How many databases should I search?

At least two for a paper that claims coverage: one multidisciplinary platform and one subject index in the field (PubMed, ERIC, PsycInfo, MLA International Bibliography). Citation indexes (Web of Science, Scopus) are a third pass when you need who cited a paper. One box in the discovery layer is a start, not a method.

Do I cite the database I searched?

It depends on the style. APA 7 §9.30 omits Academic Search, ProQuest, and JSTOR for ordinary journal articles and uses a DOI when one exists — Hallion et al. (2019) is the official specimen. MLA 9 names the database as a second container; Goldman ends with JSTOR and a stable URL. CMOS 18 prefers a DOI URL and otherwise lists the database name. Never cite "EBSCO" as the author of the article.

When does APA 7 require the database name?

When the work is proprietary (UpToDate, Cochrane Database of Systematic Reviews) or of limited circulation (many dissertations, ERIC documents, some JSTOR primary sources). Stein and Taylor (2019) keeps UpToDate and a retrieval date. Risto (2014) keeps ProQuest Dissertations and Theses Global. Riegelman and Albertine (2008) keeps ERIC. Ordinary journal articles from Academic Search do not.

Why can't I open some articles?

Your library did not license that journal year, or the database only indexed the citation. Use Find It, the catalog, Unpaywall, or interlibrary loan. The steps are on How to access academic journals. A publisher Buy PDF button is often a journal your tuition already covers.

How do I know a database result is peer-reviewed?

The Peer reviewed limiter is a journal-level flag. Open the PDF and check whether the item is a research article, review, letter, or editorial. Confirm on the journal site or in Ulrichsweb. Preprints in a database results list are still preprints.

Can I use library databases from home?

Yes, through the library's current remote-access path — OpenAthens, a new EZproxy host, or a VPN. Sign in from the A–Z list, not from a bookmarked publisher URL. Harvard's ezp-prod1 bookmarks stopped resolving after 31 May 2026. Alumni and walk-in privileges are narrower than a current student login.

Should I use Google instead of a library database?

Google is fine for vocabulary. For the set you will cite, use a licensed database or PubMed/ERIC and then Google Scholar for Cited by. Scholar has no peer-review filter and no MeSH. The operator page is How to use Google Scholar.

What is the difference between a discovery layer and a database?

A discovery layer (HOLLIS, Primo, WorldCat Discovery) searches the catalog plus many licensed indexes at once. A named database (PsycInfo, CINAHL, MLA International Bibliography) is one curated index with its own thesaurus. Discovery is a wide first pass; the subject database is where you finish a literature search.

Are EBSCO field codes the same in ProQuest and PubMed?

No. EBSCO uses TI, AU, AB, SU. PubMed uses bracket tags such as [ti], [au], [mh], and [ad] — UCLA[ad] is the User Guide's affiliation example. ProQuest and JSTOR expect Advanced Search dropdowns. EBSCO's own help says field codes are database-specific.

Should I always turn on the full-text limiter?

No. It hides citation-only records your resolver or interlibrary loan can fill. Search first, then filter to full text only when you need something you can read in the next hour.

How to Use Library Databases (EBSCO, JSTOR, PubMed) [2026]

How to Use Library Databases: Quick Answer

Open your library's database A–Z list, pick a multidisciplinary platform or a subject index, and type AND, OR, and NOT in uppercase. Put a known title in quotes. Use the vendor's field codes or Advanced Search dropdowns — TI and AU are EBSCO abbreviations, not a universal language. Apply Peer reviewed only when the assignment wants journal articles. Export a RIS or BibTeX file; never paste a session URL into a paper.

