whittaker, J., Looney, S., Reed, A., & Votta, F. (2021). Recommender systems and the amplification of extremist content. Internet Policy Review 10(3). sourceSupported

May combine two claims.

Many studies found a small echo chamber effect for Republicans, but none for Democrats before and in the aftermath of the 2016 U.S. election.

“Many studies found a small echo chamber effect for Republicans, but none for Democrats before and in the aftermath of the 2016 U.S. election [70][71][72][73].”

• cites [70] Justwan, F., Baumgaertner, B., Carlisle, J. E., Clark, A. K., & Clark, M. (2018). Social media echo chambers and satisfaction with democracy among democrats and republicans in the aftermath of the 2016 us elections. Journal of Elections Public Opinion and Parties, 28(4), 424–442. sourcePartly supported

  The Conclusion (page unavailable) reports higher satisfaction with democracy among Republicans in social-media echo chambers and no detected effect among Democrats, based on a January 2017 survey. The paper does not establish the same pattern before the election or characterize the Republican effect as small; the Results report predicted satisfaction of 96.6% for Republicans in echo chambers versus 72.0% for those outside them. (LLM, conf=0.88)

  Experimental — accuracy not yet measured
• cites [71] Neely, S. R. (2021). Politically motivated avoidance in social networks: A study of facebook and the 2020 presidential election. Social Media + Society 7(4). https://doi.org/10.1177/20563051211055438 sourceNot checked — no full text of the source
• cites [72] Eady, G., Nagler, J., Guess, A., Zilinsky, J., & Tucker, J. A. (2019). How many people live in political bubbles on social media? evidence from linked survey and twitter data. SAGE OPEN 9(1). https://doi.org/10.1177/2158244019832705 sourceNot checked — no full text of the source
• cites [73] Boutyline, A., & Willer, R. (2017). The social structure of political echo chambers: Variation in ideological homophily in online networks. Political Psychology, 38(3), 551–569. https://doi.org/10.1111/pops.12337 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Existing polarization shapes social media usage patterns, rather than social media driving societal polarization.

“Nordbrandt [76] challenged the assumption that social media drives societal polarization, instead proposing that existing polarization shapes social media usage patterns.”

• cites [76] Nordbrandt, M. (2019). Affective polarization in the digital age: Testing the direction of the relationship between social media and users’ feelings for out-group parties. New Media & Society. https://doi.org/10.1177/14614448211044393 sourceSupported

May combine two claims.

On Reddit, studies find minimal evidence of echo chambers due to platform features encouraging cross-cutting exposure.

“On Reddit, studies find minimal evidence of echo chambers due to platform features encouraging cross-cutting exposure [17,61,63].”

• cites [17] Cinelli, M., Morales, G.D.F., Galeazzi, A., Quattrociocchi, W., & Starnini, M. (2021). The echo chamber effect on social media. Proceedings of the National Academy of Sciences of the United States of America 118(9). sourcePartly supported

  The Comparative Analysis, “News Consumption on Facebook and Reddit” (page not provided), finds less segregation on Reddit and says that even users with more extreme leanings tend to interact with the majority. It does not establish that Reddit’s platform features cause this pattern or encourage cross-cutting exposure. (LLM, conf=0.85)

  Experimental — accuracy not yet measured
• cites [61] Monti, C., D’Ignazi, J., Starnini, M., & Morales, G. D. F. (2023). Evidence of demographic rather than ideological segregation in news discussion on reddit. In: Proceedings of the ACM Web Conference 2023 (WWW ’23) (pp. 1–10). ACM, New York, NY, USA. https://doi.org/10.1145/3543507.3583468 sourcePartly supported

  The Introduction reports that ideological echo chambers were absent in the studied Reddit news discussions, and the Discussion says the findings strengthen evidence for their absence in generalist Reddit communities. But the Discussion notes that niche subreddits may form echo chambers and presents platform structure as a hypothesis for the difference, not an established cause of cross-cutting exposure. (LLM, conf=0.88)

  Experimental — accuracy not yet measured
• cites [63] Morales, G. D. F., Monti, C., & Starnini, M. (2021). No echo in the chambers of political interactions on reddit. Scientific Reports 11(1). sourcePartly supported

  The Introduction reports no evidence of echo chambers and more cross-cutting political interactions than expected in the Reddit discussions studied. The Discussion suggests that Reddit’s forum structure might matter, but says further research is needed to determine whether the finding extends beyond the 2016 election; it does not establish that platform features caused the cross-cutting interactions. (LLM, conf=0.87)

  Experimental — accuracy not yet measured

May combine two claims.

Content bias varies by platform, suggesting the impact of platform design on echo chamber formation.

“Nikolov et al. [54] also observe that content bias varies by platform, suggesting the impact of platform design on echo chamber formation.”

• cites [54] Nikolov, D., Lalmas, M., Flammini, A., & Menczer, F. (2019). Quantifying biases in online information exposure. Journal of the Association for Information Science and Technology, 70(3), 218–229. https://doi.org/10.1002/asi.24121 sourcePartly supported

  The Discussion (p. 15) reports differences in exposure bias across platforms and suggests that differences in platform form and function may help explain them. But the authors say they did not examine content or specific platform mechanisms; the Discussion (p. 17) describes echo-chamber formation as a possible consequence of exposure bias, not a demonstrated effect of platform design. (LLM, conf=0.88)

  Experimental — accuracy not yet measured

May combine two claims.

Echo chamber detection can occur on individual users, groups, or entire platforms.

