How many scholarly sources do I need?

The assignment sheet wins. Common undergraduate ranges are about 5–10 sources for a short paper and 10–20 for a longer one; graduate work and theses run much higher. Count usable, cited sources, not every PDF you downloaded. A literature-review assignment is asking for coverage, not a lucky seven.

Can I use Google instead of a library database?

You can start there to learn the vocabulary. For the set you will cite, use a library catalog or a subject database (PubMed, ERIC, PsycINFO, MLA International Bibliography) and then Google Scholar for Cited by. Cornell Library's search-strategy page states the difference: databases run keyword searches across peer-reviewed literature; Google runs natural-language search across web pages. See How to use Google Scholar for the operators.

How do I know an article is peer-reviewed?

Check the article type on the PDF (research article versus letter or editorial), then the journal's site, then Ulrichsweb if your library has it. A database Peer reviewed filter is a journal-level flag — ERIC's May 2024 Selection Policy says so in those words. Preprints on arXiv or SSRN are not peer-reviewed until a journal says they are.

Should I use only sources from the last five years?

Only if the claim expires. A date limiter is a way to see current approaches, not a ban on older work. Keep foundational books and methods papers; refresh clinical, technology, and statistics claims.

What if my library does not subscribe to the database I need?

Ask a librarian for an equivalent index, a consortial login, or interlibrary loan. PubMed and ERIC remain free on their official sites. Document delivery exists for a single article you have already identified.

Why did PubMed return different results than Google Scholar for the same words?

PubMed maps terms to MeSH and requires uppercase AND, OR, and NOT. Scholar searches a wider mix of literature and does not use MeSH. The official PubMed example vitamin c common cold is translated as vitamin c AND common cold — Scholar has no such translation pane. Scholar also caps the viewable result list at 1,000 and does not disclose how those 1,000 are selected.

What is citation chaining, and is it enough by itself?

It is following references backward and Cited by forward from one good paper. McMaster Library warns that using only chaining hides work that never cited your pearl. Cochrane's Chapter 4 technical supplement (September 2024) adds that forwards citation searching should be an adjunct, not the whole search, because researchers may cite positive studies selectively. Pair chaining with a designed database search.

Do I need MEDLINE, Embase, and CENTRAL for a class paper?

No. Cochrane MECIR C24 names that trio as the minimum for a Cochrane intervention review. For an undergraduate paper, pick the one or two tools that match the question (PubMed or CINAHL for clinical nursing; ERIC for education; MLA International Bibliography for literature) and chain from one good hit. Use the Cochrane rule only when your methods section claims a systematic search.

What does PubMed proximity searching do that AND does not?

AND finds the words anywhere in the record. Proximity finds them in one field within N words of each other, in any order. The official format is "search terms"[field:~N]. PubMed allows it only in Title, Title/Abstract, and Affiliation. It ignores the operator if you put a wildcard inside the quotes.

I found Ioannidis 2005. Can I stop there?

No. Use it as a pearl: open two items from its reference list and a handful of later papers that cite it. Then evaluate the argument. The official citation is Ioannidis JPA (2005) Why Most Published Research Findings Are False. PLoS Med 2(8): e124. https://doi.org/10.1371/journal.pmed.0020124. The 25 August 2022 correction is 10.1371/journal.pmed.1004085.

How to Find Scholarly Sources: Library Search Workflow [2026]

How to find scholarly sources: quick answer

Log into your library, open a discovery catalog or a subject database, and search with uppercase Boolean operators. Use AND to require every concept, OR plus parentheses for synonyms, and quotes for a known title. If the first page is noise, change the tool or the vocabulary — do not keep paging. When one article is right, chain backward through its reference list and forward through Cited by.

Copy-ready search templates from official help pages

PubMed (official translation)

vitamin c common cold

Same search, operators written out

vitamin c AND common cold

Harvard HOLLIS (official nested example)

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

MIT Libraries (official AND example)

cloning AND humans AND ethics

MIT Libraries (official nested example)

ethics AND (cloning OR reproductive techniques)

PubMed proximity (official combine example)

"hip pain"[Title:~4] AND stretching

Google Scholar Search Help (official author example)

author:"d knuth"

Google Scholar Search Help (official title example)

"A History of the China Sea"

What this page is for

Use this URL when the job is locating research: which tool, which operators, which subject vocabulary, and how to grow one good hit into a set. Scholarly vs popular sources classifies the item after you have it. How to evaluate sources runs CRAAP on a named source. How to use library databases is the database interface in depth. How to use Google Scholar is Scholar's operators, library links, and Cite button. How to access academic journals is paywalls, interlibrary loan, and repositories. Keep those jobs off this page.

