Search

In this study, we conducted an online, between-subjects experiment (N = 2,049) to better understand the impact of warning label sources on information trust and sharing intentions. Across four warners (the social media platform, other social media users, Artificial Intelligence (AI), and fact checkers), we found that all significantly decreased trust in false information relative to control, but warnings from AI were modestly more effective. All warners significantly decreased the sharing intentions of false information, except warnings from other social media users. AI was again the most effective. These results were moderated by prior trust in media and the information itself. Most noteworthy, we found that warning labels from AI were significantly more effective than all other warning labels for participants who reported a low trust in news organizations, while warnings from AI were no more effective than any other warning label for participants who reported a high trust in news organizations., Comment: To appear at ICWSM 2025

In this study, we examine the perceived influence of others, across both strong and weak social ties, on COVID-19 vaccination decisions in the United States. We add context to social influence by measuring related concepts, such as perceived agreement of others and perceived danger of COVID-19 to others. We find that vaccinated populations perceived more influence from their social circles than unvaccinated populations. This finding holds true across various social groups, including family, close friends, and neighbors. Vaccinated participants perceived that others agreed with their decision to get vaccinated more than unvaccinated participants perceived others to agree with their decision to not get vaccinated. Despite the clear differences in perceived social influence and agreement across the groups, the majority of participants across both vaccinated and unvaccinated populations perceived no social influence from all social group in their decisions. Aligning with this result, we find through open-ended responses that both vaccinated and unvaccinated participants frequently cited fear as a motivating factor in their decision, rather than social influence: vaccinated participants feared COVID-19, while unvaccinated participants feared the vaccine itself., Comment: Preprint of paper currently under review

Pink slime news outlets automatically produce low-quality, often partisan content that is framed as authentic local news. Given that local news is trusted by Americans and is increasingly shutting down due to financial distress, pink slime news outlets have the potential to exploit local information voids. Yet, there are gaps in understanding of pink slime production practices and tactics, particularly over time. Hence, to support future research in this area, we built a dataset of over 7.9M articles from 1093 pink slime sources over 2.5 years., Comment: published at ICWSM 2024 Dataset Track

This study examines the use of embedded tweets in online news media. In particular, we add to the previous literature by exploring embedded tweets across reliable and unreliable news outlets. We use a mixed-method analysis to examine how the function and frequency of embedded tweets change across outlet reliability and news topic. We find that, no matter the outlet reliability, embedded tweets are most often used to relay the opinions of elites, to syndicate information from another news source, or to self-cite information an outlet previously produced. Our results also show some notable differences between reliable media and fringe media's use of tweets. Namely, fringe media embed tweets more and use those tweets as the source of news more than reliable media. Our work adds to the literature on hybrid media systems and the normalization of social media in journalism., Comment: MeLa Lab Preliminary Findings Report

A concern among content moderation researchers is that hard moderation measures, such as banning content producers, will push users to more extreme information environments. Research in this area is still new, but predominately focuses on one-way migration (from mainstream to alt-tech) due to this concern. However, content producers on alt-tech social media platforms are not always banned users from mainstream platforms, instead they may be co-active across platforms. We explore co-activity on two such platforms: YouTube and BitChute. Specifically, we describe differences in video production across 27 co-active channels. We find that the majority of channels use significantly more moral and political words in their video titles on BitChute than in their video titles on YouTube. However, the reasoning for this shift seems to be different across channels. In some cases, we find that channels produce videos on different sets of topics across the platforms, often producing content on BitChute that would likely be moderated on YouTube. In rare cases, we find video titles of the same video change across the platforms. Overall, there is not a consistent trend across co-active channels in our sample, suggesting that the production on alt-tech social media platforms does not fit a single narrative., Comment: This is a MeLa Lab Technical Report

In order to better support researchers, journalist, and practitioners in their use of the MeLa-BitChute dataset for exploration and investigative reporting, we provide new psycho-linguistic metadata for the videos, comments, and channels in the dataset using LIWC22. This paper describes that metadata and methods to filter the data using the metadata. In addition, we provide basic analysis and comparison of the language on BitChute to other social media platforms. The MeLa-BitChute dataset and LIWC metadata described in this paper can be found at: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/KRD1VS., Comment: This paper is a Metadata Supplement to The MeLa BitChute Dataset

In this paper, we present a dataset of over 1.4M online news articles from 313 local U.S. news outlets published over 20 months (between April 4th, 2020 and December 31st, 2021). These outlets cover a geographically diverse set of communities across the United States. In order to estimate characteristics of the local audience, included with this news article data is a wide range of county-level metadata, including demographics, 2020 Presidential Election vote shares, and community resilience estimates from the U.S. Census Bureau. The NELA-Local dataset can be found at: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/GFE66K., Comment: Published at ICWSM 2022