“The difficulties in operationalizing extend to the granularity of echo chamber detection as they can occur on individual users, groups, or entire platforms [111]”

• cites [111] Kratzke, N. (2023). How to find orchestrated trolls? a case study on identifying polarized twitter echo chambers. Computers 12(3). https://doi.org/10.3390/computers12030057 sourceSupported

May combine two claims.

User leaning was also estimated by activity in specific groups, by semantics in the posts, calculation of political slant of visited websites by the user, and shared news URLs with news outlets leaning score.

“User leaning was also estimated by activity in specific groups (e.g., [59,127,143]), by semantics in the posts (e.g., [94,99,113]), calculation of political slant of visited websites by the user [124], shared news URLs with news outlets leaning score (e.g., [8,96,129]).”

• cites [59] Rafail, P., & Freitas, I. (2019). Grievance articulation and community reactions in the men’s rights movement online. Social Media + Society 5(2). https://doi.org/10.1177/2056305119841387 sourceNot checked — no full text of the source
• cites [127] Bessi, A., Zollo, F., Del Vicario, M., Scala, A., Caldarelli, G., & Quattrociocchi, W. (2015). Trend of narratives in the age of misinformation. PLOS ONE 10(8). sourceSupported
• cites [143] Cinelli, M., Pelicon, A., Mozetic, I., Quattrociocchi, W., Novak, P. K., & Zollo, F. (2021). Dynamics of online hate and misinformation. Scientific Reports 11(1). sourcePartly supported

  In Results and discussion, “Users’ behaviour and misinformation” (page unavailable), the paper estimates user leaning from the fraction of comments a user posts on videos from questionable YouTube channels. It does not estimate leaning from shared news URLs or news outlets’ political-leaning scores, the specific method to which this citation is attached. (LLM, conf=0.84)

  Experimental — accuracy not yet measured
• cites [94] Cann, T. J. B., Weaver, I. S., & Williams, H. T. P. (2021). Ideological biases in social sharing of online information about climate change. PLOS ONE 16(4). sourcePartly supported

  The Methods section (‘Ideological coding of source domains’; page not supplied) says the authors scored the political leaning of domains whose URLs were shared. The Results section (‘Climate media sharing is polarised and politicised’; page not supplied) uses those scores to characterise sharing-network clusters. The paper does not estimate political leaning for individual users, so it supports the cited URL-and-outlet method only at the community level. (LLM, conf=0.88)

  Experimental — accuracy not yet measured
• cites [99] Asatani, K., Yamano, H., Sakaki, T., & Sakata, I. (2021). Dense and influential core promotion of daily viral information spread in political echo chambers. Scientific Reports 11(1). sourceCould not verify

  The citation to this paper appears in the clause about estimating user leaning from shared news URLs and news outlets’ leaning scores. The Results section (“Independent and steady political echo chambers”; page unavailable) identifies left- and right-leaning communities using reply/retweet relationships and notes liberal and conservative news media among influential accounts. It does not describe scoring shared news URLs or using outlet-leaning scores to estimate users’ leanings; searches for those methods found no supporting passage. (LLM, conf=0.72)

  Experimental — accuracy not yet measured
• cites [113] Colleoni, E., Rozza, A., & Arvidsson, A. (2014). Echo chamber or public sphere? predicting political orientation and measuring political homophily in twitter using big data. Journal of Communication, 64(2), 317–332. sourceNot checked — no full text of the source
• cites [124] Shore, J., Baek, J., & Dellarocas, C. (2018). Network structure and patterns of information diversity on twitter. Management Information Systems Quarterly, 42(3), 849–972. sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Users with ties on social media share partisanship for a party or not.

“This means that users with ties on social media share partisanship for a party or not[149].”

• cites [149] Mosleh, M., Martel, C., Eckles, D., & Rand, D. G. (2021). Shared partisanship dramatically increases social tie formation in a twitter field experiment. Proceedings of the National Academy of Sciences of the United States of America 118(7). sourceSupported

May combine two claims.

The EU DSA grants vetted researchers access to platform data and enables recommendation system audits.

“which grants vetted researchers access to platform data and enables recommendation system audits [88].”

• cites [88] Hartmann, D., Pereira, J. R. L., Streitbörger, C., & Berendt, B. (2024). Addressing the regulatory gap: Moving towards an EU AI audit ecosystem beyond the AIA by including civil society. sourceSupported

Probably not a substantial claim: may not pass S3 (load-bearing).

Network-level homophily measures the overall homophily of a network to understand the systemic tendencies for like-minded clustering.

“Network-Level Homophily: Measures the overall homophily of a network to understand the systemic tendencies for like-minded clustering [99,117].”

• cites [99] Asatani, K., Yamano, H., Sakaki, T., & Sakata, I. (2021). Dense and influential core promotion of daily viral information spread in political echo chambers. Scientific Reports 11(1). sourceCould not verify

  The Introduction discusses how homophily can emerge, and the Methods section describes clustering users by reply and retweet relationships. Neither section defines or uses a network-level homophily measure to assess overall like-minded clustering (searched: “network-level homophily,” “homophily measure,” and “homophily”). No page numbers are available in our copy. (LLM, conf=0.73)

  Experimental — accuracy not yet measured
• cites [117] Enjolras, B., & Salway, A. (2022). Homophily and polarization on political twitter during the 2017 Norwegian election. Social Network Analysis and Mining, 13(1), 10. https://doi.org/10.1007/s13278-022-01018-z sourceSupported

Classified as probably not a substantial claim. May combine two claims.

