[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-81857-en":3,"doc-seo-81857-105":31,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":28,"seo_description":14,"update_tm":29,"read_time":30},81857,5909877438554,"Maeve","https://ap-avatar.wpscdn.com/avatar/5600025385ad2bf12a7?_k=1778553567797529272",8,"Research & Report","Angry but Accurate: Detecting and Profiling the Counter-Misinformation Ecosystem on Twitter","Social media users frequently push back against false claims, but the identities and mechanisms behind this corrective activity shape how misinformation is contested. This study examines the counter-misinformation ecosystem at scale by applying a domain-specific NLI model to a large COVID-19 tweet corpus, classifying 264,737 posts as supporting or opposing false claims and comparing 23 user- and text-level features across groups. Anti-misinformation posts show more emotional negativity (anger, disgust, sadness). They also originate from more established users, while bot-likelihood does not meaningfully differentiate sides.","Angry but Accurate: Detecting and Profiling the Counter-Misinformation Ecosystem on Twitter  \nEun Cheol Choi  \nUniversity of Southern California Annenberg School of Communication Los Angeles, CA, United States [euncheol@usc.edu](euncheol@usc.edu)  \nEmilio Ferrara  \nUniversity of Southern California Thomas Lord Department of Computer Science Los Angeles, CA, United States [emiliofe@usc.edu](emiliofe@usc.edu)  \narXiv :2607 .02900v 1 [ cs . SI] 3 Jul 2026  \nAbstract  \nOn social media, many users actively push back against false claims. Understanding who pushes back and how they do so matters, as this corrective activity is central to how misinformation is contested. We study this counter-misinformation ecosystem at scale: applying a domain-specific NLI model from our prior work to a large corpus of COVID-19 tweets, we classify 264,737 posts as supporting or opposing false claims and compare 23 user-and text-level features across the two groups. Contrary to the dominant assumption that negative emotion is a signature of falsehood, we find that anti-misinformation posts are more emotionally negative than pro-misinformation posts, with higher levels of anger, disgust, and sadness. These differences are modest in magnitude but consistent in direction across the negative emotions. We also find that posts opposing misinformation tend to come from more established users, i.e., older accounts, more followers, and higher listed counts.  \n1 Introduction  \nMisinformation on social media poses a significant threat to public discourse and informed decision-making, but not all online activity contributes to its spread. Many users actively push back, correcting falsehoods in their feeds [32], and most such organic correction comes from ordinary users rather than professional fact-checkers [36] . These corrections matter: peer correction can reduce belief in the underlying misinformation [1], thus platforms have formalized collective pushback through community-based fact-checking [44] .  \nThe implication is that combating misinformation requires more than suppressing false content. It also requires understanding the corrective content that emerges in response [1, 38, 44] . Characterizing the corrective populations—which we term the countermisinformation ecosystem—in both linguistic and account-level terms can provide insight into how misinformation is organically challenged on social media [30, 36] .  \nWe address this by first retrieving candidate posts for each factchecked claim, detecting which posts support or oppose it at scale, and then profiling the populations they form. For detection, we apply a Natural Language Inference (NLI) model described in our previous work [11] to a large corpus of COVID-19 tweets [8], yielding 264,737 model-classified tweets that either support or oppose COVID-19 misinformation. For profiling, we compare 23 user-and text-level features previously linked to misinformation spread, including emotion scores, bot-likelihood scores [4, 20, 21, 61], and toxicity measures [22, 40, 41, 45], across the pro-and antimisinformation tweets.1  \nCorrespondence to Eun Cheol Choi \u003C[euncheol@usc.edu](euncheol@usc.edu)>.  \nWe find that posts opposing misinformation are more emotionally charged than those spreading it, with higher levels of anger, disgust, and sadness. This runs against the dominant assumption that negative emotion is primarily a feature of false rather than corrective content [27, 34, 55, 58]; our study corroborates, at scale and with discrete-emotion measurement, prior observations on a few COVID-19 topics [36], alongside a study documenting emotional complexity across both misinformation and corrections during mass-shooting events [30] . The counter-misinformation ecosystem is also anchored by more established users (older accounts, more followers, and listed more often), and bot scores do not meaningfully differentiate the two sides.  \n2 Prior Work  \n2.1 How Users Counter Misinformation  \nBeyond how misinformati","cbCaiec0IWE0Kds9","https://ap.wps.com/l/cbCaiec0IWE0Kds9","pdf",1093542,4,1,7,"English","en",105,"# Introduction\n## How misinformation is challenged\n## Study approach and detection method\n# Prior Work\n## How users counter misinformation\n## Characteristics of misinformation-related content","[{\"question\":\"How did the study detect tweets that support or oppose misinformation?\",\"answer\":\"The researchers applied a domain-specific Natural Language Inference (NLI) model from prior work to a large COVID-19 tweet corpus, producing 264,737 model-classified tweets labeled as supporting or opposing false claims.\"},{\"question\":\"What emotional pattern does the study find in anti-misinformation vs pro-misinformation posts?\",\"answer\":\"Anti-misinformation posts are more emotionally negative than pro-misinformation posts, with higher levels of anger, disgust, and sadness. The differences are modest but consistent in direction across negative emotions.\"},{\"question\":\"What user-level characteristics distinguish posts opposing misinformation?\",\"answer\":\"Posts opposing misinformation tend to come from more established users, including older accounts, more followers, and higher listed counts. Bot-likelihood scores do not meaningfully differentiate the two sides.\"}]","Angry but Accurate: Detecting and Profiling the Counter-Misinformation Ecosystem on Twitter | PDF",1784176685,18,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":29},"angry-but-accurate-detecting-and-profiling-the-counter-misinformation-ecosystem-on-twitter","",{"@graph":37,"@context":86},[38,54,69],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,49,52],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":48},"https://docshare.wps.com/document/","Document",2,{"item":50,"name":12,"@type":44,"position":51},"https://docshare.wps.com/document/research-report/",3,{"item":53,"name":13,"@type":44,"position":20},"https://docshare.wps.com/document/angry-but-accurate-detecting-and-profiling-the-counter-misinformation-ecosystem-on-twitter/81857/",{"url":53,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":42,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-07-30","2026-07-16",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"How did the study detect tweets that support or oppose misinformation?","Question",{"text":76,"@type":77},"The researchers applied a domain-specific Natural Language Inference (NLI) model from prior work to a large COVID-19 tweet corpus, producing 264,737 model-classified tweets labeled as supporting or opposing false claims.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"What emotional pattern does the study find in anti-misinformation vs pro-misinformation posts?",{"text":81,"@type":77},"Anti-misinformation posts are more emotionally negative than pro-misinformation posts, with higher levels of anger, disgust, and sadness. The differences are modest but consistent in direction across negative emotions.",{"name":83,"@type":74,"acceptedAnswer":84},"What user-level characteristics distinguish posts opposing misinformation?",{"text":85,"@type":77},"Posts opposing misinformation tend to come from more established users, including older accounts, more followers, and higher listed counts. Bot-likelihood scores do not meaningfully differentiate the two sides.","https://schema.org",{"og:url":53,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":53},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":48,"doc_module":4,"doc_module_name":47,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":20,"doc_module":4,"doc_module_name":47,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":47,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":47,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":22,"doc_module":4,"doc_module_name":47,"category_name":117,"show_sort_weight":118,"slug":119},"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":47,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":47,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":47,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":47,"category_name":137,"show_sort_weight":107,"slug":138},19,"General","general"]