In this paper, we present the fifth installment of the NELA-GT datasets, NELA-GT-2022. The dataset contains 1,778,361 articles from 361 outlets between January 1st, 2022 and December 31st, 2022. Just as in past releases of the dataset, NELA-GT-2022 includes outlet-level veracity labels from Media Bias/Fact Check and tweets embedded in collected news articles. The NELA-GT-2022 dataset can be found at: https://doi.org/10.7910/DVN/AMCV2H, Comment: Technical report documenting the NELA-GT recent update (NELA-GT-2022). arXiv admin note: substantial text overlap with arXiv:2102.04567

In this paper, we present an updated version of the NELA-GT-2019 dataset, entitled NELA-GT-2020. NELA-GT-2020 contains nearly 1.8M news articles from 519 sources collected between January 1st, 2020 and December 31st, 2020. Just as with NELA-GT-2018 and NELA-GT-2019, these sources come from a wide range of mainstream news sources and alternative news sources. Included in the dataset are source-level ground truth labels from Media Bias/Fact Check (MBFC) covering multiple dimensions of veracity. Additionally, new in the 2020 dataset are the Tweets embedded in the collected news articles, adding an extra layer of information to the data. The NELA-GT-2020 dataset can be found at https://doi.org/10.7910/DVN/CHMUYZ., Comment: 6 pages, 4 figures. arXiv admin note: text overlap with arXiv:2003.08444

Stopping the malicious spread and production of false and misleading news has become a top priority for researchers. Due to this prevalence, many automated methods for detecting low quality information have been introduced. The majority of these methods have used article-level features, such as their writing style, to detect veracity. While writing style models have been shown to work well in lab-settings, there are concerns of generalizability and robustness. In this paper, we begin to address these concerns by proposing a novel and robust news veracity detection model that uses the content sharing behavior of news sources formulated as a network. We represent these content sharing networks (CSN) using a deep walk based method for embedding graphs that accounts for similarity in both the network space and the article text space. We show that state of the art writing style and CSN features make diverse mistakes when predicting, meaning that they both play different roles in the classification task. Moreover, we show that the addition of CSN features increases the accuracy of writing style models, boosting accuracy as much as 14\% when using Random Forests. Similarly, we show that the combination of hand-crafted article-level features and CSN features is robust to concept drift, performing consistently well over a 10-month time frame., Comment: Preprint Version

In this paper, we characterize the content and discourse on BitChute, a social video-hosting platform. Launched in 2017 as an alternative to YouTube, BitChute joins an ecosystem of alternative, low content moderation platforms, including Gab, Voat, Minds, and 4chan. Uniquely, BitChute is the first of these alternative platforms to focus on video content and is growing in popularity. Our analysis reveals several key characteristics of the platform. We find that only a handful of channels receive any engagement, and almost all of those channels contain conspiracies or hate speech. This high rate of hate speech on the platform as a whole, much of which is anti-Semitic, is particularly concerning. Our results suggest that BitChute has a higher rate of hate speech than Gab but less than 4chan. Lastly, we find that while some BitChute content producers have been banned from other platforms, many maintain profiles on mainstream social media platforms, particularly YouTube. This paper contributes a first look at the content and discourse on BitChute and provides a building block for future research on low content moderation platforms., Comment: This long paper is supplemental to a short version of the paper published in ACM Conference on Hypertext and Social Media 2020

In this paper, we present an updated version of the NELA-GT-2018 dataset (N{\o}rregaard, Horne, and Adal{\i} 2019), entitled NELA-GT-2019. NELA-GT-2019 contains 1.12M news articles from 260 sources collected between January 1st 2019 and December 31st 2019. Just as with NELA-GT-2018, these sources come from a wide range of mainstream news sources and alternative news sources. Included with the dataset are source-level ground truth labels from 7 different assessment sites covering multiple dimensions of veracity. The NELA-GT-2019 dataset can be found at: https://doi.org/10.7910/DVN/O7FWPO, Comment: Updated dataset for paper NELA-GT-2018: A Large Multi-Labelled News Dataset for The Study of Misinformation in News Articles, originally published at ICWSM in 2019