Copy-ready official search strings

EBSCO AND (official example)

travel AND Europe

EBSCO OR (official example)

college OR university

EBSCO NOT (official example)

television NOT cable

EBSCO nested (official example)

(heart OR lung) AND bypass

EBSCO field-code grouping (official example)

AU (Smith AND Peters NOT Lee)

PubMed (official translation)

vitamin c common cold

PubMed affiliation tag (official example)

UCLA[ad]

Harvard HOLLIS nested (official example)

("Affordable Care Act" OR Obamacare) AND "birth control"

What this page is for

Use this URL when you are already inside a licensed product — EBSCOhost, ProQuest, JSTOR, PubMed, Web of Science — and need the buttons, limiters, thesaurus, and export. How to find scholarly sources is which tool to open and how to chain citations. How to use Google Scholar is Scholar's author: operator and Cited by. How to access academic journals is the paywall after you have a citation. Scholarly vs popular sources classifies the item. Keep those jobs off this page.

What a library database actually is

A database is an organized collection of records: title, authors, source, abstract, controlled subject terms, and often a full-text file. The Association of College and Research Libraries treats that organization as a skill, not a backdrop. Two knowledge practices from Searching as Strategic Exploration in the ACRL Framework (adopted 11 January 2016) are the rule for this page:

ACRL Framework, Searching as Strategic Exploration — knowledge practices

Searching for information is often nonlinear and iterative, requiring the evaluation of a range of information sources and the mental flexibility to pursue alternate avenues as new understanding develops.

match information needs and search strategies to appropriate search tools

understand how information systems (i.e., collections of recorded information) are organized in order to access relevant information

use different types of searching language (e.g., controlled vocabulary, keywords, natural language) appropriately

That is why typing the assignment title into one box and taking the first five PDFs fails even when the PDFs are real articles. The database already tagged those articles with a thesaurus term you have not used yet.

Choose the database before you type

Libraries license products, not "the internet." Academic Search Complete, Academic Search Premier, and Academic Search Ultimate are different EBSCO packages with different title lists. JSTOR is an archival platform, not a current-awareness feed. PubMed is a free NLM index; your PDF still comes from a library resolver. The old habit of treating every A–Z name as interchangeable is how students miss the thesaurus that would have found the paper.

Which licensed tool for which job
Job	Open this first	Do not expect it to do
A first pass across disciplines	EBSCO Academic Search (your campus's Complete / Premier / Ultimate package) or ProQuest Central	Replace PubMed in medicine or ERIC in education
Humanities back files, books, and primary sources	JSTOR	Catch this week's clinical trial
Biomedical citations and MeSH	PubMed	Guarantee a PDF on every row
Education research and ERIC documents	ERIC or the library's ERIC-on-EBSCO interface	Index every STEM journal
Psychology descriptors	APA PsycInfo / PsycARTICLES through the library	Work as a news database
Citation counts and cited-reference search	Web of Science or Scopus	Be a full-text warehouse

Multidisciplinary platforms

Start here when you do not yet know which field owns the question. Use the vendor's current product page for title counts — they change, and an old "10,000 journals" line is not a fact you should repeat.

EBSCO Academic Search family. Academic Search Complete is EBSCO's mid-tier multidisciplinary full-text database (the product page currently lists 5,601 active full-text journals). Your library may have Premier or Ultimate instead — check the A–Z name.
ProQuest Central. ProQuest still describes it as 47 of its complete databases across more than 175 subjects — journals, trade titles, news, dissertations, and market reports on one platform. That is 47 databases, not 47,000 sources. Some campuses now license ProQuest Central Premium; the A–Z name is the one that matters.
JSTOR. ITHAKA's archival platform for journals, books, and primary sources. Coverage often runs to first issues; a 1924 hit can be correct for history and stale for a 2025 clinical guideline.
Gale Academic OneFile. News, magazines, and journals. Useful when the assignment allows mixed source types; still apply a peer-review limiter if you need only research articles.
Web of Science / Scopus. Citation indexes. Use them to follow who cited a paper, not as your only full-text store.

Subject databases

These are the tools that know the field's vocabulary. A multidisciplinary search that "worked" can still miss the paper tagged only in the subject index.