Official rule: searching is strategic exploration

The Association of College and Research Libraries filed the Framework on 2 February 2015 and adopted it on 11 January 2016. The frame that governs this page is Searching as Strategic Exploration. The opening sentence is the rule, not a slogan:

ACRL Framework for Information Literacy for Higher Education (2016)

Searching as Strategic Exploration

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.

Novice versus expert searchers

Novice learners may search a limited set of resources, while experts may search more broadly and deeply to determine the most appropriate information within the project scope. Likewise, novice learners tend to use few search strategies, while experts select from various search strategies, depending on the sources, scope, and context of the information need.

That is why "I typed the assignment title into Google and took the first five PDFs" fails even when the PDFs are real articles. ACRL's knowledge practices for this frame include matching the need to a search tool, using controlled vocabulary and keywords, and refining the strategy from the results. The Framework's definition of information literacy starts with discovery:

ACRL definition of information literacy

Information literacy is the set of integrated abilities encompassing the reflective discovery of information, the understanding of how information is produced and valued, and the use of information in creating new knowledge and participating ethically in communities of learning.

What counts as a scholarly source

A scholarly source is written by researchers for researchers: named authors, methods or argument, and a reference list. Peer review is common for journal articles and uncommon for most books, dissertations, and preprints. Confirm the type on Scholarly vs popular sources before you spend an evening searching the wrong container.

Peer-reviewed journal article. Original study or review in a research journal. Best for a current empirical claim.
Scholarly book or chapter. University-press monograph or an essay in an edited volume. Best for theory, history, and a field's long argument.
Dissertation or thesis. Graduate research. Useful for a literature map and methods detail; check whether a later book or article superseded it.
Conference paper. Faster than a journal in some STEM fields; quality varies by venue.
Preprint or working paper. arXiv, SSRN, or a lab site. Label it as a preprint until a version of record exists. See grey literature.

Pick the right search tool first

ACRL's practice is to match the need to the tool. A discovery catalog is a wide first pass. A subject database is the tool that knows the field's vocabulary. Google Scholar is a free citation index, not a peer-review filter. Open indexes are for items that publishers have already released.

Which tool for which job
Job	Start here	Then go here if you need more
A book, edition, or known title your library owns	Library catalog / discovery (HOLLIS, WorldCat, Primo)	Interlibrary loan if the catalog record has no local copy
Biomedical or clinical articles	PubMed	CINAHL, Cochrane Library, or Embase through the library
Education research	ERIC (IES / U.S. Department of Education)	Education Source or a vendor ERIC interface with extra full text
Psychology	PsycINFO / APA PsycNet through the library	PubMed if the claim is clinical
Literature and languages	MLA International Bibliography	Library catalog for the book the article is about
A known article title, then who cited it	Google Scholar	Web of Science, Scopus, or OpenAlex for citation data
A paper you can read with no login	PubMed Central, DOAJ, arXiv	How to access academic journals for the rest

Library catalog or discovery layer

Harvard Library's HOLLIS is a typical discovery layer: catalog records plus a large article index. The HOLLIS search-operator page (last updated 14 September 2026, 11:55 a.m.) is the wording this page treats as official for catalog syntax. Two rules students miss: Boolean words must be ALL CAPS, and the Peer-Reviewed Articles filter hides books on purpose.

Official HOLLIS operators (Harvard Library, last updated 14 September 2026)

Exact phrase

"The color purple"

Synonyms

NSA OR "National Security Agency"

Disambiguation

cellular NOT phone

Require both terms

genetic AND variation

Harvard Library, Using Search Operators

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

Your campus tool has a different name — SearchWorks, Articles+, OneSearch — and the same job. Start there when you need a book or a panoramic sweep. Switch to a subject database when the discovery set is a mix of news, reviews, and off-topic hits.

