SHARE Score

About
Framework
/

Harmonization

/

H4

H4

Bibliographic References

Cited works with DOIs, linked methods papers

Harmonization (H)
Interoperable (I3)

Justification

References to methods papers and related works create knowledge graphs. DataCite Recommends RelatedIdentifier with rich relation types. Dublin Core includes Relation and Source as core elements. schema.org includes citation as recommended. RDA-I3-01M and RDA-I3-03M (Important) require qualified references.

Practical Guide

good-to-have

Link referenced works. Builds knowledge graph connections.

Bibliographic references create knowledge graph connections between datasets and published works. The citation impact is neutral (RR = 1.01, p = 0.666) — references don't directly predict citations. But their value is in contextual discovery: linking your dataset to methods papers and related literature helps users understand what your data is for and how to use it.

For Repositories

  • Support structured reference fields with DOI/PMID identifiers
  • Map to DataCite #12 RelatedIdentifier with rich relation types
  • Auto-extract references from uploaded papers when possible

For Depositors

  • Link to methods papers and key publications that informed your data
  • Use DOIs for references when available for machine readability
  • Include both the dataset's source papers and downstream analysis publications

Knowledge graph value is real but citation evidence is neutral. Low prevalence (0.9%) means high differentiation potential.

Standards Sources

Convergence score: 4/4 independent sources —

Strongly justified

StandardField / PropertyObligation Level
DataCite 4.6#12 RelatedIdentifier
Recommended
Dublin CoreRelation / Source
Core Element
schema.orgcitation
Recommended

FAIR Principle Alignment

Primary mapping: Interoperable (I3)

  • I3: (Meta)data include qualified references to other (meta)data

RDA FAIR Data Maturity Model Indicators:

  • RDA-I3-01M: Metadata includes references to other metadata
  • RDA-I3-03M: Metadata includes qualified references to other metadata

How This Signal Is Measured

Presence of references with identifiers (DOIs, PMIDs). Binary: at least one reference present.

Empirical Evidence (Zenodo, n=1.3M)

Per-signal statistics use Zenodo as the primary validation source because it is the largest general-purpose repository with structured DataCite metadata, natural variance across all 25 signals, and available citation/usage data. Domain-specific repositories exhibit ceiling effects or restricted variance that preclude per-signal discrimination. Cross-repository validation is reported separately.

Prevalence

0.9%

of Zenodo datasets

Citation Lift

1.0x

vs. datasets without

Data Source

Zenodo (CERN)

1,328,100 records analyzed

Interpretation: Neutral citation lift but bibliographic references create knowledge graphs enabling contextual discovery. Value is in network effects, not direct citation prediction.

Quantitative Evidence

Scoring Formula

bibliographic_references.length ≥ 1 → 4 pts

Contribution: 4 of 100 points · Harmonization bucket (0–20)

With Signal Present

11,436

datasets (0.9%)

μ = 0.246 citations/dataset

Without Signal

1,316,664

datasets (99.1%)

μ = 0.244 citations/dataset

Rate Ratio

1.01

95% CI: [0.971.05]

P-value

0.666

z = 0.43

Significance

Not significant

Method: Poisson rate ratio · Source: Zenodo (n = 1,328,100)

Note: Not statistically significant (p = 0.666). Value is in knowledge graph connectivity and contextual discovery, not direct citation prediction.

ShareScore