The echo chamber construct is controversial and was ill-defined from the beginning.

“Firstly, the echo chamber construct is controversial, and researchers claim that it was ill-defined from the beginning and that researchers should either abandon or conceptualize it properly [91].”

• cites [91] Mahrt, M. (2019). Beyond Filter Bubbles and Echo Chambers: The Integrative Potential of the Internet. Digital Communication Research 5, 246. Digital Communications Research, Berlin. https://doi.org/10.17174/dcr.v5.0 sourcePartly supported

  Section 3.4 reports mixed evidence for online echo chambers, supporting the claim that the idea is contested. Section 4.3 says research lacks a clear threshold for *fragmentation*, but our copy does not say the echo-chamber construct was ill-defined from the beginning or recommend abandoning or reconceptualizing it. No page numbers were available for these passages. (LLM, conf=0.79)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

Group-level homophily aggregates the homophily of individual group members to assess group polarization.

“Group-Level Homophily: Aggregates the homophily of individual group members to assess group polarization [25].”

• cites [25] Interian, R., Marzo, R. G., Mendoza, I., & Ribeiro, C. C. Network polarization, filter bubbles, and echo chambers: an annotated review of measures and reduction methods. International Transactions in Operational Research. https://doi.org/10.1111/itor.13224. sourceSupported

Probably not a substantial claim: may not pass S3 (load-bearing).

Machine learning-based bot detection offers a way to mitigate this issue.

“Machine learning-based bot detection, as proposed by Rusche [110], and implemented by Choi et al. [100] and Gaumont et al. [120], offers a way to mitigate this issue.”

• cites [110] Rusche, F. (2022). Few voices, strong echo: Measuring follower homogeneity of politicians’ twitter accounts. New Media & Society, 14614448221099860. https://doi.org/10.1177/14614448221099860 sourceSupported
• cites [100] Choi, D., Chun, S., Oh, H., Han, J., & Kwon, T. (2020). Rumor propagation is amplified by echo chambers in social media. Scientific Reports, 10(1), 310. sourcePartly supported

  In Methods (Data collection), Choi et al. report using a machine-learning bot detector to identify bot accounts and remove their tweets and retweets when building rumor cascades. This supports the statement that they implemented bot detection, but the paper does not show that it mitigated rumor propagation; the Discussion (Limitation and future work) says detecting and preventing rumor spreads was not covered. (LLM, conf=0.85)

  Experimental — accuracy not yet measured
• cites [120] Gaumont, N., Panahi, M., & Chavalarias, D. (2018). Reconstruction of the socio-semantic dynamics of political activist twitter networks-method and application to the 2017 French presidential election. PLOS ONE 13(9). sourceCould not verify

  The Abstract (Method) describes automated categorization of Twitter accounts by political affiliation, not bot detection. The Results (Propagation of political information) discuss strategies for mitigating fake news, but do not propose or implement machine-learning-based bot detection. Searches for bot detection, automated accounts, and bots found no passage supporting the attributed claim. (LLM, conf=0.76)

  Experimental — accuracy not yet measured

May combine two claims.

Latent models can be biased by parameter choices and assumptions and may fall into circular reasoning.

“Latent models can be biased by parameter choices and assumptions and may fall into circular reasoning, such as inferring political positions from network structures and then using those positions to explain the same structures [155].”

• cites [155] Barbera, P. (2013). Birds of the same feather tweet together: Bayesian ideal point estimation using twitter data. SSRN Electronic Journal 23. https://doi.org/10.2139/ssrn.2108098 sourcePartly supported

  Section 2.3 (Identification; page unavailable) describes how parameter restrictions and starting values affect identification of the latent scale, and Section 2.1 states the model’s key assumption about whom users follow. The paper does not say these choices bias latent models or warn that inferring positions from a network and then using them to explain that same network is circular reasoning. (LLM, conf=0.78)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

From these, unipartite projections or multiple bipartite graphs are derived to study user-to-user connections or user-content dynamics.

“From these, unipartite projections or multiple bipartite graphs are derived to study user-to-user connections or user-content dynamics [94,96].”

• cites [94] Cann, T. J. B., Weaver, I. S., & Williams, H. T. P. (2021). Ideological biases in social sharing of online information about climate change. PLOS ONE 16(4). sourcePartly supported

  Methods > Network construction describes user–URL bipartite networks and a unipartite projection used to analyse sharing patterns. But that projection connects URLs, not users, so this paper does not support the claim that it studies user-to-user connections through the projection. (LLM, conf=0.88)

  Experimental — accuracy not yet measured
• cites [96] Del Vicario, M., Zollo, F., Caldarelli, G., Scala, A., & Quattrociocchi, W. (2017). Mapping social dynamics on facebook: The brexit debate. Social Networks, 50, 6–16. sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Node-level homophily focuses on individual tendencies to connect with similar users.

“Node-Level Homophily: Focuses on individual tendencies to connect with similar users (e.g., [73,99]).”

• cites [73] Boutyline, A., & Willer, R. (2017). The social structure of political echo chambers: Variation in ideological homophily in online networks. Political Psychology, 38(3), 551–569. https://doi.org/10.1111/pops.12337 sourceNot checked — no full text of the source
• cites [99] Asatani, K., Yamano, H., Sakaki, T., & Sakata, I. (2021). Dense and influential core promotion of daily viral information spread in political echo chambers. Scientific Reports 11(1). sourcePartly supported

  The Introduction links homophily to individual cognitive and psychological behaviour, and the Results section on the connected core assumes that users with the same opinion connect with each other. The paper does not identify or define a distinct concept called “node-level homophily,” so it supports the underlying idea but not that specific framing. The indexed copy provides no page numbers for these passages. (LLM, conf=0.78)

  Experimental — accuracy not yet measured

May combine two claims.

