CoChat
Sign Up
Cited by 29Privacy, Security, and Data ProtectionMisinformation and Its ImpactsMedia and Digital Communication

What Do We Mean When We Talk about Trust in Social Media? A Systematic Review

Open access
Zhang, Yixuan, Gaggiano, Joseph D, Yongsatianchot, Nutchanon, Suhaimi, Nurul M, Kim, Miso, Sun, Yifan, Griffin, Jacqueline, Parker, Andrea G·2023·Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems
DOI 10.1145/3544548.3581019
Open this paper in CoChat — free accountOpen PDF

Abstract

Do people trust social media? If so, why, in what contexts, and how does that trust impact their lives? Researchers, companies, and journalists alike have increasingly investigated these questions, which are fundamental to understanding social media interactions and their implications for society. However, trust in social media is a complex concept, and there is conflicting evidence about the antecedents and implications of trusting social media content, users, and platforms. More problematic is that we lack basic agreement as to what trust means in the context of social media. Addressing these challenges, we conducted a systematic review to identify themes and challenges in this field. Through our analysis of 70 papers, we contribute a synthesis of how trust in social media is defined, conceptualized, and measured, a summary of trust antecedents in social media, an understanding of how trust in social media impacts behaviors and attitudes, and directions for future work.

AI snapshot

Written from the paper’s full text.

TL;DR

This systematic review synthesizes 70 empirical papers on trust in social media, showing that trust is inconsistently defined, measured, and conceptualized across the literature. It maps antecedents and consequences of trust, highlights mixed evidence, and recommends clearer definitions, better measurement, and more diverse methods.

Key findings

  • The review included 70 papers; 54 (77%) were published between 2017 and 2021. Surveys were the most common method (61 studies), followed by experiments (13) and interviews (6); Facebook was the most studied specific platform (13 studies, 19%), while 27 studies (39%) addressed social media generally.
  • Only 34 papers (49%) clearly defined trust concepts, and only 13 provided tailored social media definitions. Trust was most often conceptualized as willingness (n=12), followed by belief (n=6), feeling (n=4), confidence (n=2), expectation (n=2), perception (n=2), and decision (n=1). Only 47% of studies mapped definitions to measurements, and none specifically measured “willingness.”
  • Antecedents were grouped into demographic (n=13), trustor–trustee interaction (n=29), and trustee-related (n=21) categories. Evidence was mixed for age, gender, social media use, familiarity, and homophily; race/ethnicity, geo-location, and political ideology were important antecedents, while transparency, social credibility, and visual/emotional appeal generally increased trust.
  • Consequences were examined in 23 studies: trust positively influenced online behaviors such as social media use and information sharing, and offline event participation; it negatively influenced information verification; health behavior evidence was mixed. No paper solely examined distrust or mistrust; 16 mentioned both, and only 2 treated trust and distrust as distinct concepts.

Methods

  • Systematic review following PRISMA: searched six databases (ACM Digital Library, IEEE Xplore, SAGE Journals, ScienceDirect, Taylor & Francis Online, Emerald Insight) from October to November 2021 using Boolean strings combining social media platforms with “trust,” “distrust,” or “mistrust.”
  • Screening reduced 2,780 records to 70 included papers after duplicate removal, title screening (467 remaining), abstract screening (238 remaining), full-text screening, and quality assessment.
  • Data extraction covered study design, population, platform, definitions, measurements, antecedents, and consequences; two researchers extracted independently with Cohen’s kappa = 0.81, then used frequency analysis and clustering.

Limitations

  • Stated in the paper: the database search was scoped to six databases, so some relevant studies may have been missed; the corpus was restricted to English-language papers.
  • Evident from the text: most included studies used observational or survey methods, limiting causal conclusions; almost all were single time-point (67, 96%), only three had two waves, and none were longitudinal.
  • Evident from the text: many included papers lacked clear trust definitions or aligned measurements, which limits comparability and synthesis across studies.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1145/3544548.3581019

Cites

Statistics

Nothing to recompute.

Claims the paper makes

0 of 129 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Antecedents of trust identified in the review were organized into three categories: demographic antecedents (n=13), trustor-trustee interaction antecedents (n=29), and trustee-related antecedents (n=21).

    “As shown in Figure 5, antecedents of trust identifed within the papers of our review were organized into 1) trust antecedents related to specifc characteristics of the trustor, which we call demographic antecedents (n=13); 2) trust antecedents related to the interaction and/or relationship between trustor and trustee, which we call trustor-trustee interaction antecedents (n=29), and 3) trust antec”
    Quote in the paperNot a substantial claim
  • Findingown data

    Most papers did not name a specific trustee, instead opting to use the phrase 'another party' as a general term to describe the trustee.

    “Most papers did not name a specifc trustee, instead opting to use the phrase”
    Quote in the paperNot a substantial claim
  • Findingreasoning

    In studies that integrated multiple trustee measurements, the survey questions intended to measure a single trustee in fact ask about more than one trustee.

