Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
51
datasets available to search
ShareScore release 0.7.1
Dataset results
51 results for “Misinformation”
Data from: Processing political misinformation: comprehending the Trump phenomenon
Open the record for dataset details and reuse information.
PolitiKweli Misinformation Classification Dataset
<p>PolitiKweli is a Swahili-English Code-switched Twitter Political Misinformation Classification Dataset</p>
EEG data in the tasks of misinformed sentence reading
<p><span>The present study assumes that the two indispensable meaning factors, the Familiarity and the Outcome of a substance in misinformation, might affect the neural responses of sentence processing and trustworthiness validation. The behavioral and EPRs data support this assumption. It demonstrated that participants tend to rate the perceived trustworthiness around a happy medium, yet the score of Unfamiliar stimuli was slightly higher than the Familiar. The onset of an Unfamiliar subject elicited more negative N400 at a left occipital site, and a word rendering a Good Outcome elevated N400 at a right occipital site. At the same time, a Familiar subject elicited higher LPC on the right peripheral lines from the temporal-parietal lobe to the occipital in contrast to an Unfamiliar Term. Furthermore, the amplitudes of N400 observed on some sites strongly correlated with the trustworthiness scores, wherein the more negative N400 was evoked, the less trustworthy a statement rated.</span></p>
Top FIB Misinformation Superspreaders
The Top FIBers dashboard (https://osome.iu.edu/tools/topfibers/) tracks and reports on the top superspreaders of low-credibility information on Twitter and Facebook each month. This repository serves as a record of the top 50 we find each time we update the dashboard. Superspreaders are identified using the FIB index. Please see the paper (https://arxiv.org/abs/2207.09524) for more details. You can learn more about the project here: https://osome.iu.edu/tools/topfibers/about
Leveraging Community Health Workers to Combat COVID-19 and Mental Health Misinformation in Haiti, Malawi, and Rwanda
ClinicalTrials.gov study NCT05997511. IPD Sharing: NO. Countries: 3. Publications: 0.
Protecting Against HIV Vaccine Misinformation With Adolescent Girls and Young Women in South Africa
ClinicalTrials.gov study NCT06700447. IPD Sharing: UNDECIDED. Countries: 1. Publications: 0.
Can Doctors Reduce COVID-19 Misinformation and Increase Vaccine Uptake in Ghana? A Cluster-randomised Controlled Trial
ClinicalTrials.gov study NCT06191211. IPD Sharing: YES. Countries: 1. Publications: 0.
EEG data in the tasks of misinformed sentence reading
Open the record for dataset details and reuse information.
MisInformation
<p>Misinformation</p>
Dataset for the paper: "Monant Medical Misinformation Dataset: Mapping Articles to Fact-Checked Claims"
<p><strong>Overview</strong></p> <p>This dataset of medical misinformation was collected and is published by <a href="https://kinit.sk/">Kempelen Institute of Intelligent Technologies (KInIT)</a>. It consists of approx. 317k news articles and blog posts on medical topics published between January 1, 1998 and February 1, 2022 from a total of 207 reliable and unreliable sources. The dataset contains full-texts of the articles, their original source URL and other extracted metadata. If a source has a credibility score available (e.g., from Media Bias/Fact Check), it is also included in the form of annotation. Besides the articles, the dataset contains around 3.5k fact-checks and extracted verified medical claims with their unified veracity ratings published by fact-checking organisations such as Snopes or FullFact. Lastly and most importantly, the dataset contains 573 manually and more than 51k automatically labelled mappings between previously verified claims and the articles; mappings consist of two values: <em>claim presence </em>(i.e., whether a claim is contained in the given article) and <em>article stance </em>(i.e., whether the given article supports or rejects the claim or provides both sides of the argument).</p> <p>The dataset is primarily intended to be used as a training and evaluation set for machine learning methods for claim presence detection and article stance classification, but it enables a range of other misinformation related tasks, such as misinformation characterisation or analyses of misinformation spreading.</p> <p>Its novelty and our main contributions lie in (1) focus on medical news article and blog posts as opposed to social media posts or political discussions; (2) providing multiple modalities (beside full-texts of the articles, there are also images and videos), thus enabling research of multimodal approaches; (3) mapping of the articles to the fact-checked claims (with manual as well as predicted labels); (4) providing source credibility labels for 95% of all articles and other potential sources of weak labels that can be mined from the articles' content and metadata.