API and scraping studies are predominantly platform-centered, while data donations and tracking data are mainly user-centered.

“API and scraping studies are predominantly platform-centered, while data donations and tracking data are mainly user-centered [107].”

• cites [107] Ohme, J., Araujo, T., Boeschoten, L., Freelon, D., Ram, N., Reeves, B. B., & Robinson, T. N. (2023). Digital trace data collection for social media effects research: Apis, data donation, and (screen) tracking. Communication Methods and Measures 1–18. sourceSupported

Probably not a substantial claim: may not pass S3 (load-bearing).

Users can be connected in a graph through edges if they are friends or have shared content.

“Thus, users can be connected in a graph through edges if they are friends (e.g., [136]) or have shared content.”

• cites [136] Urman, A. (2019). News consumption of Russian Vkontakte users: Polarization and news avoidance. International Journal of Communication, 13, 5158–5182. sourceCould not verify

  The citation is attached to the claim about connecting users who are friends, not the separate claim about shared content. In “Polarization Is Real” (page unavailable), the paper says it examines subscription patterns, not friendship ties; its network description under Figure 1 (page unavailable) defines edges between users and pages they follow. Neither passage addresses edges between users who are friends. (LLM, conf=0.76)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

Centrality measures identify influential nodes, such as disseminators (high closeness centrality) or bridges (high betweenness centrality), offering insights into individual roles within echo chambers.

“Centrality measures identify influential nodes, such as disseminators (high closeness centrality) or bridges (high betweenness centrality), offering insights into individual roles within echo chambers [6,111].”

• cites [6] Torregrosa, J., Panizo-Lledot, A., Bello-Orgaz, G., & Camacho, D. (2020). Analyzing the relationship between relevance and extremist discourse in an alt-right network on twitter. Social Network Analysis and Mining 10(1). sourceNot checked — no full text of the source
• cites [111] Kratzke, N. (2023). How to find orchestrated trolls? a case study on identifying polarized twitter echo chambers. Computers 12(3). https://doi.org/10.3390/computers12030057 sourcePartly supported

  Materials and Methods (section 2; page unavailable) says centrality measures can identify critical actors in echo chambers and explicitly describes high-betweenness actors as bridges between groups. The paper mentions closeness centrality but does not identify high-closeness actors as disseminators; section 2.2.3 instead calls high-HITS-hub accounts content disseminators. (LLM, conf=0.88)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

Metrics like polarity scores measure the distance between users based on their attributes when group membership is unavailable.

“Metrics like polarity scores measure the distance between users based on their attributes when group membership is unavailable (e.g., [73,100]).”

• cites [73] Boutyline, A., & Willer, R. (2017). The social structure of political echo chambers: Variation in ideological homophily in online networks. Political Psychology, 38(3), 551–569. https://doi.org/10.1111/pops.12337 sourceNot checked — no full text of the source
• cites [100] Choi, D., Chun, S., Oh, H., Han, J., & Kwon, T. (2020). Rumor propagation is amplified by echo chambers in social media. Scientific Reports, 10(1), 310. sourcePartly supported

  The Results section (‘Properties of rumor echo chambers’; page unavailable) describes polarity scores derived from users’ follows and tweets, and a homogeneity score comparing two users’ political positions. It does not present these scores as a way to measure distance when group membership is unavailable; its comparison explicitly distinguishes echo-chamber members from non-members. (LLM, conf=0.84)

  Experimental — accuracy not yet measured

Probably not a substantial claim: may not pass S3 (load-bearing).

After removing duplicates and including only peer-reviewed work, 1,051 unique publications were screened for eligibility.

“This selection resulted in 1,051 unique publications that were then screened for eligibility.”

May combine two claims.

Studies differ in outcomes, focus, construct conceptualization, operationalization, and granularity.

“We find that studies differ in (1) outcomes, (2) focus, (3) construct conceptualization and (4) operationalization, and (5) granularity.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

A total of 1,051 studies were screened based on title, excluding 655 studies that did not meet the three criteria.

“A total of 1,051 studies were screened based on title, excluding 655 studies that did not meet the three criteria.”

May combine two claims.

Studies based on homophily and computational social science methods often support the echo chamber hypothesis, while research on content exposure and broader media environments, such as surveys, tends to challenge it.

“Studies based on homophily and computational social science methods often support the echo chamber hypothesis, while research on content exposure and broader media environments, such as surveys, tends to challenge it.”

May combine two claims.

The review of 129 studies identifies variations in measurement approaches, as well as regional, political, cultural, and platform-specific biases, as key factors contributing to the lack of consensus.

“The review of 129 studies identifies variations in measurement approaches, as well as regional, political, cultural, and platform-specific biases, as key factors contributing to the lack of consensus.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The 129 studies included in the systematic review were all published between 2014 and 2023.

“The 129 studies included in the systematic review were all published between 2014 and 2023, as Fig. 4a illustrates.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Database searches yielded a total of 1,706 results.

“Database searches yielded a total of 1,706 results.”

Classified as probably not a substantial claim. May combine two claims.

Group behavior, cultural influences, instant messaging platforms, and short video platforms remain underexplored.

“Group behavior, cultural influences, instant messaging platforms, and short video platforms remain underexplored.”

