03 · What You Need to Know
Start by Asking What the Identifier Actually Identifies
An identifier is useful because it distinguishes an entity from other entities in a particular system. The difficulty is that scholarly communication contains many kinds of entities: articles, books, book chapters, journals, researchers, institutions, datasets, grants, and database records, among others.
Consequently, the presence of several identifiers is not redundant by itself. Each may describe a different relationship.
| Identifier |
What it commonly identifies |
Where it is particularly useful |
| DOI |
A registered digital object or scholarly object, such as an article, chapter, book, dataset, preprint, or other registered content |
Persistent linking, metadata retrieval, source identification, reference management, citation |
| PMID |
A record in PubMed |
Finding and distinguishing literature indexed in PubMed |
| ISBN |
A particular title, edition, and product format of a monographic publication |
Books, editions, formats, catalogs, ordering, and bibliographic identification |
| ISSN |
A continuing resource or serial publication |
Identifying journals and other serials |
| ORCID iD |
A researcher or contributor |
Distinguishing people and connecting them with research activities and outputs |
The table also reveals why asking whether DOI is “better” than PMID or ISBN is usually the wrong comparison. They do not necessarily identify the same type of entity.
DOI: a persistent identifier for a registered object
DOI stands for Digital Object Identifier. In scholarly communication, DOIs are widely used to identify research objects and provide persistent links to them.
Crossref is one of the major DOI registration agencies serving scholarly communication. Crossref members register metadata about content and associate each registered metadata record with a unique DOI. Crossref explains that the DOI remains with the work even when its online location changes, provided that the responsible organization keeps the DOI's destination information updated.
This is why a DOI is generally more durable than copying the publisher's ordinary webpage URL. If the content moves, the DOI can continue resolving to its updated location.
A DOI commonly looks like this:
10.1234/example.5678
When displayed as a link, Crossref recommends the form:
https://doi.org/10.1234/example.5678
The characters before the slash form the DOI prefix and the characters after it form the suffix. You generally do not need to interpret either component when using the DOI as a researcher.
A DOI does not mean that a source is good, peer reviewed, or trustworthy
This is one of the most important distinctions to preserve.
A DOI identifies an object within the DOI infrastructure. It is not a quality badge. Crossref explicitly states that a DOI does not signify the value or accuracy of the content it locates.
Having a DOI therefore does not establish that an article was peer reviewed, that its methods are sound, that the journal is reputable, or that its conclusions are correct. Those are separate questions requiring evaluation of the publication and research itself.
Watch Out
Do not use the presence or absence of a DOI as a shortcut for judging research quality or legitimacy. An identifier tells you what something is within an identification system; it does not tell you whether the research deserves your confidence.
Not every legitimate scholarly source has a DOI
DOIs are common, but they are not universal.
Older journal literature may lack one. Some journals, books, reports, theses, conference materials, and other legitimate research outputs may not participate in a DOI registration system. Other works may use different persistent identifiers.
Do not invent a DOI because a citation style commonly includes one. If a work does not have a DOI, cite it according to the applicable style using the bibliographic information that actually exists.
A DOI can identify more than a journal article
Researchers sometimes use “DOI” almost as though it meant “journal article number.” The DOI system is much broader.
Crossref registers metadata for many research-object types, including journal articles, books, book chapters, conference papers, datasets, preprints and other posted content, reports, grants, peer reviews, and additional scholarly objects.
DOIs can also exist at different levels. Crossref documentation notes that a journal may have a title-level DOI while individual articles have their own DOIs. A book can similarly have a book-level DOI while individual chapters also have separate chapter-level DOIs.
This creates an important practical rule: use the identifier for the object you are actually citing.
If you cite a specific book chapter that has its own DOI, the chapter DOI is more directly associated with the cited object than a DOI identifying the book as a whole.
PMID: the identifier for a PubMed record
PMID stands for PubMed Identifier. It is the unique identifier associated with a record in PubMed, the biomedical and life-sciences literature database maintained by the U.S. National Library of Medicine.
The National Library of Medicine states that PMIDs do not change over time during processing and are never reused. A PMID can be entered directly into PubMed to retrieve the corresponding record.
A PMID is therefore especially useful when you work within biomedical and health-related literature and want to retrieve or exchange an exact PubMed record.
For example:
PMID: 12345678
The number identifies the PubMed record. It is not itself a DOI.
One article can have both a DOI and a PMID
This is not duplication.
Suppose a journal article is registered with a DOI and is also indexed in PubMed. Its DOI identifies the scholarly work through the DOI infrastructure, while its PMID identifies the record representing that work in PubMed.
DOI
Identifies a registered scholarly object and supports persistent resolution to its location.
PMID
Identifies the corresponding bibliographic record within PubMed.
This distinction becomes useful when verifying imported citation metadata. A DOI can help retrieve information from DOI infrastructure, while a PMID can take you directly to the relevant PubMed record. They provide different routes for checking the source.
Do not confuse PMID with PMCID
PMID and PMCID look similar but serve different purposes.
A PMID identifies a PubMed record. A PMCID is an identifier associated with an article in PubMed Central, the National Library of Medicine's full-text archive.
An article may have a PMID without having a PMCID. When an article is represented in both systems, it may have both identifiers.
This distinction can matter particularly in funder and public-access workflows, where a PMCID may be requested specifically. Supplying a PMID when a system asks for a PMCID does not satisfy the same identification requirement.
ISBN: an identifier for books, editions, and formats
ISBN stands for International Standard Book Number. The International ISBN Agency describes it as a product identifier used for monographic publications.
