Reddit Stories Annotated for Stance and Veracity

A collection of a series of Reddit stories and surrounding conversations annotated for stance and veracity. It is sourced from the ACL anthology and was last updated in October 2022. The dataset size is categorized as between 1,000 and 10,000 entries.

Use Cases

• Train a text classification model for stance detection using the annotated stance labels on Reddit conversations.
• Analyze the relationship between conversation patterns and the annotated veracity labels of the original stories.
• Build a multi-task model that jointly predicts stance and veracity from the text of Reddit stories and comments.

Strengths

• Annotations for both stance and veracity were created by experts.
• The dataset size is categorized as between 1,000 and 10,000 entries.
• The data is monolingual, focusing on a single language for consistency.

Limitations

• The specific number of rows and columns is unknown, limiting precise assessment of scale and features.
• The geographic region is indicated as US, which may limit generalizability to other linguistic or cultural contexts.
• The dataset was last updated in 2022, potentially making it temporally stale for current online discourse analysis.

Provenance

Source

Reddit, via ACL anthology.

Collection Method

Original dataset with expert-generated annotations.

Freshness

Last updated 2022-10-25.

Geography

Region indicated as US.