    “while these studies appear to use multiple survey questions to measure a single trustee, these questions in fact ask about more than one trustee.”
    Quote in the paperSubstantial claim
  • Findingown data

    Only 47% of the studies mapped the definitions and/or conceptualizations of trust to the measurements.

    “Our results showed that overall, only 47% of the studies mapped the”
    Quote in the paperSubstantial claim
  • Findingown data

    Other dimensions, such as race and ethnicity, income, and political affiliation, were less commonly reported.

    “In contrast, other dimensions, such as race and ethnicity, income, and political afliation, were less commonly reported.”
    Quote in the paperNot a substantial claim
  • May combine two claims. 60
  • Classified as probably not a substantial claim. 53
  • The quote only partly states this claim. 3

  • May combine two claims.

    Findingown data

    The majority of studies (n=67, 96%) were single time-point research; three studies (n=3, 4%) gathered data in two waves.

    “The majority of studies (n=67, 96%) were single time-point research (i.e., research carried out over a single period of time).Three studies (n=3, 4%) gathered data in two waves (i.e., time points) of measurement.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Among the papers that did define trust, trust has been most commonly conceptualized as willingness, with some conceptualizations framing trust as a belief, feeling, confidence, expectation, perception, or decision.

    “Among the papers that did defne it, trust has been most commonly conceptualized as willingness, with some conceptualizations framing trust as a belief, feeling, confdence, expectation, perception, or decision.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Trust has mostly been conceptualized as a kind of willingness on the part of the trustor (n=12), followed by a belief (n=6), feeling (n=4), confidence (n=2), expectation (n=2), perception (n=2), and decision (n=1).

    “Overall, trust has mostly been conceptualized as manifesting as a kind of willingness on the part of the trustor (n=12), followed by a belief (n=6), feeling (n=4), confdence (n=2), expectation (n=2), perception (n=2), and decision (n=1).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Survey methods were the most popular approach, followed by experimental studies; very few qualitative studies were conducted.

    “Overall, survey methods were the most popular approach, followed by experimental studies; very few qualitative studies were conducted.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Most measurements of trust in social media focused on a single trustee, while eight studies integrated multiple trustee measurements into one variable.

    “Most measurements of trust specifcally focused on a single trustee, except for eight studies that integrated multiple trustee measurements into one variable.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The vast majority of surveyed articles used a single research method (n=62, 89%), while the remaining used multiple methods (n=8, 11%).

    “The vast majority of surveyed articles used a single research method (n=62, 89%), while the remaining used multiple methods (n=8, 11%).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Among specific platforms examined, Facebook was by far the most commonly-reported individual platform, followed by Twitter, WeChat, Instagram, Weibo, YouTube, Snapchat, and others.

    “Among the specifc platform(s) examined, Facebook was by far the most commonly-reported individual platform, followed by Twitter, WeChat, Instagram, Weibo, YouTube, Snapchat, and others.”
    Quote in the paperNot a substantial claim

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

    Findingcitation

    A few of the dimensions used to measure trust in social media were directly borrowed from early trust literature, such as integrity and benevolence.

    “A few of these dimensions were directly borrowed from early trust literature, such as integrity (e.g., making good faith agreement) and benevolence (e.g., caring and being motivated to act in one's interest rather than acting opportunistically) [43].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [43]

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Most papers in the corpus were published between 2017 and 2021 (n=54, 77%); the rest were from 2009 to 2016 (n=16, 23%).

    “Most of the papers in our corpus were published in the last fve years between 2017 and 2021 (n=54, 77%); the rest were from 2009 to 2016 (n=16, 23%).”
    Quote in the paperNot a substantial claim

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

    Findingown data

    The review included 70 papers.

    “a total of 70 papers were included in our review”
    • Also stated in: § 2.6 Data Extraction and Analysis · plain
      “Records included in our systematic review (n = 70)”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Less than half of the papers in the corpus explicitly defined trust, and very few studies provided tailored definitions of trust in the context of social media, making the concept of trust ambiguous.

    “Less than half of the papers in our corpus explicitly defned trust, and very few studies provided tailored defnitions of trust in the context of social media, making the concept of trust ambiguous.”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    Only about half of the papers (n=34, 49%) provided clear definitions of trust concepts.

    “Among the papers that focused on trust in social media, only about half of them (n=34, 49%) provided clear defnitions of trust concepts”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Most studies (n=43, 61%) focused on a specific platform(s), while the rest addressed social media as a general concept (n=27, 39%).

    “Most studies (n= 43, 61%) focused on a specifc platform(s), while the rest addressed social media as a general concept (n=27, 39%)”
    Quote in the paperNot a substantial claim

  • May combine two claims. The quote only partly states this claim.

    Findingown data

    Other commonly-cited definitions of trust were from Moorman et al. (n=2) and Gefen et al. (n=2).