</p> <p>The dataset is associated with the research paper "Monant Medical Misinformation Dataset: Mapping Articles to Fact-Checked Claims" accepted and presented at ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR '22). </p> <p>The accompanying <a href="https://github.com/kinit-sk/medical-misinformation-dataset">Github repository</a> provides a small static sample of the dataset and the dataset's descriptive analysis in a form of Jupyter notebooks.</p> <p>In order to obtain an access to the full dataset (in the CSV format), please, request the access by following the instructions provided below.</p> <p> </p> <p><strong>Note: </strong>Please, check also our <a href="https://doi.org/10.5281/zenodo.7737982">MultiClaim Dataset</a> that provides a more recent, a larger, and a highly multilingual dataset of fact-checked claims, social media posts and relations between them.</p> <p> </p> <p><strong>References</strong></p> <p>If you use this dataset in any publication, project, tool or in any other form, please, cite the following papers:</p> <pre><code>@inproceedings{SrbaMonantPlatform, author = {Srba, Ivan and Moro, Robert and Simko, Jakub and Sevcech, Jakub and Chuda, Daniela and Navrat, Pavol and Bielikova, Maria}, booktitle = {Proceedings of Workshop on Reducing Online Misinformation Exposure (ROME 2019)}, pages = {1--7}, title = {Monant: Universal and Extensible Platform for Monitoring, Detection and Mitigation of Antisocial Behavior}, year = {2019} }</code></pre> <pre><code>@inproceedings{SrbaMonantMedicalDataset, author = {Srba, Ivan and Pecher, Branislav and Tomlein Matus and Moro, Robert and Stefancova, Elena and Simko, Jakub and Bielikova, Maria}, booktitle = {Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR '22)}, numpages = {11}, title = {Monant Medical Misinformation Dataset: Mapping Articles to Fact-Checked Claims}, year = {2022}, doi = {10.1145/3477495.3531726}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, url = {https://doi.org/10.1145/3477495.3531726}, } </code></pre> <p><br><strong>Dataset creation process</strong></p> <p>In order to create this dataset (and to continuously obtain new data), we used our research platform <a href="https://rome2019.github.io/papers/Srba_etal_ROME2019.pdf">Monant</a>. The Monant platform provides so called data providers to extract news articles/blogs from news/blog sites as well as fact-checking articles from fact-checking sites. General parsers (from RSS feeds, Wordpress sites, Google Fact Check Tool, etc.) as well as custom crawler and parsers were implemented (e.g., for fact checking site Snopes.com). All data is stored in the unified format in a central data storage.<br><strong> <br> <br>Ethical considerations</strong></p> <p>The dataset was collected and is published for research purposes only. We collected only publicly available content of news/blog articles. The dataset contains identities of authors of the articles if they were stated in the original source; we left this information, since the presence of an author's name can be a strong credibility indicator. However, we anonymised the identities of the authors of discussion posts included in the dataset. </p> <p>The main identified ethical issue related to the presented dataset lies in the risk of mislabelling of an article as supporting a false fact-checked claim and, to a lesser extent, in mislabelling an article as not containing a false claim or not supporting it when it actually does. To minimise these risks, we developed a labelling methodology and require an agreement of at least two independent annotators to assign a claim presence or article stance label to an article. It is also worth noting that we do not label an article as a whole as false or true. Nevertheless, we provide partial article-claim pair veracities based on the combination of claim presence and article stance labels.</p> <p>As to the veracity labels of the fact-checked claims and the credibility (reliability) labels of the articles' sources, we take these from the fact-checking sites and external listings such as Media Bias/Fact Check as they are and refer to their methodologies for more details on how they were established.