Probably not a substantial claim: may not pass S3 (load-bearing).

514 duplicates were removed.

“514 duplicates were removed.”

Probably not a substantial claim: may not pass S3 (load-bearing).

From an initial set of 1,706 studies, 129 studies were selected based on criteria.

“From an initial set of 1,706 studies, we selected these 129 studies based on criteria explained in Sect.3.4.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The survey by Dubois et al. found that political involvement and media variety reduce the risk of being in an echo chamber.

“The survey by Dubois et al. [65] found that political involvement and media variety reduce the risk of being in an echo chamber.”

• cites [65] Dubois, E., Minaeian, S., Paquet-Labelle, A., & Beaudry, S. (2020). Who to trust on social media: How opinion leaders and seekers avoid disinformation and echo chambers. Social Media + Society 6(2). https://doi.org/10.1177/2056305120913993 sourceNot checked — no full text of the source

May combine two claims.

Recommendation algorithms significantly influence echo chambers, but gaps remain in understanding the interplay between moderation practices and echo chamber dynamics.

“While evidence suggests that recommendation algorithms significantly influence echo chambers, gaps remain in understanding the interplay between moderation practices and echo chamber dynamics.”

Classified as probably not a substantial claim.

After analysis of studies, the authors differentiate between single-user, group, platform, cross-platform, and holistic granularity.

“After analysis of studies, we differentiate between single-user, group, platform, cross-platform, and holistic.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Fragmentation, measured through the size and density of communities across entire platforms, has been comparatively understudied.

“On the other hand, fragmentation, measured through the size and density of communities across entire platforms, has been comparatively understudied.”

May combine two claims.

On Facebook, Beam et al. found that users exposed to newsfeeds experienced depolarization over time, suggesting that social media could reduce rather than reinforce ideological segregation.

“On Facebook, Beam et al. [74] found that users exposed to newsfeeds experienced depolarization over time, suggesting that social media could reduce rather than reinforce ideological segregation.”

• cites [74] Beam, M. A., Hutchens, M. J., & Hmielowski, J. D. (2018). Facebook news and (de)polarization: Reinforcing spirals in the 2016 us election. Information, Communication & Society, 21(7), 940–958. https://doi.org/10.1080/1369118X.2018.1444783 sourceNot checked — no full text of the source

May combine two claims.

14 studies in the corpus investigate the relationship between echo chambers and extremism, examining how homophily and algorithmic amplification contribute to radicalization.

“14 studies in the corpus investigate the relationship between echo chambers and extremism, examining how homophily and algorithmic amplification contribute to radicalization.”

Classified as probably not a substantial claim.

Four key concepts are integral to echo chamber conceptualizations: homophily, content exposure, individual user behaviors (such as selective exposure, confirmation bias or fear), and group behaviors (including exclusion of outsiders and collective unfriending).

“We identified four key concepts integral to echo chamber conceptualizations: (1) homophily, (2) content exposure, (3) individual user behaviors, such as selective exposure, confirmation bias or fear, and (4) group behaviors, including exclusion of outsiders and collective unfriending.”

May combine two claims.

Few papers looked at the specific group behavior of these groups but mainly examined homophily over time and information flow to the group as opposed to outside the group.

“However, few papers looked at the specific group behavior of these groups but mainly examined homophily over time and information flow to the group as opposed to outside the group.”

May combine two claims.

Eady et al. and Barbera et al. observed significant ideological overlap among the most liberal and conservative users, contradicting the assumption of rigid polarization.

“Similarly, Eady et al. [72] and Barbera et al. [75] observed significant ideological overlap among the most liberal and conservative users, contradicting the assumption of rigid polarization.”

• cites [72] Eady, G., Nagler, J., Guess, A., Zilinsky, J., & Tucker, J. A. (2019). How many people live in political bubbles on social media? evidence from linked survey and twitter data. SAGE OPEN 9(1). https://doi.org/10.1177/2158244019832705 sourceNot checked — no full text of the source
• cites [75] Barbera, P., Jost, J. T., Nagler, J., Tucker, J. A., & Bonneau, R. (2015). Tweeting from left to right: Is online political communication more than an echo chamber? Psychological Science, 26(10), 1531–1542. https://doi.org/10.1177/0956797615594620 sourceNot checked — no full text of the source

May combine two claims.

Selective exposure, demographics, emotions, and personality traits are key factors in echo chamber formation.

“Selective exposure, demographics, emotions, and personality traits have been identified as key factors in echo chamber formation”

May combine two claims.

The relationship between societal polarization, fragmentation, and echo chambers is complex and often correlational rather than causal.

“The relationship between societal polarization, fragmentation, and echo chambers is complex and often correlational rather than causal.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Recommendation systems and content moderation are frequently cited as primary drivers of echo chambers due to their role in shaping user content exposure.

“recommendation systems and content moderation are frequently cited as primary drivers of echo chambers due to their role in shaping user content exposure.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Many studies put more focus on the content that an individual sees through their news feed curated by recommendation systems.

“This is why many studies put more focus on the content that an individual sees through their news feed curated by recommendation systems.”

May combine two claims.

Polarization may stem from other factors, such as traditional media, cultural differences, demographics, or other offline influences, complicating the attribution of causality solely to social media.

“However, novel studies from 2023 also highlight that polarization may stem from other factors, such as traditional media, cultural differences, demographics, or other offline influences, complicating the attribution of causality solely to social media.”

May combine two claims.