Since 1 January 2007, ISBNs have consisted of 13 digits. More importantly for researchers, an ISBN identifies a particular title, edition, and format. Different product forms of the same book, such as paperback, EPUB, or PDF editions, should generally have separate ISBNs.
That means an ISBN is more specific than simply identifying the intellectual idea of “this book.” A revised edition can have a different ISBN, and a different format may also carry a different ISBN.
This can be useful when you need to determine exactly which edition or manifestation of a book you consulted.
An ISBN is not the book equivalent of a DOI in every respect
Both can identify books, but their systems and purposes differ.
The ISBN system functions extensively as a product-identification system for publishers, booksellers, libraries, retailers, and other parts of the book supply chain. A DOI is designed as a persistent identifier and resolution mechanism for registered objects.
A book may have an ISBN, a DOI, or both. Crossref can register DOIs at the book and chapter levels, and its current guidance recommends including DOIs in citations when they are available for registered content.
Do not replace one identifier with the other merely because both happen to be present. Store them in the appropriate metadata fields and let the applicable citation style determine what appears in the final reference.
ISSN identifies the serial, not an individual journal article
ISSN stands for International Standard Serial Number. It is used to identify continuing resources such as journals and other serial publications.
This is fundamentally different from an article DOI.
If an article appears in a journal, the ISSN identifies the journal or serial manifestation, while the article's DOI, where one exists, identifies the article or other registered object at the relevant level.
Researchers do not normally substitute an ISSN for a missing article DOI. They identify different things.
ORCID iD identifies a person, not a paper
ORCID addresses another identification problem: researchers can share names, publish under name variants, change names, move institutions, and work across different scholarly systems.
An ORCID iD is a persistent identifier for an individual researcher or contributor. ORCID describes it as a name-independent person identifier designed to distinguish people and connect them reliably with their contributions and affiliations.
An ORCID iD is represented as a 16-character identifier and should be expressed as a URI, for example:
https://orcid.org/0000-0001-2345-6789
The number does not encode the researcher's country, institution, discipline, or other career information. That is deliberate because those characteristics can change while the identifier is intended to persist.
DOI
Answers: “Which research object is this?”
ORCID iD
Answers: “Which researcher or contributor is this?”
They can also be connected. Crossref metadata can contain contributor ORCID iDs, allowing scholarly metadata to connect an identified work with an identified contributor.
Scopus Author ID and Web of Science ResearcherID identify researchers within other systems
You may also encounter identifiers such as Scopus Author ID and Web of Science ResearcherID. These are person identifiers associated with particular scholarly information systems.
They should not be confused with identifiers for research outputs. ORCID can also connect with other person identifiers. ORCID's documentation lists Scopus Author ID and ResearcherID among identifiers that trusted organizations may add or connect to an ORCID record.
For routine reference management, you are usually more concerned with identifiers attached to the source itself. Person identifiers become particularly useful when distinguishing authors, maintaining researcher profiles, submitting manuscripts, applying for grants, or connecting research activities across systems.
Identifiers can make importing references more reliable
When a reference manager supports identifier-based lookup, an identifier can provide a more precise starting point than manually typing a title.
A DOI can retrieve the metadata associated with a registered scholarly object. A PMID can identify a specific PubMed record. An ISBN can help identify a book product. Reference-management software may support some or all of these identifiers depending on the platform.
This does not eliminate the need to inspect the imported metadata. Identifier lookup can retrieve the correct record while still exposing incomplete or inconsistent metadata from the underlying source.
The useful sequence is therefore:
Identify Use the appropriate identifier to locate the intended object or record.
Import Allow the reference manager to retrieve available bibliographic metadata.
Verify Confirm that the resulting record accurately represents the work.
Cite Let the citation style determine how the verified bibliographic information and identifier should appear.
Identifiers are especially useful for duplicate detection
Two imported records may have slightly different titles or author formatting while sharing the same DOI. That common identifier is strong evidence that the records may represent the same object.
Reference managers can therefore use identifiers as part of duplicate detection. Zotero, for example, considers DOI and ISBN among the fields used by its duplicate-detection system.
Identifiers do not eliminate the need for judgment. A book may legitimately have different ISBNs for different formats, while different versions of a scholarly work may need separate records. Use identifiers as evidence when resolving duplicate references, not as an instruction to merge records blindly.
Do not put every identifier you find into the DOI field
A reference manager's DOI field is for a DOI. A PMID, ISBN, ISSN, accession number, database identifier, or ordinary URL does not become a DOI because it uniquely identifies something somewhere.
Misplacing identifiers can interfere with citation formatting, metadata lookup, linking, and deduplication.
If your reference manager has a dedicated field for the identifier, use it. If it does not, follow the software's documentation for storing additional identifiers rather than repurposing an unrelated field.
The citation style determines which identifiers appear in your bibliography
Keeping an identifier in your reference library and displaying it in a bibliography are separate questions.
A citation style may require or prefer a DOI for a journal article while omitting identifiers that are irrelevant to the reference format. Some disciplinary or database-specific styles may use PMIDs or PMCIDs in particular circumstances. Book references may or may not display an ISBN, depending on the purpose and style.
Do not manually delete useful metadata from your library merely because a particular citation style does not display it. The record can contain information useful for identification and retrieval even when the final bibliography does not show every field.
Conversely, do not force an identifier into the formatted citation merely because your reference manager stores it. Before submission, check what the required bibliography should actually contain.