</p> <p>Lastly, the dataset also contains automatically predicted labels of claim presence and article stance using our baselines described in the next section. These methods have their limitations and work with certain accuracy as reported in this paper. This should be taken into account when interpreting them. <br><strong> <br> <br>Reporting mistakes in the dataset</strong><br>The mean to report considerable mistakes in raw collected data or in manual annotations is by creating a new issue in the accompanying <a href="https://github.com/kinit-sk/medical-misinformation-dataset">Github repository</a>. Alternately, general enquiries or requests can be sent at info [at] kinit.sk.</p> <p><br><strong>Dataset structure</strong></p> <p><em><strong>Raw data</strong></em></p> <p>At first, the dataset contains so called raw data (i.e., data extracted by the Web monitoring module of Monant platform and stored in exactly the same form as they appear at the original websites). Raw data consist of articles from news sites and blogs (e.g. naturalnews.com), discussions attached to such articles, fact-checking articles from fact-checking portals (e.g. snopes.com). In addition, the dataset contains feedback (number of likes, shares, comments) provided by user on social network Facebook which is regularly extracted for all news/blogs articles.</p> <p>Raw data are contained in these CSV files:</p> <ul> <li>sources.csv</li> <li>articles.csv</li> <li>article_media.csv</li> <li>article_authors.csv</li> <li>discussion_posts.csv</li> <li>discussion_post_authors.csv</li> <li>fact_checking_articles.csv</li> <li>fact_checking_article_media.csv</li> <li>claims.csv</li> <li>feedback_facebook.csv</li> </ul> <p><em>Note: Personal information about discussion posts' authors (name, website, gravatar) are anonymised.</em></p> <p><br><em><strong>Annotations</strong></em></p> <p>Secondly, the dataset contains so called annotations. Entity annotations describe the individual raw data entities (e.g., article, source). Relation annotations describe relation between two of such entities.</p> <p>Each annotation is described by the following attributes:</p> <ol> <li>category of annotation (`annotation_category`). Possible values: label (annotation corresponds to ground truth, determined by human experts) and prediction (annotation was created by means of AI method).</li> <li>type of annotation (`annotation_type_id`). Example values: Source reliability (binary), Claim presence. The list of possible values can be obtained from enumeration in annotation_types.csv.</li> <li>method which created annotation (`method_id`). Example values: Expert-based source reliability evaluation, Fact-checking article to claim transformation method. The list of possible values can be obtained from enumeration methods.csv.</li> <li>its value (`value`). The value is stored in JSON format and its structure differs according to particular annotation type.</li> </ol> <p><br>At the same time, annotations are associated with a particular object identified by:</p> <ol> <li>entity type (parameter `entity_type` in case of entity annotations, or `source_entity_type` and `target_entity_type` in case of relation annotations). Possible values: sources, articles, fact-checking-articles.</li> <li>entity id (parameter `entity_id` in case of entity annotations, or `source_entity_id` and `target_entity_id` in case of relation annotations).</li> </ol> <p><br>The dataset provides specifically these entity annotations:</p> <ul> <li>Source reliability (binary). Determines validity of source (website) at a binary scale with two options: reliable source and unreliable source.</li> <li>Article veracity. Aggregated information about veracity from article-claim pairs.</li> </ul> <p>The dataset provides specifically these relation annotations:</p> <ul> <li>Fact-checking article to claim mapping. Determines mapping between fact-checking article and claim.</li> <li>Claim presence. Determines presence of claim in article.</li> <li>Claim stance. Determines stance of an article to a claim. </li> </ul> <p><br>Annotations are contained in these CSV files:</p> <ul> <li>entity_annotations.csv</li> <li>relation_annotations.csv</li> </ul> <p><em>Note: Identification of human annotators authors (email provided in the annotation app) is anonymised.</em></p> <p> </p> <p><em><strong>Enumerations</strong></em></p> <p>Finally, the dataset provides additional CSV files with enumerations:</p> <ul> <li>media_types.csv</li> <li>source_types.csv</li> <li>annotation_types.csv</li> <li>methods.csv</li> </ul>
Resistant and Misinformed: Understanding the Role of Elite Perceptions and Resistance Strategies against COVID-19 Reporting in Legacy Media in Adopting COVID-19 Misperceptions
<p>The following dataset underlies the study " Resistant and Misinformed: Understanding the Role of Elite Perceptions and Resistance Strategies against COVID-19 Reporting in Legacy Media in Adopting COVID-19 Misperceptions". The data is published to allow for replicating and reviewing the analysis of our study.</p>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.