Twenty-three studies in the corpus address the connection between echo chambers and misinformation, with seven focusing explicitly on misinformation in the context of echo chambers.

“Twenty-three studies in the corpus address this connection, with seven focusing explicitly on misinformation in the context of echo chambers.”

May combine two claims.

Such an echo chamber conceptualization is hardly reflected in the corpus, although there is research that examines particular group behavior and presents it as echo chamber-specific.

“However, such an echo chamber conceptualization is hardly reflected in the corpus, although there is research that examines particular group behavior and presents it as echo chamber-specific.”

Probably not a substantial claim: may not pass S3 (load-bearing).

It remains unclear how selective exposure and recommendation systems interact with each other.

“it remains unclear how selective exposure and recommendation systems interact with each other”

May combine two claims.

Facebook users tend to consume more partisan news than Reddit users, who encounter greater content diversity.

“Comparative studies, such as Kitchens et al. [24], found that Facebook users tend to consume more partisan news than Reddit users, who encounter greater content diversity.”

• cites [24] Kitchens, B., Johnson, S. L., & Gray, P. (2020). Understanding echo chambers and filter bubbles: The impact of social media on diversification and partisan shifts in news consumption. MIS Quarterly, 44(4), 1619–1649. sourceNot checked — no full text of the source

May combine two claims.

User engagement has been found to correlate positively with echo chambers and political affiliation.

“User engagement has been found to correlate positively with echo chambers and political affiliation.”

May combine two claims.

The majority of results are consistent with the echo chamber hypothesis that echo chambers contribute to both polarization and societal fragmentation.

“Despite diverse operationalizations, the majority of results are consistent with the echo chamber hypothesis: echo chambers contribute to both polarization and societal fragmentation.”

May combine two claims.

Although group behaviors play a role in echo chamber formation, few studies have investigated them comprehensively or as antecedents of the construct.

“Although group behaviors play a role in echo chamber formation, few studies have investigated them comprehensively or as antecedents of the construct.”

Probably not a substantial claim: may not pass S4 (based).

Previous studies have focused on either selective exposure or recommendation systems, not their interaction.

“given that previous studies have focused on one or the other.”

May combine two claims.

Out of all survey studies, 17 looked at general social media usage or the whole media diet, while only two looked at specific platforms.

“Out of all survey studies, 17 looked at general social media usage or the whole media diet, while only two looked at specific platforms.”

May combine two claims.

In the identified studies, nodes often represent users and edges denote interactions between users and interaction of users with content such as posts or hyperlinks to news pages.

“In the identified studies, nodes often represent users and edges denote interactions between users and interaction of users with content such as posts or hyperlinks to news pages (e.g., [108]).”

• cites [108] Aruguete, N., Calvo, E., & Ventura, T. (2023). News by popular demand: Ideological congruence, issue salience, and media reputation in news sharing(1). International Journal of Press-Politics. https://doi.org/10.1177/19401612211057068 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

About half of the survey studies were conducted in the USA, a total of 8.

“About half of the surveys were conducted in the USA (a total of 8), and only 5 of these studies addressed the echo chamber hypothesis directly.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Each user can be characterized by a vector of node attributes, such as demographic information, interests, or activity patterns.

“Each user can be characterized by a vector of node attributes, 8 such as demographic information, interests, or activity patterns.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

While these metrics provide user-level insights, they do not directly capture broader interaction patterns or group dynamics.

“While these metrics provide user-level insights, they do not directly capture broader interaction patterns or group dynamics.”

May combine two claims.

Studies using CSS methods typically conceptualize echo chambers as attribute-based homophilic structures, whereas surveys and experimental studies focus on content exposure and user behavior.

“For instance, studies using CSS methods typically conceptualize echo chambers as attribute-based homophilic structures, whereas surveys and experimental studies focus on content exposure and user behavior.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The general procedure was to build a graph for every time-dependent sample and use interaction (structural data sets) or semantic networks (content-considering data sets).

“Based on these two types of data sets, the general procedure was to build a graph for every time-dependent sample as social networks are dynamic networks [142] and use interaction (structural data sets) or semantic networks (content-considering data sets)”

• cites [142] Light, R., & Moody, J. (2021). 16network basics: Points, lines, and positions. In The Oxford handbook of social networks. Oxford University Press. https://doi.org/10.1093/oxfordhb/9780190251765.013.2 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

API restrictions have made access more challenging in recent years.

“However, API restrictions have made access more challenging in recent years.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Purely structural data sets consist of interactions between users such as 'likes', 'posts', 'comments', 'friendship', 'shares', etc.; contentconsidering data sets use the content of 'posts' and 'comments' to gain insights into the users political leaning, polarization, radicalism or other attributes.

“The former data sets consist of interactions between users such as 'likes', 'posts', 'comments','friendship', 'shares', etc.; the latter use the content of 'posts' and 'comments' to gain insights into the users political leaning, polarization, radicalism or other attributes.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The underlying data sets can be classified into API and scraping data, crowdsourced or donated data, and tracking data.

“The underlying data sets can be classified into API and scraping data, crowdsourced or donated data, and tracking data.”

May combine two claims.

Survey studies use between 5 and 41 items to gather data from a participant pool of 198 to 11,052 individuals.

“The studies use between 5 and 41 items to gather data from a participant pool of 198 to 11,052 individuals.”

The quote only partly states this claim.

Only 5 of the survey studies addressed the echo chamber hypothesis directly.

“About half of the surveys were conducted in the USA (a total of 8), and only 5 of these studies addressed the echo chamber hypothesis directly.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The next methodological step for most of the social data methods was to use a community detection algorithm.