    “Other commonly-cited defnitions of trust were from Moorman et al. [75] who defned trust as”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [75] source

      Not checked in the free run.

    • cites [29] source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Studies that examined populations in North America were the most popular, followed by Asia and Europe; few studies focused on people in South America and Australia.

    “In terms of the study locations, studies that examined populations in North America were the most popular, followed by Asia and Europe; few studies focused on people in South America and Australia.”
    Quote in the paperNot a substantial claim

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

    Findingown data

    Of the papers that reported participants' sex or gender, sex (i.e., female, male) was far more assessed across studies than gender (i.e., women, men, etc.).

    “Of the papers that reported participants' sex or gender, sex (i.e., female, male) was far more assessed across studies than gender (i.e., women, men, etc.).”
    Quote in the paperNot a substantial claim

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

    Findingcitation

    Perceived critical mass is positively related to users' trust in the social media platform WeChat.

    “This study found that perceived critical mass was positively related to users' trust in the social media platform, WeChat [34].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [34] source

      Not checked in the free run.

  • May combine two claims.

    Findingcitation

    Perceived entertainment benefits do not influence trust in social media platforms, and perceived entertainment is not a predictor of trust in information on social media.

    “This same study also suggested that perceived entertainment benefts (e.g., browsing social media for fun or relaxation) did not infuence trust in social media platforms [50], and perceived entertainment was not a predictor of trust in information on social media.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [50] source

      Not checked in the free run.

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

    Findingcitation

    Privacy awareness positively influences users' trust in Facebook.

    “Finally, Malik et al. found that privacy awareness, or the cognizance and understanding of various aspects of privacy on social media, positively infuenced users' trust in Facebook [70].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [70] source

      Not checked in the free run.

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

    Findingcitation

    The level of detail of an advertisement on Facebook is not directly correlated with participants' trust in the Facebook platform as a whole.

    “And yet, another study found that the level of detail of an advertisement on Facebook was not directly correlated with participants' trust in the Facebook platform as a whole [23].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [23] source

      Not checked in the free run.

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

    Findingcitation

    Twitter users who believed their personality traits aligned with Twitter's online brand personality were more likely to trust Twitter in general.

    “In a diferent context, Pentina et al. [79] found that Twitter users who believed their personality traits aligned with Twitter's online brand personality were more likely to trust Twitter in general.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [79] source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Three studies focused on Black/African American populations, and one examined Asian populations.

    “Three studies focused on Black/African American populations, and one examined Asian populations.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation

    Facebook advertisements with a medium level of detail lead to more trust in the platform than advertisements with a high level of detail, especially if the detailed advertisement is highly personalized.

    “The study found that Facebook advertisements with a medium level of detail led to more trust in the platform than advertisements with a high level of detail (the assessment of levels of detail was determined by participants), especially if the detailed advertisement was highly personalized based on users' online activities (e.g., the websites they visited).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation

    Social ties are a significant trust antecedent, though one study disagreed.

    “Three studies in our corpus suggested that social ties were a signifcant trust antecedent [34,34,109], whereas one study in our corpus disagreed [26].”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [34] source

      Not checked in the free run.

    • cites [109] source

      Not checked in the free run.

    • cites [26] source

      Not checked in the free run.

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

    Findingcitation

    The timeliness of information has no significant effect on information trust in the context of COVID-19 information behaviors on social media.

    “However, one study focused on COVID-19 information behaviors on social media found that the timeliness of information had no signifcant efect on information trust [104].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [104] source

      Not checked in the free run.

  • May combine two claims.

    Findingcitation

    Gender was not a significant antecedent of trust in information, trust in online social media groups, or trust in social media platforms.

    “In general, our results show that gender was not a signifcant antecedent of trust in information [95,122], trust in online social media groups [69], nor trust in social media platforms [27].”
    Quote in the paperNot a substantial claim0 of 4 sources checked
    • cites [95] source

      Not checked in the free run.

    • cites [122] source

      Not checked in the free run.

    • cites [69] source

      Not checked in the free run.

    • cites [27] source

      Not checked in the free run.

  • May combine two claims.

    Findingcitation

    Perceived risks, risk-taking attitudes, and privacy-related risks have an impact on trust.

    “Work has explored several nuances that exist between risk factors and trust, including the impact that perceived risks, risk-taking attitudes, and privacy-related risks have on trust.”
    Quote in the paperNot a substantial claim

  • May combine two claims. The quote only partly states this claim.

    Findingcitation

    Higher levels of perceived benefits, such as perceived entertainment benefits and/or social support from social media, help increase trust in social media platforms.

    “Higher levels of perceived benefts, such as perceived entertainment benefts and/or social support from social media, help increase trust in social media platforms.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation

    Geo-location is an antecedent of trust in social media platforms, indicating that cultural differences play a role in influencing trust in social media.

    “Three studies found geo-location to be an antecedent of trust in social media platforms [66,79,95], indicating that cultural diferences play a role in infuencing trust in social media.”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites [66] source

      Not checked in the free run.