Health and medicine: PubMed, MEDLINE, CINAHL, Cochrane Library
Psychology: APA PsycInfo, PsycARTICLES
Education: ERIC, Education Source
Business: Business Source Complete, ABI/INFORM
Sciences: SciFinder, ScienceDirect, Scopus
Engineering: IEEE Xplore, Compendex
Humanities: MLA International Bibliography, Project MUSE, JSTOR
Social sciences: SocINDEX, Sociological Abstracts
Law: Nexis Uni, Westlaw, HeinOnline

Get in from the library, not the publisher homepage

Off campus, the same ScienceDirect URL that worked on wifi will show a paywall because the browser never passed through the library. Open the database from the library A–Z list, then authenticate with OpenAthens, EZproxy, or a campus VPN. Harvard Library retired EZproxy on 31 May 2026 and moved remote login to OpenAthens — bookmarked ezp-prod1.hul.harvard.edu links now error. The full access order (resolver, Unpaywall, interlibrary loan) lives on How to access academic journals.

Go to the library website and open Databases, Articles, or A–Z Databases.
Browse by subject or search the database name.
Click the name so the session starts on a licensed URL.
Sign in with campus credentials if the product asks.
If a publisher page still paywalls, search the title in the catalog or click Institutional login — do not pay the Buy PDF button yet.

Boolean operators: EBSCO and PubMed wording

Boolean logic is how a database combines sets. EBSCO's help article is the wording to use on Academic Search, CINAHL, and most EBSCOhost indexes. Type the operators in uppercase; some libraries configure EBSCO so lowercase and is treated as a word.

EBSCO, Searching with Boolean Operators

AND

AND combines search terms so that each search result contains all of the terms. For example, travel AND Europe finds articles that contain both travel and Europe.

OR combines search terms so that each search result contains at least one of the terms. For example, college OR university finds results that contain either college or university.

NOT

NOT excludes terms so that each search result does not contain any of the terms that follow it. For example, television NOT cable finds results that contain television but not cable.

AND before OR

When executing a search, AND takes precedence over OR.

Parentheses

When you enter (mouse OR rat) AND trap, the search engine retrieves results containing the word mouse or the word rat together with the word trap in the fields searched by default.

Why the parentheses matter

For example, (heart OR lung) AND bypass will return different results than heart OR lung AND bypass.

Without parentheses, EBSCO processes AND and NOT first, then OR — the help page's own expansion of mouse OR rat AND trap NOT mousetrap is mouse OR ((rat AND trap) NOT mousetrap). Put the synonym group in parentheses every time you mix operators.

PubMed is not EBSCO. It applies AND between bare concepts and requires uppercase operators. The PubMed User Guide is the official translation:

PubMed User Guide, Combining search terms with Boolean operators

PubMed applies an AND operator between concepts, e.g., "vitamin c common cold" is translated as vitamin c AND common cold. Enter Boolean operators in uppercase characters to combine or exclude search terms:

AND retrieves results that include all the search terms.

OR retrieves results that include at least one of the search terms.

NOT excludes the retrieval of terms from your search.

PubMed processes searches in a left-to-right sequence. Use parentheses to "nest" concepts that should be processed as a unit and then incorporated into the overall search.

Phrases, truncation, and wildcards

Quotation marks hold a phrase together. Truncation and wildcards expand a stem — and the symbol is not the same in every product. Harvard Library's HOLLIS search-operator page (last updated 19 August 2026) documents child* and wom?n for HOLLIS. PubMed uses * only after at least four characters.

Official truncation and wildcard strings

EBSCO / most discovery layers — exact phrase

"machine learning"

Harvard HOLLIS exact phrase (official)

"The color purple"

Harvard HOLLIS truncation (official)

child*

Harvard HOLLIS single-character wildcard (official)

wom?n

PubMed User Guide, Wildcards and truncation

To create a wildcard search in PubMed, use an asterisk (*) to substitute for 0 or more characters in a term or phrase.

PubMed minimum stem

Terms must begin with at least 4 characters before the first wildcard: colo*

Harvard Library, Using Search Operators

HOLLIS will only recognize Boolean operators (AND, OR, NOT) that are in ALL CAPS.

Copy the stem from the product's help page. EBSCO, HOLLIS, and PubMed all use *, but PubMed refuses a stem shorter than four characters, and a wildcard turns off PubMed's Automatic Term Mapping. The User Guide's own warning is that "heart attack*" will not map to Myocardial Infarction.

