eduKateSG · Why Science?
A paper–data connection becomes reusable when both ends and the link are described
Use the Scholix framework to read literature–data links through persistent identifiers, object types, relationship direction, publishers, source systems and link provenance.
Scientific papers often depend on datasets stored elsewhere. Repositories and publishers may know about those relationships, yet their systems use different metadata and workflows. Scholix, short for Scholarly Link Exchange, provides a high-level interoperability framework for exchanging information about links between scholarly literature and data.
The Research Data Alliance and World Data System working group describes Scholix as a lightweight set of guidelines with a conceptual information model and link metadata schema. It is designed for exchange among hubs and aggregators such as DataCite, Crossref, OpenAIRE and EMBL-EBI rather than replacing each community’s deposit process.
Did you know? A useful Scholix link is more than two URLs. It describes source and target objects, their persistent identifiers and types, the relationship, and provenance about who asserted or supplied the link. That structure lets another system interpret the connection without guessing from page text.
This guide uses Scholix to teach connection literacy. A link can help find data associated with a publication; it does not prove that the files are complete, that they reproduce every analysis, or that the paper’s conclusions are correct.
Chapter index
Chapters 1–9: link objects, identifiers and direction
- 1. A scholarly link has two endpoints
- 2. Literature and data are different objects
- 3. Persistent identifiers reduce ambiguity
- 4. Object types shape interpretation
- 5. Relationship direction matters
- 6. The information model is conceptual
- 7. The schema carries exchange metadata
- 8. Hubs exchange at wholesale scale
- 9. Publishers and repositories see different sides
Chapters 10–17: exchange hubs, provenance and evidence limits
- 10. Provenance names the assertion route
- 11. Link dates are not research dates
- 12. One paper can connect to many datasets
- 13. One dataset can support many papers
- 14. Missing links are not proof of absence
- 15. Duplicate links need reconciliation
- 16. Links support reproducibility checks
- 17. Exchange standards support careers
A literature–data link is evidence about a relationship, not a scientific conclusion by itself. The framework represents a relationship between a source object and a target object. Each endpoint should remain individually identifiable. This gives science learners a precise way to connect papers with supporting data.
Try this
Evidence question: Which object is the source and which is the target? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Write both identifiers on separate lines before naming the relationship. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Keep the identifier, retrieval date and source page in the note so the route can be repeated.
Evidence checkpoint
Record the evidence object for a scholarly link has two endpoints, its identifier or version context, and one limit that must remain visible.
Shared link metadata lets separate scholarly hubs exchange connections consistently. A publication tells a narrative and presents claims, while a dataset packages observations or derived material. Their records may have different custodians and licences. It also prevents a visible link from being mistaken for complete reproducibility.
Try this
Evidence question: What evidence role does each endpoint play? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Describe the paper and dataset without collapsing them into one citation. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. If a field is missing, write “not shown” and follow the official documentation rather than inventing a value.
Evidence checkpoint
Record the evidence object for literature and data are different objects, its identifier or version context, and one limit that must remain visible.
The source object, target object and relationship direction must travel together. DOIs and other persistent identifiers make it easier for exchange systems to refer to the same object. A title-only link is more vulnerable to spelling and version changes. The useful habit is to inspect both endpoints and the party that supplied the link.
Try this
Evidence question: Does each endpoint have a resolvable identifier? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Resolve both IDs and compare the landing-page metadata. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Before comparing two records, check that their object types, methods, versions and units really align.
Evidence checkpoint
Record the evidence object for persistent identifiers reduce ambiguity, its identifier or version context, and one limit that must remain visible.
Persistent identifiers make a cross-system link more durable and less ambiguous. Link metadata identifies whether an endpoint is literature, data or another scholarly object category. Type helps receiving systems decide how to display and validate it. A connection can then support discovery without hiding its provenance or limits.
Try this
Evidence question: Do the declared types match the landing pages? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Record any mismatch as a data-quality issue. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Finish by naming one supported statement and one stronger statement that still needs another evidence layer.
Evidence checkpoint
Record the evidence object for object types shape interpretation, its identifier or version context, and one limit that must remain visible.
A literature–data link is evidence about a relationship, not a scientific conclusion by itself. The same pair of objects can be described from either direction, and relationship terms should be read as complete sentences. Reversing direction can change the meaning. This gives science learners a precise way to connect papers with supporting data.
Try this
Evidence question: Which object references, is supplemented by or is documented by the other? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Read the relation aloud with both titles inserted. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Keep the identifier, retrieval date and source page in the note so the route can be repeated.
Evidence checkpoint
Record the evidence object for relationship direction matters, its identifier or version context, and one limit that must remain visible.
