Determine whether journals manipulate metadata to improve indexing

Determine whether journals intentionally manipulate metadata deposited to Crossref—such as selecting English as a primary language, front‑loading English abstracts, or omitting language attributes—to improve indexing outcomes and discoverability across downstream platforms.

Background

The paper identifies widespread gaps and inconsistencies in language-related metadata within Crossref records, including the near absence of article-level language attributes (95.7% missing) and frequent discrepancies between declared and detected languages. It also notes that the Crossref REST API exposes only the first listed abstract, potentially incentivizing journals to lead with English abstracts for indexing purposes.

Because these issues complicate reliable language detection and reporting, the authors explicitly raise whether journals might be gaming metadata to improve indexing, highlighting a need to empirically establish the presence and extent of such practices.