    • cites [79] source

      Not checked in the free run.

    • cites [95] source

      Not checked in the free run.

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

    Findingcitation

    Facebook pages containing detailed product information significantly increase users' trust in that page.

    “For example, in the context of e-commerce, Sari et al. [89] found that Facebook pages containing detailed product information signifcantly increased users' trust in that page.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [89] source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Only two papers in the corpus specifically examined both trust and distrust, and both considered trust and distrust as two distinct concepts.

    “Only two papers [16,40] specifcally examined both trust and distrust, and both studies considered trust and distrust as two distinct concepts.”
    Quote in the paperNot a substantial claim

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

    Findingnone given

    Trust in social media negatively influences users' information verification behaviors (the probability of verifying fake news seen on social media).

    “However, trust in social media negatively infuences users' information verifcation behaviors (i.e., the probability of verifying fake news seen on social media).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Race and ethnicity, geo-location, and political ideology were found to be important antecedents of trust across studies.

    “Unlike age and gender, race and ethnicity, geo-location, and political ideology were found to be important antecedents of trust across studies.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    Group size is the most significant predictor of trust in Facebook groups, with trust inversely correlated with group size.

    “Their study found that group size was the most signifcant predictor of trust in Facebook groups, as trust was inversely correlated with the group size.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [69] source

      Not checked in the free run.

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

    Findingcitation

    Trust in information on social media does not affect user verification behaviors, but it decreases users' probability of verifying fake news on social media.

    “One study found that trust in information on social media does not afect user verifcation behaviors, though this trust decreases users' probability of verifying fake news on social media [4].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [4] source

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Among trustor-trustee interaction antecedents, social media use was the most prevalent factor examined, followed by prior experiences and attitudes, familiarity, and risk factors; fewer studies explored social ties, homophily, and perceived benefits.

    “Furthermore, among the trustor-trustee interaction antecedents, social media use was the most prevalent factor examined, followed by prior experiences and attitudes, familiarity with the trustee, and risk factors; fewer studies explored social ties, homophily, and perceived benefts as antecedents of trust in social media.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    More than half of the surveyed papers did not provide clear definitions of trust and/or did not contextualize the trustee in the context of social media.

    “Our results show that more than half of the surveyed papers did not provide clear defnitions of trust and/or did not contextualize the trustee in the context of social media.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    Cheng et al. [16] assessed trust in Facebook by measuring integrity and competence, and assessed distrust in Facebook by measuring skepticism and vulnerability.

    “Cheng et al. [16] assessed trust in Facebook by measuring integrity and competence, while also assessing distrust in Facebook by measuring skepticism and vulnerability.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [16] source

      Not checked in the free run.

  • May combine two claims.

    Findingnone given

    Transparency, social credibility of information, and visual and emotional appeal embedded into information overall positively influenced trust in social media information.

    “Finally, regarding the trustee-related antecedents, factors such as transparency (availability of URLs to information sources), social credibility of information, and visual and emotional appeal embedded into information overall positively infuenced trust in social media information.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Findingnone given

    Consequences of trust in social media can be categorized into three broad dimensions: online behaviors, offline behaviors, and attitudes spurred by trust.

    “These consequences can be categorized into three broad dimensions: online behaviors, ofine behaviors, and attitudes spurred by trust, as shown in Figure 6.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Trust in social media significantly positively impacts people's online behaviors and offline participation in events.

    “Our fndings show that trust in social media signifcantly positively impacts people's online behaviors (e.g., social media use frequency, intention to share information and news on social media) and ofine participation in events.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    Trust in information on social media and trust in social media platforms both play a role in increasing people's use of social media.

    “The studies in our corpus show that trust in the information on social media [2,65], as well as trust in the social media platforms themselves [33,79], both play a role in increasing people's use of social media.”
    Quote in the paperSubstantial claim0 of 4 sources checked
    • cites [2] source

      Not checked in the free run.

    • cites [65] source

      Not checked in the free run.

    • cites [33] source

      Not checked in the free run.

    • cites [79] source

      Not checked in the free run.

  • May combine two claims.

    Findingcitation

    Trust in information on social media and trust in the platforms themselves positively influence people's intention to share information, news, and opinions on social media.

    “All the studies in our corpus suggest that trust in the information on social media [10,56,65,68] and trust in the platforms themselves [30,59,66,70] positively infuence people's intention to share information, news, and opinions on social media [122].”
    Quote in the paperSubstantial claim0 of 9 sources checked
    • cites [10] source

      Not checked in the free run.

    • cites [56] source

      Not checked in the free run.

    • cites [65] source

      Not checked in the free run.

    • cites [68] source

      Not checked in the free run.

    • cites [30] source

      Not checked in the free run.

    • cites [59] source

      Not checked in the free run.

    • cites [66] source

      Not checked in the free run.

    • cites [70] source

      Not checked in the free run.

    • cites [122] source

      Not checked in the free run.

