What Is a DOI vs. a URL: Which One Goes in Your Citation?

What a DOI actually is

DOI stands for Digital Object Identifier: a string beginning with 10, then a registrant code, then a slash and the work's own suffix — 10.1037/per0000431 identifies one article permanently. The identifier is registered with an international agency, and registrants contractually keep it resolving: put https://doi.org/ in front and the system forwards to whichever page currently hosts the work, surviving publisher mergers, website redesigns and journal ownership changes that kill ordinary URLs. The persistence is contractual, not technical luck: Crossref and DataCite registrants pay annual fees and agree to maintain resolution, which is why a 2003 DOI still opens while the 2015 publisher URL behind it has died twice.

The persistence is engineered, not accidental, and it explains why citation styles trust it. Studies of link rot have measured roughly one dead link per fifty per year among cited web sources, while DOIs keep resolving at rates above 99 percent because the identifier never depends on one server staying up. Where both a DOI and a URL exist on the same article — the normal case — the DOI is the one the style guides want, and the URL is duplication rather than an upgrade.

The numbers at a glance

Which locator to print, by situation — the range, and where each row actually shows up in coursework. Skim it first; the sections below explain why the range exists and which row applies to you.

The fallback chain every style uses

When an article has no DOI — older journals, print-only titles, some book chapters — the styles agree on an order of substitutes: the stable publisher URL first, then the database name for content available only through a library platform. What never belongs in a reference is the session URL the browser showed while you were logged in: proxy links stuffed with ezproxy and accountid fragments authenticate you, not the article, and they fail for the grader the first time they click.

Database permalinks are the legitimate middle case. EBSCO, ProQuest and JSTOR each provide a 'permalink,' 'stable URL' or 'share' button that generates a durable link — use that when the database is genuinely the only access route, and name the database itself when the style calls for it, as APA 7 does for works from research databases beyond the library. The two-second habit: before pasting any URL into a reference, imagine clicking it after graduation, from off campus, without your login.

APA 7 rules: DOIs as links, retrieval dates rare

APA 7 formats the DOI as a live hyperlink and dropped the old 'DOI:' label — the reference ends with https://doi.org/10.xxxx/xxxxx and nothing else. URLs print in full as plain links, 'Retrieved from' is gone, and retrieval dates appear only when the content is designed to change over time — a wiki entry, a live dashboard, an unarchived page. A journal article's reference is finished the moment its DOI is attached; there is nothing to update later. For a worked example, the entry ends: https://doi.org/10.1037/per0000431 — no period after the link, no 'Retrieved from,' no database name, and the same identifier repeated in every in-text mention that needs one.

The presentation details graders check: the DOI link should not be styled differently from the reference text by your formatting choices, long DOIs are allowed to break across lines, and the reference list keeps its hanging indent regardless of links. If a form or citation manager asks for 'the DOI,' give the bare 10.xxxx form; if the style manual shows the link form, print the full https://doi.org/ version. Both identify the same object — contexts differ, the identifier does not. When no DOI exists, the fallback order is stable database URL, then publisher URL, then nothing — a print-era article reference ending after the page range is complete, and bolting a database link onto one is an error, not an upgrade.

MLA 9 and Chicago: the same logic, two dialects

MLA 9 wants the DOI presented through its resolver — https://doi.org/10.xxxx — and for plain URLs it does the opposite of APA: it omits the https:// prefix, printing www.example.org/page rather than the full address. MLA also avoids 'Retrieved from' phrasing entirely and asks for access dates only in special cases. The philosophy matches APA's — persistent locators first, session URLs never — with different punctuation habits on top. Chicago 18 notes print the DOI at the end of the citation, usually as the full resolver link, and the bibliography follows the same pattern — one more reason the identifier is worth capturing at reading time.

Chicago handles locators through its notes system: a DOI may print as a DOI or a full link at the end of a note or bibliography entry, and Chicago's guidance favors DOIs over URLs for journal articles exactly as the other styles do. Turabian, the classroom edition, teaches the same preference. The cross-style constant is worth stating plainly: no current style wants a database's session URL, and every current style prefers a DOI when the work has one, which makes this the rare citation question with one answer.

Finding the DOI when it hides

Publisher pages usually print the DOI beside the abstract or under the article's header information, and database records carry it in a labeled field — look for 'DOI' in EBSCO's detailed record or ProQuest's citation tools. Crossref's free search at search.crossref.org resolves a DOI from an article title, and Google Scholar often shows it under the result. If every route comes up empty, the article probably lacks one, which is a fact about older publishing rather than a search failure. Capture it the moment you confirm the source is usable: paste the DOI into your notes beside the page numbers, because hunting it again during citation formatting is the task everyone underestimates until it costs an hour.

Then decide, by the chain: DOI if it exists, stable publisher URL if it does not, database name or permalink as the last resort. The citation generator on this page handles the formatting once you paste the right locator, which is why the searching is the part worth doing carefully — a generator fed a proxy link produces a polished reference to a dead address. Fix the locator first; the format is mechanical, and the grader's click is the real test.

Three minutes to a clean citation

Edge-case citations punish hesitation, not ignorance: the student who spends twenty minutes googling a no-author format loses more time to doubt than the format ever costs. Decide the pattern once — title moves forward, organization counts as author, n.d. when no date exists anywhere — and apply it mechanically. Consistency across your list matters more to graders than perfect recall of manual arcana.

Automate what's mechanical. The citation generator below formats journal articles, books, chapters and websites in all three major styles, including the awkward cases on this page. Paste the metadata, copy the output, done — the minutes you save belong to analysis, which is where rubric points actually live. Cross-check in-text forms against the table above and the reference list is submission-ready.