Shared link metadata lets separate scholarly hubs exchange connections consistently. Scholix defines the pieces needed to describe a link independently of one repository’s local database design. Implementations map their fields into this shared model. It also prevents a visible link from being mistaken for complete reproducibility.
Try this
Evidence question: Which local field was mapped to the common relation? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Keep the original term and mapped term together. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. If a field is missing, write “not shown” and follow the official documentation rather than inventing a value.
Evidence checkpoint
Record the evidence object for the information model is conceptual, its identifier or version context, and one limit that must remain visible.
The source object, target object and relationship direction must travel together. A link metadata schema gives systems a consistent representation of endpoints, relationships and provenance. Schema validity supports interoperability, not scientific truth. The useful habit is to inspect both endpoints and the party that supplied the link.
Try this
Evidence question: Is the record structurally valid and scientifically supported? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Separate a syntax check from an evidence check. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Before comparing two records, check that their object types, methods, versions and units really align.
Evidence checkpoint
Record the evidence object for the schema carries exchange metadata, its identifier or version context, and one limit that must remain visible.
Persistent identifiers make a cross-system link more durable and less ambiguous. The RDA recommendation describes Scholix as a wholesaler-to-wholesaler framework for major link aggregators. Community repositories and publishers can continue using familiar workflows with their hubs. A connection can then support discovery without hiding its provenance or limits.
Try this
Evidence question: Which hub collected the original link? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Trace from the aggregator back to the contributing community source. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Finish by naming one supported statement and one stronger statement that still needs another evidence layer.
Evidence checkpoint
Record the evidence object for hubs exchange at wholesale scale, its identifier or version context, and one limit that must remain visible.
CHAPTER 9 OF 17 · Trace the provider
9. Publishers and repositories see different sides
A literature–data link is evidence about a relationship, not a scientific conclusion by itself. A publisher may deposit a paper-to-data reference, while a repository may record the dataset’s related article. Combining both can improve discovery but can also reveal inconsistent direction or identifiers. This gives science learners a precise way to connect papers with supporting data.
Try this
Evidence question: Do both sides describe the same relationship? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Compare publisher and repository records before reconciling them. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Keep the identifier, retrieval date and source page in the note so the route can be repeated.
Evidence checkpoint
Record the evidence object for publishers and repositories see different sides, its identifier or version context, and one limit that must remain visible.
Shared link metadata lets separate scholarly hubs exchange connections consistently. Link provenance tells a consumer who supplied or aggregated the connection. That information helps diagnose errors and choose a correction route. It also prevents a visible link from being mistaken for complete reproducibility.
Try this
Evidence question: Who asserted, collected or exposed this link? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Add a provenance column to the endpoint table. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. If a field is missing, write “not shown” and follow the official documentation rather than inventing a value.
Evidence checkpoint
Record the evidence object for provenance names the assertion route, its identifier or version context, and one limit that must remain visible.
The source object, target object and relationship direction must travel together. Creation or update timestamps for a link record describe metadata events. They may differ from data collection, publication or analysis dates. The useful habit is to inspect both endpoints and the party that supplied the link.
Try this
Evidence question: Which event does each date represent? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Build separate link-record and research timelines. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Before comparing two records, check that their object types, methods, versions and units really align.
Evidence checkpoint
Record the evidence object for link dates are not research dates, its identifier or version context, and one limit that must remain visible.
Persistent identifiers make a cross-system link more durable and less ambiguous. A study may use raw data, processed data, calibration files and supplementary tables. A single generic link may not show which object supports a particular figure. A connection can then support discovery without hiding its provenance or limits.
Try this
Evidence question: Which exact dataset or file supports the claim? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Map one result to the narrowest available research object. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Finish by naming one supported statement and one stronger statement that still needs another evidence layer.
Evidence checkpoint
Record the evidence object for one paper can connect to many datasets, its identifier or version context, and one limit that must remain visible.
A literature–data link is evidence about a relationship, not a scientific conclusion by itself. A reusable dataset may be analysed by later studies. Those links do not make every paper an original data producer. This gives science learners a precise way to connect papers with supporting data.
Try this
Evidence question: Is the relationship creation, reuse, citation or commentary? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Label producer and secondary-use papers separately. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Keep the identifier, retrieval date and source page in the note so the route can be repeated.
Evidence checkpoint
Record the evidence object for one dataset can support many papers, its identifier or version context, and one limit that must remain visible.
Shared link metadata lets separate scholarly hubs exchange connections consistently. Exchange coverage depends on what providers deposit and what hubs collect. No retrieved link does not prove that no relationship exists. It also prevents a visible link from being mistaken for complete reproducibility.