  • May combine two claims.

    Findingnone given

    Social group category and platform differences influence people's trust in information from social media platforms.

    “Studies also found that social group category and platform diferences infuence people's trust in information from social media platforms.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Prior positive experiences and attitudes have been shown to positively influence trust in social media across studies, but there is mixed evidence as to whether other trustor-trustee interaction factors are truly trust antecedents.

    “Besides prior (positive) experiences and attitudes, which have been shown to positively infuence trust in social media across studies, there is mixed evidence as to whether these trustor-trustee interaction factors are truly trust antecedents.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    Group category is an antecedent of trust; users often trusted family/friend groups more than interest- or location-based groups.

    “Finally, the group category was shown to be another antecedent of trust, as users often trusted family/friend groups more than interest-or location-based groups.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingnone given

    The smaller the social group and the longer the group tenure, the more trust users placed in the social group.

    “In terms of social group attributes, the smaller the social group and the longer the group tenure, the more trust users placed in the social group.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Among the 16 papers mentioning both trust and distrust, most (n=13, 81%) viewed trust and distrust as two extremes of the same dimension and did not specifically examine distrust.

    “Among these 16 papers, most of them (n=13, 81%) viewed trust and distrust as two extremes of the same dimension and did not specifcally examine distrust (only briefy mentioned distrust).”
    • Also stated in: § A Conceptual Framework of Trust in Social Media · plain
      “On top of that, of the papers that did mention both trust and distrust, most considered trust and distrust to be two opposites of one concept.”
    Quote in the paperSubstantial claim

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

    Findingcitation

    Trust in Facebook significantly impacts privacy concerns about Facebook.

    “Specifcally, Kumar et al. [59] suggested that trust in Facebook signifcantly impacted privacy concerns about Facebook.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [59] source

      Not checked in the free run.

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

    Findingcitation

    One study in the corpus considered mistrust to be the opposite of trust.

    “Furthermore, we found that one study [5] considered mistrust to be the opposite of trust.”
    Quote in the paperNot a substantial claim

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

    Findingnone given

    There is mixed evidence as to how trust in social media influences people's health behaviors.

    “Additionally, there is mixed evidence as to how trust in social media infuences people's health behaviors.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingcitation

    Although timely information positively influences people's trust in information from social media, the timeliness of a social media post had no significant effect on people's trust in that post during COVID-19.

    “Additionally, our review has also shown that although timely information positively infuences people's trust in information from social media, this timeliness of a social media post had no signifcant efect on people's trust in that post during COVID-19 [104].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [104] source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Findingown data

    The review summarizes a set of personal factors that influence people's trust towards information on social media, social media users and groups, and the platforms themselves.

    “Our review summarizes a set of personal factors that infuence people's trust towards the information found on social media, social media users and groups, and the platforms themselves.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Findingnone given

    The review provides insights into factors that promote trust in information on social media, other users on the platform, and the social media platforms themselves.

    “Our work provides insights into the factors that promote trust in information on social media, other users on the platform, and the social media platforms themselves.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Only three studies in the corpus conducted two-wave measurement, and no studies adopted a longitudinal approach to understanding temporal facets of trust in social media.

    “In contrast, only three studies conducted the two-wave measurement, and no studies in our corpus have adopted a longitudinal approach to understanding the temporal facets of trust in the context of social media.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingnone given

    Some studies suggested that higher usage of social media increases trust in information found on social media, whereas other studies found the exact opposite relationship or no significant relationship at all.

    “Specifcally, when examining the antecedents of trust, some studies suggested that higher usage of social media increases trust in the information found on social media, whereas other studies either found the exact opposite relationship or no signifcant relationship at all.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingnone given

    Users' trust in social media platforms is influenced by the credibility and quality of information on social media; this trust consequently influences users' future engagement with those platforms.

    “Our fndings demonstrate how users' trust in social media platforms is infuenced by the credibility and quality of information on social media; this trust (or lack thereof) consequently infuences users' future engagement with those platforms.”
    Quote in the paperSubstantial claim

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

    Study design and conductnone given

    Each article was evaluated by two researchers independently, and any disagreement regarding quality was discussed until consensus was achieved.

    “Each article was evaluated by two researchers independently.Any disagreement regarding the quality of the article was discussed until consensus was achieved.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    Prior work found that people who held a higher level of trust in health information on social media tended to engage in health behaviors, such as preventive behaviors like avoiding crowds during the COVID-19 pandemic.

    “prior work has found that people who held a higher level of trust in health information on social media tended to engage in health behaviors (e.g., preventive behaviors, like avoiding crowds during the COVID-19 pandemic) [112].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [112] source

      Not checked in the free run.

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

    Study design and conductnone given

    Two researchers independently extracted information from each paper and recorded it into spreadsheets.

    “Two researchers independently extracted the information from each paper and recorded it into spreadsheets.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    The majority of studies examining antecedents of trust in the corpus used observational or survey methodology, which limits causal conclusions.

