Verification Log for Anchoring Elements in AI and Human News Texts

Replication data supports an evidence-logged verifiability audit comparing anchoring elements in AI-generated versus human-written news texts. The dataset includes a derived verification log in CSV format with evidence fields, scripts for reproducing reported statistics and figures, and a SHA256 manifest for integrity checks. It was authored by Zhilong Zhao and last updated in January 2026.

Use Cases

• Reproduce reported statistics and figures from the audit using the provided scripts and verification log.
• Analyze evidence fields within the verification log to audit anchoring elements in text.
• Verify data integrity for the derived verification log using the included SHA256 manifest.

Strengths

• Includes a SHA256 manifest for integrity verification of the derived data.
• Contains scripts to fully regenerate the reported statistics and figures from the research article.
• Dataset was updated in January 2026, indicating recent maintenance.

Limitations

• The dataset contains derived data for replication, not the original source texts used in the audit.
• Specific data volume (rows, columns) and file formats are not documented.
• Geographic or temporal scope of the underlying news texts is not specified.

Provenance

Source

zhilong zhao Dataverse

Collection Method

Derived data and code created to replicate results from a research article on anchoring elements in news texts.

Freshness

Last updated 2026-01-20.