Field codes are vendor-specific

Students copy TI climate change into every box. That syntax is EBSCO's. EBSCO's field-code help says codes are database-specific and must be uppercase so they are not read as words:

EBSCO, Searching with Field Codes

EBSCO uses two-character abbreviations for field codes such as SU-Subject, AU-Author, or TI- Article_Title.

When entering searches in the search box, use lowercase for search terms. Use uppercase only for field codes and Boolean operators (AND, OR, NOT).

Official grouping example

AU (Smith AND Peters NOT Lee)

On EBSCOhost, the common codes are TI (title), AU (author), AB (abstract), SU (subject terms), and SO (source / journal name). PubMed uses square-bracket tags instead — UCLA[ad] for affiliation is the User Guide's example. ProQuest and JSTOR expose the same idea as Advanced Search dropdowns. If you are not sure, use the dropdown rather than guessing a two-letter code.

Subject headings and the thesaurus

A keyword search matches words anywhere. A subject heading is a tag an indexer assigned. ACRL's practice is to use both. The workflow is the same on EBSCO, ProQuest, and PubMed:

Run a keyword search that returns at least one relevant record.
Open the record and copy the subject headings, descriptors, or MeSH terms.
Click a heading, or paste it into a Subject / SU / [mh] search.
Open the database thesaurus (CINAHL Headings, APA Thesaurus, MeSH Database, ERIC Thesaurus) for broader, narrower, and related terms.
Rebuild the search with the heading OR the keywords you still need.

PubMed User Guide, Search PubMed using the MeSH database

MeSH(Medical Subject Headings) is the NLM controlled vocabulary thesaurus used for indexing PubMed citations.

Use the MeSH database to find MeSH terms, including Subheadings, Publication Types, Supplementary Concepts and Pharmacological Actions, and then build a PubMed search.

Affiliation field tag

You can search for a term in a specific field by including a search field tag after the term; for example, UCLA[ad] will search for the term “UCLA” in the affiliation field only.

A keyword search for "heart attack" and a MeSH search for Myocardial Infarction are not the same set. PubMed's Automatic Term Mapping often adds the MeSH term for you; a quoted phrase or a wildcard turns that mapping off. Check Search Details before you trust the first page.

Filters that help — and filters that hide good work

Limiters sit in the sidebar after the search. Use them to correct a result set, not to replace a designed string. The Peer reviewed / Scholarly journals box is a journal-level flag: it can still return editorials, letters, and book reviews from a refereed journal, and it will hide books, newspapers, and many reports you may need.

When to touch each limiter

Peer reviewed / Scholarly. Use for a research-article assignment. Confirm the item type on the PDF. Ulrichsweb, if your library licenses it, checks the journal. See peer review.
Full text. Convenient and incomplete. It hides citation-only records your resolver or interlibrary loan can still fill.
Date. Use for a claim that expires (a 2024 drug trial, last quarter's market). Do not apply a five-year cap to a history paper or a methods classic.
Source type. Academic journals versus magazines, newspapers, dissertations, reports. A current-events paper needs the news file and a scholarly analysis.
Language. Limit to languages you can read. Do not assume English is the whole literature.
Subject. Use after you have seen which headings the database actually assigned.

PubMed's species, sex, and age filters only apply to MEDLINE-indexed records that received those MeSH terms. The User Guide warns they exclude preprints and some ahead-of-print citations. That is a precision tool, not a default.

Full text, Find It, and citation-only hits

A database row is a record, not a promise of a PDF. HTML full text, a native PDF, a Find It / Check for full text resolver button, and a citation-only line are four different outcomes. If the PDF is missing, leave this page and follow How to access academic journals: catalog, Unpaywall, PubMed Central, interlibrary loan, author copy — in that order.

Click PDF Full Text or HTML Full Text when it is on the record.
Use the library's Find It / OpenURL button if the vendor only indexed the article.
Search the title in the catalog or Google Scholar with library links on.
Request interlibrary loan with the DOI if you have one.
Do not treat a publisher "Buy PDF" button as the only legal copy.