    “Note that the majority of studies examining the antecedents of trust in our corpus utilize observational or survey methodology, limiting any causal conclusions that can be drawn from the research.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    Two researchers closely examined the remaining 238 articles to determine final inclusion according to inclusion and exclusion criteria.

    “Two researchers then closely examined the remaining 238 articles to determine the fnal inclusion of papers according to inclusion and exclusion criteria constructed by the researchers.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    The databases were chosen because they are well-known digital libraries within social sciences and social media and contain a large amount of research related to trust in social media.

    “These databases were chosen because they are well-known digital libraries within the felds of social sciences and social media and contain a large amount of research related to trust in social media.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    The research team held regular meetings to validate the inclusion and exclusion of articles.

    “The research team held regular meetings (e.g., weekly meetings and day-to-day online communication) to validate the inclusion and exclusion of articles.”
    Quote in the paperNot a substantial claim

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

    Study design and conductown data

    Three papers were marked as 'poor' based on critical appraisal tools and were removed from the corpus.

    “Three papers were marked as being”
    Quote in the paperNot a substantial claim

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

    Study design and conductown data

    A high inter-rater reliability, with a Cohen's kappa coefficient of 0.81, was achieved.

    “A high inter-rater reliability (IRR), with a Cohen's kappa () coefcient of 0.81, was achieved.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Premisecitation

    Trust is multifaceted and is a notoriously complex and nuanced concept.

    “trust is multifaceted and is a notoriously complex and nuanced concept [64]”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [64]

      Not checked in the free run.

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

    Premisecitation

    Trust is a multi-dimensional and complex construct.

    “Trust is a multi-dimensional and complex construct [64]”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [64]

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Trust might be better understood through multiple dimensions.

    “indicating that trust might be better understood through multiple dimensions.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    The two researchers compared and discussed discrepancies in data extraction to achieve a consensus.

    “Then the two researchers compared and discussed discrepancies in data extraction to achieve a consensus.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Other work has shown that trust in the information found on social media negatively influences people's health behaviors, for example by increasing vaccine hesitancy.

    “other work has shown that trust in the information found on social media negatively infuences people's health behaviors (e.g., by increasing vaccine hesitancy) [18].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [18]

      Not checked in the free run.

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

    Interpretationreasoning

    A potential explanation for the discrepancy in results between the two studies is differences in measurements.

    “One potential explanation for the discrepancy in results from these two studies may be the diferences in measurements.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationcitation

    Multi-media and visual components invoke positive emotional sentiments among users, which contributes to increased trust in that information.

    “This is because these components often invoke positive emotional sentiments among users, which contributes to increased trust in that information [13,40,56].”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites [13] source

      Not checked in the free run.

    • cites [40] source

      Not checked in the free run.

    • cites [56] source

      Not checked in the free run.

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

    Interpretationcitation

    Visually- and emotionally-appealing stimuli better satisfy users' media needs, leading to more trust, according to gratification theory.

    “According to the authors, this increase in trust can be attributed to gratifcation theory, as visually-and emotionally-appealing stimuli better satisfy users' media needs, leading to more trust [56].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [56] source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Information shared by others in one's network is more likely to be perceived as credible and less likely to be personally verified.

    “The authors explained that perceived social credibility decreases users' verifcation behaviors online, as information shared by others in one's network is more likely to be perceived as credible and, thus, less likely to be personally verifed.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    When people trust a political system, they have accompanying feelings of confidence in, and commitment toward, that system.

    “Specifcally, when people trust a political system, they have accompanying feelings of confdence in, and commitment toward, that system.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [48] source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    Most studies that examined the consequences of trust in social media used observational or survey methodology, which limits causal conclusions.

    “Similar to what we mentioned about the fndings on trust antecedents, it is important to note that most studies that examined the consequences of trust in social media utilized observational or survey methodology, which ultimately limits any causal conclusions that can be drawn from the research.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    The directionality of the relationship between social media use and trust in information on social media may be important to consider.

    “This suggests that the directionality of this relationship may be important to consider and should be further assessed in future research.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    There is no 'one size fits all' in trust and social media research; distinct features, characteristics, and limitations of each platform need to be more heavily considered.

    “In essence, the distinct features, characteristics, and limitations present among each individual platform need to be more heavily considered in the future, as there is no”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Nuanced features and characteristics of the trustee are essential to unpack to better understand trust.

    “Nuanced features and characteristics of the trustee, such as the specifc social media platform used, the nature of the social media group, and the characteristics of the individuals sharing the information online, are essential to unpack to better understand trust.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    The notion of generalized trust in social media is called into question because users' trust across platforms can drastically differ.

    “These results collectively call into question the notion of generalized trust in social media, as users' trust across platforms can drastically differ.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    In political trust, Jennings et al. [48] differentiate trust and distrust by describing diverging attitudes and behaviors.