“The next methodological step for most of the social data methods was to use a community detection algorithm.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The edges can be weighted by the strength of the social tie between users, such as the frequency, intensity of interactions, or a friendship tie.

“The edges can be weighted by the strength of the social tie between users, such as the frequency, intensity of interactions, or a friendship tie.”

May combine two claims.

In the reviewed literature, many studies identify associations but few establish causal relationships, such as whether echo chambers amplify polarization or whether pre-existing polarization fosters echo chambers.

“While many studies identify associations, few establish causal relationships, such as whether echo chambers amplify polarization or whether pre-existing polarization fosters echo chambers.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The user is connected to posts or news pages that they have shared, thus also creating a semantic network.

“At the same time, the user is connected to posts or news pages that they have shared, thus also creating a semantic network.”

Probably not a substantial claim: may not pass S3 (load-bearing).

There is another distinction between the underlying observational trace data sets for echo chamber research: purely structural data (interaction-based) and contentconsidering data sets.

“We find another distinction between the underlying observational trace data sets for echo chamber research: purely structural data (interaction-based) and contentconsidering data sets.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Ten experiments are part of the review corpus.

“Ten experiments are part of the corpus.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

To assess the similarity between different users in communities, different metrics are used.

“To asses the similarity between different users in communities, different metrics are used.”

May combine two claims.

Studies that included data donated by users included additional information to the traces, but were smaller in size.

“Studies that included data donated by users included additional information to the traces, but were smaller in size.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Out of all the analyzed studies in the review corpus, 20 were surveys.

“Out of all the analyzed studies, 20 were surveys.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Data consisted of posts, comments, likes, and other digital traces.

“Data consisted of posts, comments, likes, and other digital traces.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Adjustable recommender systems like those tested on Reddit allow users to customize content diversity, reducing algorithmic reinforcement of homogeneous content and introducing contrasting perspectives.

“Adjustable recommender systems like those tested on Reddit allow users to customize content diversity, reducing algorithmic reinforcement of homogeneous content and introducing contrasting perspectives.”

May combine two claims.

Most studies affirming the echo chamber hypothesis are conducted in the United States or use data concerning the United States population and are not representative of other regions.

“Concurrent with previous research, we find that most studies affirming the echo chamber hypothesis are conducted in the United States or use data concerning the United States population and are not representative of other regions.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Most of the studies look at a certain event, such as an election, to investigate emerging echo chambers around that event.

“most of the studies look at a certain event, such as an election to investigate emerging echo chambers around that event.”

May combine two claims.

Content-considering data sets either had to be labeled, cleaned, or had to apply resource-intensive Machine Learning (ML) algorithms to make sense of their data.

“whereas content-considering data sets either had to be labeled, cleaned or had to apply resource-intensive Machine Learning (ML) algorithms to make sense of their data.”

Classified as probably not a substantial claim.

To detect and analyze echo chambers, a variety of metrics are employed, categorized into interaction-based metrics, content-considering metrics, homophily and polarization measures, and modularity measures.

“To detect and analyze echo chambers, a variety of metrics are employed, categorized into interaction-based metrics, content-considering metrics, homophily”

Probably not a substantial claim: may not pass S3 (load-bearing).

Different amounts of users and data were used for the analysis.

“Different amounts of users and data were used for the analysis”

Probably not a substantial claim: may not pass S3 (load-bearing).

Data gathering periods differ among studies.

“Data gathering periods differ among studies as well”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Polarization and homophily measures are frequently calculated on these networks to assess ideological clustering.

“Polarization and homophily measures are frequently calculated on these networks to assess ideological clustering.”

May combine two claims.

Studies using homophily as a conceptualization tend to have affirming results on the echo chamber hypothesis, and studies using context exposure tend to have negative findings.

“Here again, studies using homophily as a conceptualization tend to have affirming results on the echo chamber hypothesis, and studies using context exposure tend to have negative findings.”

May combine two claims.

More studies from the US and UK affirm the echo chamber hypothesis and less from countries such as Germany, France, and the Netherlands.

“We find that more studies from the US and UK affirm the echo chamber hypothesis and less from countries such as Germany, France, and the Netherlands.”

May combine two claims.

Both Masip et al. and Beam et al. found no echo chambers or polarization through social media, contrary to homophily studies.

“Interestingly, both studies found no echo chambers or polarization through social media, contrary to homophily studies.”

May combine two claims.

Research on platforms such as Instagram, WhatsApp, Telegram, TikTok, and WeChat is notably absent, largely due to challenges in data accessibility.

“Research on platforms such as Instagram, WhatsApp, Telegram, TikTok, and WeChat is notably absent, largely due to challenges in data accessibility.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Some content-considering data sets first transformed the network into a weighted, undirected network by computing pairwise similarities between nodes of the same type.

“Some of the content-considering data sets first transformed the network into a weighted, undirected network by computing the pairwise similarities between nodes of the same type (e.g., similarity between users based on their shared interactions with content or friendship ties).”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Homophily and polarization metrics are used interchangeably in many studies to evaluate the similarity between users within echo chambers.

“Homophily and Polarization Metrics.These are used interchangeably in many studies to evaluate the similarity between users within echo chambers.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

In the context of the bipartite graph B, these algorithms are used to detect communities of users who are more likely to interact with content items within their community than with content items outside their community.

“In the context of the bipartite graph B, these algorithms are used to detect communities of users who are more likely to interact with content items within their community than with content items outside their community.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Some used scraping or tracking via plugins (e.g., Bing toolbar, YouGov).