Sort the list after you have a usable set. Relevance is the default and is what most assignments want. Date (newest) is for current methods, not for a literature's origin story. Times cited (Web of Science, Scopus, some EBSCO rows) surfaces influential work and older reviews — it is not a quality stamp. Related articles / Similar articles (PubMed's Similar articles uses title, abstract, and MeSH) is a way to grow one good hit, not a substitute for a designed string.

The old shortcut "never name the database" is APA's default, not a universal rule. APA 7 §9.30 — reprinted by Alamo Colleges when the Style Center is behind a login — omits the database for most journal articles, books, and chapters because those works are widely available. Name it for proprietary works (UpToDate, Cochrane Database of Systematic Reviews) and for limited-circulation items (many dissertations, ERIC documents). MLA 9 treats the database as a second container. CMOS 18 prefers a DOI URL and otherwise allows the database name.

APA Style, Database Information in References (Alamo reprint of §9.30)

Database information is seldom provided in reference list entries.

Do not include database information for works obtained from most academic research databases or platforms because works in these resources are widely available.

Likewise, do not include URLs from these academic research databases in reference list entries because these URLs will not resolve for readers.

If a widely available work (e.g., journal article, book, book chapter) from an academic research database does not have a DOI, treat the work as a print version.

Provide the name of the database or archive when it publishes original, proprietary works available only in that database or archive (e.g., UpToDate or the Cochrane Database of Systematic Reviews).

If the URL is session-specific (meaning it will not resolve for readers), provide the URL of the database home page or login page instead.

The official APA examples below are real works. Hallion et al. (2019) is a Sport, Exercise, and Performance Psychology article with DOI 10.1037/spy0000150 — APA names neither PsycINFO nor EBSCOhost. Stein and Taylor (2019) is an UpToDate topic that exists only there, so the database name and a retrieval date stay. Risto (2014) is Angela Risto's Tennessee State University dissertation, ProQuest publication number 3683242; the Style Center shortens the title to "social media and texting," while the TSU / ProQuest record reads "Texting and Social Media." Quote the official specimen when you are matching APA's example; use the title on the PDF when you are citing the dissertation itself. Riegelman and Albertine (2008) is ERIC document ED504790, still at https://files.eric.ed.gov/fulltext/ED504790.pdf.

Official APA 7 §9.30 examples (keep English)

Journal from an ordinary research database (Hallion et al., 2019)

Hallion, M., Taylor, A., Roberts, R., & Ashe, M. (2019). Exploring the association between physical activity participation and self-compassion in middle-aged adults. Sport, Exercise, and Performance Psychology, 8(3), 305–316. https://doi.org/10.1037/spy0000150

Proprietary database (Stein & Taylor, 2019)

Stein, M. B., & Taylor, C. T. (2019). Approach to treating social anxiety disorder in adults. UpToDate. Retrieved September 13, 2019, from https://www.uptodate.com/contents/approach-to-treating-social-anxiety-disorder-in-adults

Limited-circulation dissertation (Risto, 2014)

Risto, A. (2014). The impact of social media and texting on students' academic writing skills (Publication No. 3683242) [Doctoral dissertation, Tennessee State University]. ProQuest Dissertations and Theses Global.

ERIC monograph (Riegelman & Albertine, 2008)

Riegelman, R. K., & Albertine, S. (2008). Recommendations for undergraduate public health education (ED504790). ERIC. https://files.eric.ed.gov/fulltext/ED504790.pdf

MLA Style Center — official database examples (keep English)

MLA Style Center, Citations by Format — Journal Article in a Database

Goldman, Anne. “Questions of Transport: Reading Primo Levi Reading Dante.” The Georgia Review, vol. 64, no. 1, spring 2010, pp. 69-88. JSTOR, www.jstor.org/stable/41403188.

MLA Style Center, Databases as containers (24 January 2022)

MLA style encourages writers to document the facts they observe.