    “For example, in the context of political trust, Jennings et al. [48] diferentiate these concepts by describing the diverging attitudes and behaviors that accompany trust versus distrust.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [48] source

      Not checked in the free run.

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

    Premisecitation

    Trust may inspire loyalty to a political system, while distrust can foster disengagement from or rebellion towards that system.

    “While trust may inspire loyalty to a system, distrust can foster disengagement from, or even rebellion towards, that same political system.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [48] source

      Not checked in the free run.

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

    Study design and conductreasoning

    Although measures were taken to ensure the databases contained a majority of relevant studies, it is possible that some studies were not included.

    “Although measures were taken to ensure that these databases contained a majority of the studies available that focus on trust in social media, it is possible that some studies were not included in the selected databases.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    Trust transference has been explored in other contexts, such as business and online banking.

    “Trust transference has been explored in other contexts, such as in business and online banking [37,98].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [37] source

      Not checked in the free run.

    • cites [98] source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Trust can impact the use of social media and could, in turn, lead to a change in trust antecedents, such as a change in social media use.

    “As our fndings have shown, trust can impact the use of social media and could, in turn, lead to a change in trust antecedents, such as a change in social media use.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    Distrust in political systems is accompanied by negative emotions such as contempt, insecurity, fear, anger, alienation, and cynicism.

    “Rather, distrust in political systems is accompanied by negative emotions, such as contempt, insecurity, fear, anger, alienation, and cynicism [48].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [48] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationnone given

    The paper's recommendations will help increase the validity of trust research as researchers clearly establish what they mean by trust in social media.

    “Our recommendations will help to increase the validity of trust research moving forward as researchers begin to clearly establish what they truly mean when they talk about trust in social media.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Proper translation of trust into assessment is paramount to ensure construct validity.

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

    Premisecitation

    Distrust in political systems comes with a separate set of attitudes that are more than just the opposite of the attitudes accompanying trust.

    “In contrast, these same researchers explain that distrust comes with a separate set of attitudes that are more than just the opposite of the attitudes that accompany trust [48].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [48] source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Trust can evolve over time, shaping users' behaviors and their perception of social media.

    “This reciprocal relationship highlights that trust can evolve over time, shaping users' behaviors and their perception of social media.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    If users have seen a large amount of misinformation on social media, they could lose trust in that particular platform.

    “For instance, if users have seen a large amount of misinformation on social media, they could lose trust in that particular platform [118].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [118] source

      Not checked in the free run.

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

    Premisecitation

    These strategies may also be used for nefarious purposes, such as inflating the perceived credibility of misinformation and disinformation online.

    “Unfortunately, these strategies may also be used for more nefarious purposes, such as infating the perceived credibility of misinformation and disinformation online [108].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [108]

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Hispanic communities tend to trust doctors and community leaders of the same ethnic background, but are less likely than White populations to trust the government and social media influencers in the context of COVID-19.

    “For example, Hispanic communities tend to trust doctors and community leaders of the same ethnic background, but they are less likely than White populations to trust the government and social media infuencers in the context of COVID-19 [58].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [58] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Premisecitation

    Excluding marginalized groups from study samples is fundamentally unethical and can lead to biased findings and systems with embedded bias.

    “Excluding these groups from study samples is fundamentally unethical and can lead to biased fndings and, subsequently, the development of systems with embedded bias [117].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [117] source

      Not checked in the free run.

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

    Study design and conductnone given

    The database search for this review was scoped to include six databases.

    “One limitation of this work is that our database search was scoped to include six databases.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    The corpus is restricted to English-language papers.

    “Additionally, our corpus is restricted to English-language papers.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Social media companies have been criticized for failing to combat the spread of misinformation even though they have promised to take action.

    “However, social media companies have been criticized for failing to combat the spread of misinformation (e.g., failing to label and remove misinformation) even though they have promised to take action to address these issues [28].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [28]

      Not checked in the free run.

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

    Premisecitation

    Recent research has suggested that utilizing crowdsourced data to identify misinformation-producing sources and imputing this data into social media ranking algorithms has promise for reducing misinformation on social media platforms.

    “For example, recent research has suggested that utilizing crowdsourced data (e.g., general public rating of information outlets) to identify misinformation-producing sources and imputing this data into social media ranking algorithms that rank content within a platform has promise for reducing the amount of misinformation on social media platforms [24].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [24] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing). May combine two claims.

    Premisecitation

    Some trust scholars have argued that mistrust is misplaced trust whereby a trustee defaults or betrays a person's trust, and mistrust does not imply the opposite of trust.

    “However, some trust scholars have argued that mistrust is misplaced trust whereby a trustee”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [72] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Culturally-competent approaches, such as enlisting community-trusted Hispanic doctors and community leaders to develop and disseminate public health information, may be much more effective than current, more general approaches.

    “These trends suggest that culturallycompetent approaches [116], such as enlisting community-trusted Hispanic doctors and community leaders to develop and disseminate public health information, may be much more efective than current, more general approaches.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [116] source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationreasoning

    Building trust with users is a complex task that requires effort to ensure that these attributes are fully realized.