Try this
Evidence question: Which other source could hold the connection? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Check the article, repository record and identifier metadata before concluding absence. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. If a field is missing, write “not shown” and follow the official documentation rather than inventing a value.
Evidence checkpoint
Record the evidence object for missing links are not proof of absence, its identifier or version context, and one limit that must remain visible.
The source object, target object and relationship direction must travel together. Several providers can report the same endpoint pair. Aggregators may normalise or combine those reports while keeping provenance. The useful habit is to inspect both endpoints and the party that supplied the link.
Try this
Evidence question: Are these repeated assertions or distinct relationships? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Deduplicate by endpoint IDs and relation while preserving every source. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Before comparing two records, check that their object types, methods, versions and units really align.
Evidence checkpoint
Record the evidence object for duplicate links need reconciliation, its identifier or version context, and one limit that must remain visible.
Persistent identifiers make a cross-system link more durable and less ambiguous. A visible paper–data route helps readers locate candidate inputs. Reproducibility still requires versions, files, software, parameters, methods and access conditions. A connection can then support discovery without hiding its provenance or limits.
Try this
Evidence question: Which required evidence lies outside the link record? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Create a reproducibility checklist that begins after resolving both endpoints. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Finish by naming one supported statement and one stronger statement that still needs another evidence layer.
Evidence checkpoint
Record the evidence object for links support reproducibility checks, its identifier or version context, and one limit that must remain visible.
A literature–data link is evidence about a relationship, not a scientific conclusion by itself. Metadata specialists, repository developers, publishers, data stewards and research software engineers implement and audit scholarly links. This gives science learners a precise way to connect papers with supporting data.
Try this
Evidence question: Which hand-off fails when relationship direction is lost? Answer in three parts: identify the visible field or file, explain how it bears on the question, and state the boundary of that evidence. Preserve the database’s own object names and accession rather than flattening everything into the word ‘data’.
Try this: Design a Scholix-style link receipt for a school investigation. Turn the result into a compact claim-evidence-reasoning note, then ask a partner to retrace it. Keep the identifier, retrieval date and source page in the note so the route can be repeated.
Evidence checkpoint
Record the evidence object for exchange standards support careers, its identifier or version context, and one limit that must remain visible.
Contents · Previous chapter · Continue at the Science Learning Hub
Follow both endpoints before trusting the bridge
Scholix helps large scholarly systems exchange literature–data connections in a shared form. Its strength is interoperability: a link can travel while keeping objects, direction and provenance visible.
Save both persistent identifiers, object types, relation direction, provider, aggregator and retrieval date. Then inspect the paper, dataset files, methods and version context before making a reproducibility claim.
| Reading object | What it can support | Question to keep visible |
|---|---|---|
| Source object | First endpoint in the exchanged link | What object and identifier does it name? |
| Target object | Second endpoint | Does its type and version match? |
| Relationship | Directed connection | How should it be read as a sentence? |
| Provenance | Supplier or aggregator of the link | Where should an error be corrected? |
Sources and next reading
Substantive resource claims in this article were checked against the primary pages below. A structured record is a route into evidence; it does not by itself prove a scientific conclusion. Recheck the live record, evidence labels and release context before using it in assessed work or research.
For a classroom enquiry, save a short research receipt containing the question, search terms, selected object, exact identifier, version, access date and the evidence boundary you observed. Add the file or page that supports each statement. This small routine turns browsing into accountable science learning and gives a teacher, classmate or future analyst a route they can independently follow.
Before sharing the conclusion, ask another reader to retrace the route without verbal hints. If they reach a different version, file or interpretation, improve the receipt and narrow the claim. Reproducibility begins with this practical check. It teaches careful reading, supports fair collaboration and keeps uncertainty visible while evidence moves from an online service into notes, analysis and a final explanation.
A useful final review separates five questions. What object was found? Which organisation or repository maintains it? What exact evidence was inspected? Which statement does that evidence support? What remains uncertain? Answer each question in one sentence, then connect every answer to a visible field, file or source page. This review catches common mistakes: citing a collection when a file supplied the value, treating the newest release as the analysed release, confusing permission with scientific suitability, or turning a discovery count into a conclusion. It also gives students a calm way to improve work. A missing detail becomes a next research step instead of an invitation to guess. When the route is complete, the final explanation can stay simple because the identifiers, dates, versions, limitations and supporting evidence are already organised underneath it.
Primary sources
- RDA Scholix recommendation — accessed 25 October 2026.
- Scholix metadata schema PDF — accessed 25 October 2026.
- RDA Scholix working-group outputs — accessed 25 October 2026.
- RDA Scholix recommendation focus — accessed 25 October 2026.
- RDA open literature-data cross-linking service — accessed 25 October 2026.