Official EBSCOhost container example

Corral, Eduardo C. “Watermark.” New England Review, vol. 30, no. 4, fall 2009, pp. 24–25. EBSCOhost, search.ebscohost.com.

The Goldman essay is a real Georgia Review article (Spring 2010, 64.1, pp. 69–88). MLA names JSTOR as container 2. APA 7 of the same article ends after the page range because no DOI is published and Academic Search / JSTOR are ordinary research databases:

Same real article, APA 7 default (no database name)

Goldman, A. (2010). Questions of transport: Reading Primo Levi reading Dante. The Georgia Review, 64(1), 69–88.

CMOS 18 Notes and Bibliography Quick Guide — journal articles

For articles consulted online, include a URL (preferably one based on a DOI; see CMOS 13.7); alternatively, list the name of the database.

Same page, books consulted online

To cite a book consulted online, include either a URL or the name of the database.

Applied to Goldman, CMOS 18 therefore prefers a stable JSTOR URL or the database name — not an ezproxy session string. Do not invent a DOI; The Georgia Review did not assign one.

Same real article, CMOS 18 applied (not an official CMOS specimen)

Notes-Bibliography bibliography

Goldman, Anne. "Questions of Transport: Reading Primo Levi Reading Dante." The Georgia Review 64, no. 1 (Spring 2010): 69–88. JSTOR.

Do you name the database in the reference?
Style	Ordinary journal from Academic Search / JSTOR	When the database does belong
APA 7 §9.30	No. DOI if one exists (Hallion); otherwise stop after the pages (Goldman).	UpToDate (Stein), Cochrane, ERIC documents (Riegelman), dissertations of limited circulation (Risto)
MLA 9	Yes. Journal is container 1; JSTOR, EBSCOhost, or the named database is container 2 plus a DOI, permalink, or URL.	Always, when that is the version you consulted
CMOS 18	DOI URL preferred. If the URL is a session link, write the database name instead.	Subscription databases whose URLs will not resolve for a reader

Never paste an ezproxy…/login?url= or search.ebscohost.com session string into APA as if it were a DOI. Use the DOI from the record, a JSTOR stable URL in MLA, or the database name in Chicago. How to find a DOI is the lookup. Format the finished entry in the citation generator from the article record, not from the search box.

Export, permalinks, folders, and alerts

Vendor Cite buttons are a draft. They drop italics, invent a retrieval date, and sometimes emit a session URL. Export RIS or BibTeX into Zotero, Mendeley, EndNote, or RefWorks, then open the record and fix the DOI, page range, and article type. Import citations covers the file formats.

Select the records you will actually read.
Click Cite, Export, or Save and choose RIS or BibTeX — not "email me a pretty APA string."
Import the file into a reference manager.
Copy the Permalink / Stable URL / JSTOR stable/ link if MLA or CMOS needs a location. Session URLs die. See permalink.
Create a search alert or a Cited by alert only after the string is worth repeating.

EBSCOhost and ProQuest folders are session or account storage, not a bibliography. Items in an unsigned folder vanish. Download the PDF when you find it, keep the DOI in the filename, and back up the folder. Zotero can attach the file to the imported record.

Common database mistakes

One box, no operators. A discovery layer may AND your words; PubMed will too, but EBSCO may parse lowercase and as a term. Write AND / OR / NOT in uppercase.
Mixing OR and AND without parentheses. EBSCO processes AND before OR. (heart OR lung) AND bypass is not heart OR lung AND bypass.
Too broad. education is not a search. Add a population, a method, or a heading.
Too narrow. Four AND clauses and a five-year limit on a thin topic returns zero. Drop the weakest limiter first.
Ignoring the thesaurus. You will miss papers tagged Myocardial Infarction if you only typed "heart attack."
Treating Peer reviewed as a quality stamp. It is a journal flag. Read the article type.
Citing the database in every style. APA usually omits it; MLA usually includes it; CMOS wants a DOI or the database name — not a dead proxy URL.
Stopping in one product. Search a subject database after the multidisciplinary pass. Indexes do not overlap perfectly.