    “As such, building trust with users is a complex task that requires efort to ensure that these attributes are fully realized.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    An emerging body of work has begun to explore the use of automated conversational agents and language models to identify unique traits in social media users' communication styles.

    “In addition, an emerging body of work has begun to explore the use of automated conversational agents and language models to identify unique traits in social media users' communication styles [93].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [93] source

      Not checked in the free run.

102 of 102

  • The paper relies on this without support.

    There is a lack of a synthesized understanding of the learnings, challenges, open questions, and implications for future work across the body of research on trust in social media.

    “Despite a large body of work focused on trust in social media, we lack a synthesized understanding of the learnings, challenges, open questions, and implications for future work across this body of research.”

  • The paper relies on this without support.

    There is a lack of a synthesized understanding regarding the antecedents of trust in social media.

    “However, there is a lack of a synthesized understanding regarding these antecedents of trust in social media.”

  • The paper relies on this without support.

    It is unclear the extent to which antecedents of trust identified in other contexts can be applied to social media, and what other unique antecedents may exist in this context.

    “Specifcally, it is unclear the extent to which the antecedents of trust identifed in other contexts can be applied to social media, and what other unique antecedents may exist in this context.”

  • The paper relies on this without support.

    There is a need for a comprehensive synthesis and analysis of the literature regarding how trust impacts people's behaviors and attitudes.

    “Given the varied and conficting evidence regarding the consequences of trust, there is a need for a comprehensive synthesis and analysis of this body of literature regarding how trust impacts people's behaviors and attitudes.”

  • The paper relies on this without support.

    Proper translation of trust into assessment is paramount to ensure construct validity.

    “Furthermore, we advocate for future research to ensure that trust defnitions align with and guide the trust measurements used in each study, as proper translation of trust into assessment is paramount to ensure construct validity.”

  • The paper relies on this without support.

    Subjugating communities to unwanted research is detrimental to their well-being and autonomy.

    “After all, subjugating these communities to unwanted research is detrimental to their well-being and autonomy.”

  • The paper relies on this without support.

    Choosing a proper conceptualization and definition of trust is largely dependent upon the research goals.

    “Choosing a proper conceptualization and defnition is largely dependent upon the research goals.”

  • The paper relies on this without support.

    By specifying the precise conceptualization and definition of trust, more pointed research questions arise.

    “By specifying the precise conceptualization and defnition of trust, more pointed research questions arise.”
8 of 8

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Citation for Woko et al. points to different authors

    Minor

    The text attributes this finding to Woko et al. and cites reference 111, but the entry numbered 111 is by Wu and Shen.

  • Social-ties citations repeat one study and omit another

    Minor

    The first passage says three studies support the claim but cites reference 34 twice, yielding only two distinct references. The later passage identifies both references 33 and 34 as relevant Gong et al. studies.

2 of 2

The citation lists haven’t been fetched yet

References 109 · Cited by 29

Research works printed but not linked: 4

  1. Kirsimarja Blomqvist. 1997. The many faces of trust. Scandinavian journal of management 13, 3 (1997), 271–286.
  2. Kenneth Bordens and Bruce Barrington Abbott. 2014. Ebook: Research Design and Methods: A Process Approach.
  3. Shannon E Clark, Megan C Bledsoe, and Christopher J Harrison. 2022. The role of social media in promoting vaccine hesitancy. Current opinion in pediatrics 34, 2 (2022), 156–162.
  4. Mark S Granovetter. 1973. The strength of weak ties. American journal of sociology 78, 6 (1973), 1360–1380.

Authors
Zhang, Y., Gaggiano, J. D., Yongsatianchot, N., Suhaimi, N. M., Kim, M., Sun, Y., Griffin, J., & Parker, A. G.
Affiliations
  • Georgia Institute of Technology (US)
  • Northeastern University (US)
  • Universiti Malaysia Pahang Al-Sultan Abdullah (MY)
  • William & Mary (US)
ORCID
Venue
Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems
Publisher
ACM
Published
2023-04-19
Pages
1-22
DOI
10.1145/3544548.3581019
OpenAlex
W4319654023
Cited by
29
References
109
Citations by year
2026: 11 · 2025: 11 · 2024: 5 · 2023: 2
FWCI
1.44
Type
journal article
Language
en
Open access
Licence
Locations
  • CHI Conference on Human Factors in Computing Systems (publishedVersion, cc-by-nd)
  • arXiv (Cornell University) (submittedVersion, cc-by)
  • arXiv (Cornell University) (cc-by)
Topics
  • Social Sciences › Sociology and Political Science › Privacy, Security, and Data Protection
  • Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
  • Physical Sciences › Computer Science › Computer Networks and Communications › Media and Digital Communication
Keywords
trust in social media, antecedents of trust, trust measurement

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-10.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-10.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review