Research in combating misinformation reports many negative results: facts may not change minds, especially if they come from sources that are not trusted. Individuals can disregard and justify lies told by trusted sources. This problem is made even worse by social recommendation algorithms which help amplify conspiracy theories and information confirming one's own biases due to companies' efforts to optimize for clicks and watch time over individuals' own values and public good. As a result, more nuanced voices and facts are drowned out by a continuous erosion of trust in better information sources. Most misinformation mitigation techniques assume that discrediting, filtering, or demoting low veracity information will help news consumers make better information decisions. However, these negative results indicate that some news consumers, particularly extreme or conspiracy news consumers will not be helped. We argue that, given this background, technology solutions to combating misinformation should not simply seek facts or discredit bad news sources, but instead use more subtle nudges towards better information consumption. Repeated exposure to such nudges can help promote trust in better information sources and also improve societal outcomes in the long run. In this article, we will talk about technological solutions that can help us in developing such an approach, and introduce one such model called Trust Nudging., Comment: Published at IEEE TPS 2019

In this paper, we present a dataset of 713k articles collected between 02/2018-11/2018. These articles are collected directly from 194 news and media outlets including mainstream, hyper-partisan, and conspiracy sources. We incorporate ground truth ratings of the sources from 8 different assessment sites covering multiple dimensions of veracity, including reliability, bias, transparency, adherence to journalistic standards, and consumer trust. The NELA-GT-2018 dataset can be found at https://doi.org/10.7910/DVN/ULHLCB., Comment: Published at ICWSM 2019

In this paper, we analyze content sharing between news sources in the alternative and mainstream media using a dataset of 713K articles and 194 sources. We find that content sharing happens in tightly formed communities, and these communities represent relatively homogeneous portions of the media landscape. Through a mix-method analysis, we find several primary content sharing behaviors. First, we find that the vast majority of shared articles are only shared with similar news sources (i.e. same community). Second, we find that despite these echo-chambers of sharing, specific sources, such as The Drudge Report, mix content from both mainstream and conspiracy communities. Third, we show that while these differing communities do not always share news articles, they do report on the same events, but often with competing and counter-narratives. Overall, we find that the news is homogeneous within communities and diverse in between, creating different spirals of sameness., Comment: Published at ICWSM 2019

With the spread of false and misleading information in current news, many algorithmic tools have been introduced with the aim of assessing bias and reliability in written content. However, there has been little work exploring how effective these tools are at changing human perceptions of content. To this end, we conduct a study with 654 participants to understand if algorithmic assistance improves the accuracy of reliability and bias perceptions, and whether there is a difference in the effectiveness of the AI assistance for different types of news consumers. We find that AI assistance with feature-based explanations improves the accuracy of news perceptions. However, some consumers are helped more than others. Specifically, we find that participants who read and share news often on social media are worse at recognizing bias and reliability issues in news articles than those who do not, while frequent news readers and those familiar with politics perform much better. We discuss these differences and their implication to offer insights for future research., Comment: Published at ICWSM 2019

In today's news ecosystem, news sources emerge frequently and can vary widely in intent. This intent can range from benign to malicious, with many tactics being used to achieve their goals. One lesser studied tactic is content republishing, which can be used to make specific stories seem more important, create uncertainty around an event, or create a perception of credibility for unreliable news sources. In this paper, we take a first step in understanding this tactic by exploring verbatim content copying across 92 news producers of various characteristics. We find that content copying occurs more frequently between like-audience sources (eg. alternative news, mainstream news, etc.), but there consistently exists sparse connections between these communities. We also find that despite articles being verbatim, the headlines are often changed. Specifically, we find that mainstream sources change more structural features, while alternative sources change many more content features, often changing the emotional tone and bias of the titles. We conclude that content republishing networks can help identify and label the intent of brand-new news sources using the tight-knit community they belong to. In addition, it is possible to use the network to find important content producers in each community, producers that are used to amplify messages of other sources, and producers that distort the messages of other sources., Comment: Peer-reviewed by NECO 2018 Workshop

The complexity and diversity of today's media landscape provides many challenges for researchers studying news producers. These producers use many different strategies to get their message believed by readers through the writing styles they employ, by repetition across different media sources with or without attribution, as well as other mechanisms that are yet to be studied deeply. To better facilitate systematic studies in this area, we present a large political news data set, containing over 136K news articles, from 92 news sources, collected over 7 months of 2017. These news sources are carefully chosen to include well-established and mainstream sources, maliciously fake sources, satire sources, and hyper-partisan political blogs. In addition to each article we compute 130 content-based and social media engagement features drawn from a wide range of literature on political bias, persuasion, and misinformation. With the release of the data set, we also provide the source code for feature computation. In this paper, we discuss the first release of the data set and demonstrate 4 use cases of the data and features: news characterization, engagement characterization, news attribution and content copying, and discovering news narratives., Comment: Published at ICWSM 2018. Dataset: https://github.com/BenjaminDHorne/NELA2017-Dataset-v1 Feature Code: https://github.com/BenjaminDHorne/Language-Features-for-News