Search strategies by assignment type

Literature review

Open the subject database for the field, not only Academic Search.
Build a heading-plus-keyword string and save it.
Use Cited by / Times cited on two or three central papers.
Search a second index (PubMed plus CINAHL; MLA plus JSTOR) so you are not hostage to one vendor's coverage.
The writing job is on How to write a literature review.

Current events

Use a news file (Nexis Uni, ProQuest news, Newspaper Source) with a recent date limit.
Read more than one outlet.
Then search a scholarly database for analysis — a newspaper is not a methods paper.

Historical research

Use JSTOR, historical newspaper archives, and primary-source collections.
Do not apply a "last five years" limiter; older items are the evidence.
Subject headings in a catalog (Library of Congress) are often more precise than article keywords.
See Primary sources for history.

Frequently asked questions

How many databases should I search?: At least two for a paper that claims coverage: one multidisciplinary platform and one subject index in the field (PubMed, ERIC, PsycInfo, MLA International Bibliography). Citation indexes (Web of Science, Scopus) are a third pass when you need who cited a paper. One box in the discovery layer is a start, not a method.
Do I cite the database I searched?: It depends on the style. APA 7 §9.30 omits Academic Search, ProQuest, and JSTOR for ordinary journal articles and uses a DOI when one exists — Hallion et al. (2019) is the official specimen. MLA 9 names the database as a second container; Goldman ends with JSTOR and a stable URL. CMOS 18 prefers a DOI URL and otherwise lists the database name. Never cite "EBSCO" as the author of the article.
When does APA 7 require the database name?: When the work is proprietary (UpToDate, Cochrane Database of Systematic Reviews) or of limited circulation (many dissertations, ERIC documents, some JSTOR primary sources). Stein and Taylor (2019) keeps UpToDate and a retrieval date. Risto (2014) keeps ProQuest Dissertations and Theses Global. Riegelman and Albertine (2008) keeps ERIC. Ordinary journal articles from Academic Search do not.
Why can't I open some articles?: Your library did not license that journal year, or the database only indexed the citation. Use Find It, the catalog, Unpaywall, or interlibrary loan. The steps are on How to access academic journals. A publisher Buy PDF button is often a journal your tuition already covers.
How do I know a database result is peer-reviewed?: The Peer reviewed limiter is a journal-level flag. Open the PDF and check whether the item is a research article, review, letter, or editorial. Confirm on the journal site or in Ulrichsweb. Preprints in a database results list are still preprints.
Can I use library databases from home?: Yes, through the library's current remote-access path — OpenAthens, a new EZproxy host, or a VPN. Sign in from the A–Z list, not from a bookmarked publisher URL. Harvard's ezp-prod1 bookmarks stopped resolving after 31 May 2026. Alumni and walk-in privileges are narrower than a current student login.
Should I use Google instead of a library database?: Google is fine for vocabulary. For the set you will cite, use a licensed database or PubMed/ERIC and then Google Scholar for Cited by. Scholar has no peer-review filter and no MeSH. The operator page is How to use Google Scholar.
What is the difference between a discovery layer and a database?: A discovery layer (HOLLIS, Primo, WorldCat Discovery) searches the catalog plus many licensed indexes at once. A named database (PsycInfo, CINAHL, MLA International Bibliography) is one curated index with its own thesaurus. Discovery is a wide first pass; the subject database is where you finish a literature search.
Are EBSCO field codes the same in ProQuest and PubMed?: No. EBSCO uses TI, AU, AB, SU. PubMed uses bracket tags such as [ti], [au], [mh], and [ad] — UCLA[ad] is the User Guide's affiliation example. ProQuest and JSTOR expect Advanced Search dropdowns. EBSCO's own help says field codes are database-specific.
Should I always turn on the full-text limiter?: No. It hides citation-only records your resolver or interlibrary loan can fill. Search first, then filter to full text only when you need something you can read in the next hour.

Format the article, not the search session

Export RIS or BibTeX from the database, then generate the citation from the article record. APA usually omits the database (Hallion); MLA names it as a container (Goldman, Corral). Do not paste a session URL into the paper.

Citation generator Import BibTeX or RIS