“Some used scraping or tracking via plugins (e.g., Bing toolbar, YouGov).”

Probably not a substantial claim: may not pass S3 (load-bearing).

There were also works that examined longer time spans.

“There were also works that examined longer time spans.”

May combine two claims.

Most CSS studies used a form of social network analysis and followed a specific procedure.

“In general, most CSS studies used a form of social network analysis and followed this procedure:”

Classified as probably not a substantial claim.

Echo chambers can still emerge under specific conditions, such as heightened political interest, selective exposure tendencies, or high user activity.

“However, they also find that echo chambers can still emerge under specific conditions, such as heightened political interest, selective exposure tendencies, or high user activity.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The size of the data sets varied by platform and by study.

“The size of the data sets varied by platform and by study as Fig. 7 in the”

May combine two claims.

Studies employing CSS methods are more likely to identify echo chambers, while surveys and experiments tend to produce mixed or less conclusive results.

“Studies employing CSS methods are more likely to identify echo chambers, while surveys and experiments tend to produce mixed or less conclusive results.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Most survey studies and experiments did not research one platform in particular.

“It is important to note that most survey studies and experiments did not research one platform in particular and are therefore not considered in this plot”

Classified as probably not a substantial claim.

Figure 6b illustrates a clear relationship between methods and outcomes.

“Figure 6b illustrates a clear relationship between methods and outcomes.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Content-considering metrics estimate user attributes, such as political leanings, or analyze cross-cutting content exposure to understand how individuals engage with ideologically similar or diverse content.

“Content-Considering Metrics.They estimate user attributes, such as political leanings, or analyze cross-cutting content exposure to understand how individuals engage with ideologically similar or diverse content.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Interaction-based methods rely heavily on centrality measures to identify influential users and assess the structural dynamics of networks.

“Interaction-based methods.These metrics rely heavily on centrality measures to identify influential users and assess the structural dynamics of networks:”

May combine two claims.

After codebook establishment, the third coder independently coded 19 random studies (roughly 15% sample), and ICR was substantial (Cohen's = 0.66).

“After the codebook was established through an initial coding round and discussions between the first author and second author, the third coder independently coded 19 random studies (roughly a 15 % sample [39]).Here, ICR was substantial (Cohen's = 0.66foot_2 ).”

• cites [39] O’Connor, C., & Joffe, H. (2020). Intercoder reliability in qualitative research: Debates and practical guidelines. International Journal of Qualitative Methods, 19, 1609406919899220. https://doi.org/10.1177/1609406919899220 sourceNot checked — no full text of the source

May combine two claims.

The review only used English keywords and included studies written in English, which may exclude studies relating to non-English speaking countries and the Global South.

“In this review, we only used English keywords and included studies written in English for transparency and reproducibility reasons.This means that studies relating to certain countries might be excluded from our corpus and there could be results relating to non-English speaking countries and the Global South that are not included in our analysis.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Standardized methods for search, evaluation, and selection of studies reduce subjective influences.

“We use standardized methods for search, evaluation, and selection of studies reduce subjective influences [30].”

• cites [30] Biondi-Zoccai, G., Lotrionte, M., Landoni, G., & Modena, M. G. (2011). The rough guide to systematic reviews and meta-analyses. HSR Proceedings in Intensive Care and Cardiovascular Anesthesia, 3(3), 161–173.Not checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

There is no evidence that social media increases support for right-wing populism.

“Boulianne et al. [16] found no evidence that social media increases support for right-wing populism”

• cites [16] Boulianne, S., Koc-Michalska, K., & Bimber, B. (2020). Right-wing populism, social media and echo chambers in western democracies. New Media & Society, 22(4), 683–699. https://doi.org/10.1177/1461444819893983 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Network relevance is associated with extremist content, highlighting the role of user connections in spreading radical ideologies.

“Similarly, Torregrosa et al. [6] observed that network relevance is associated with extremist content, highlighting the role of user connections in spreading radical ideologies.”

• cites [6] Torregrosa, J., Panizo-Lledot, A., Bello-Orgaz, G., & Camacho, D. (2020). Analyzing the relationship between relevance and extremist discourse in an alt-right network on twitter. Social Network Analysis and Mining 10(1). sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Weak ties expose users to radical ideas, but the lack of reciprocal connections may limit deeper engagement.

“While these weak ties expose users to radical ideas, the lack of reciprocal connections may limit deeper engagement.”

• cites [101] Wolfowicz, M., Weisburd, D., & Hasisi, B. (2023). Examining the interactive effects of the filter bubble and the echo chamber on radicalization. Journal of Experimental Criminology. sourceNot checked — no full text of the source

May combine two claims.

Challenge avoidance and reinforcement seeking are cognitive mechanisms that explain the persistence of polarized behavior.

“Challenge avoidance and reinforcement seeking: Both are cognitive mechanisms that explain the persistence of polarized behavior.”

May combine two claims.

Significant ideological differences in self-censorship practices and fears of isolation exist, and political identity influences individuals' willingness to engage in public discourse.

“Burnett et al. [69] highlights significant ideological differences in self-censorship practices and fears of isolation, showing how political identity influences individuals' willingness to engage in public discourse”

• cites [69] Burnett, A., Knighton, D., & Wilson, C. (2022). The self-censoring majority: How political identity and ideology impacts willingness to self-censor and fear of isolation in the united states. Social Media + Society 8(3). sourceNot checked — no full text of the source