Subject databases

A subject database is curated for one field and usually offers a thesaurus, a peer-review limiter, and export to a citation manager. Multidisciplinary products (Academic Search, JSTOR, Scopus, Web of Science, ProQuest) are fine first stops. They are not a substitute for PubMed in medicine or ERIC in education. The U.S. Institute of Education Sciences describes ERIC this way on its official database page (opened 17 September 2026):

Institute of Education Sciences, Education Resources Information Center (ERIC)

ERIC is an internet-based digital library of education research and information sponsored by the Institute of Education Sciences (IES) of the U.S. Department of Education, providing access to bibliographic records of journal and non-journal literature from 1966 to the present.

ERIC Selection Policy (May 2024)

ERIC accepts peer-reviewed and non-peer-reviewed material.

ERIC Selection Policy, journal-level flag

For journals, the peer-review designation is determined at the journal level and applied to all ERIC records created for the source.

ERIC Selection Policy, editorial review

Editorial reviews are not recognized as an acceptable type of peer review.

The same IES page lists what the collection holds: journal articles, books, research syntheses, conference papers, technical reports, policy papers, and other education-related materials. Free official indexes you can open without a campus login include PubMed and ERIC. Library-licensed indexes that still matter: PsycINFO, MLA International Bibliography, SciFinder, IEEE Xplore, EconLit, Historical Abstracts, and America: History and Life. JSTOR is an archive plus current content for some titles — a hit there can be 95 years old and still correct for history, and stale for a current clinical guideline.

Cochrane 6.5.1: one database is not enough for a systematic search

A class paper does not need the full Cochrane search. A systematic review does. Chapter 4 of the Cochrane Handbook (Lefebvre et al., last updated March 2025; Cochrane Handbook for Systematic Reviews of Interventions version 6.5.1, 2025) is the official statement that bibliographic databases are not interchangeable and that topic — not habit — picks the file. The Chapter 4 technical supplement (PDF last updated September 2024) is the citation-searching detail.

Cochrane Handbook 6.5.1, Chapter 4 Searching for and selecting studies (March 2025)

§4.3.1 Bibliographic databases

Database selection should be guided by the review topic.

§4.3.1 Bibliographic databases

Searching two or more databases lowers the risk of missing eligible studies.

MECIR C24 (Mandatory)

Searches for studies should be as extensive as possible in order to reduce the risk of publication bias and to identify as much relevant evidence as possible.

Chapter 4 key point, minimum health-database set

The Cochrane Central Register of Controlled Trials (CENTRAL) and MEDLINE, together with Embase (if access to Embase is available to the review team), should be searched for all Cochrane reviews.

MECIR C25 (Highly desirable)

Search appropriate national, regional and subject-specific bibliographic databases.

MECIR C32 (Mandatory)

Ensure correct use of the ‘AND’ and ‘OR’ operators.

Chapter 4 on NOT

The ‘NOT’ operator should be avoided where possible to avoid the danger of inadvertently removing records that are relevant from the search set.

MECIR C24 names CENTRAL, MEDLINE, and Embase (when Embase is available) as the minimum for a Cochrane intervention review. That is a review standard, not an undergraduate quota. The transferable lesson is the same ACRL practice: PubMed alone is not ERIC, ERIC is not the MLA International Bibliography, and Google Scholar is not a substitute for either.

Google Scholar — use it, do not stop there

Scholar is excellent for a known title and for Cited by. It is a poor only-tool because it mixes peer-reviewed articles with theses, slides, and citations it cannot guarantee. Official coverage and operators live on Google Scholar Search Help; the worked Scholar workflow is How to use Google Scholar. Set Library links in Scholar settings so a campus resolver appears next to hits you can actually open. Cochrane's Chapter 4 technical supplement (September 2024 PDF, opened 17 September 2026) states the reproducibility limit in those words:

Official Google Scholar Search Help examples, plus the Cochrane 1,000-result limit

Google Scholar Search Help, author search

Use the "author:" operator, e.g., author:"d knuth" or author:"donald e knuth".

Google Scholar Search Help, title search

Put the paper's title in quotations: "A History of the China Sea".

Cochrane Handbook 6.5 technical supplement (September 2024)

Finally, Google Scholar limits the number of viewable results to 1,000 and does not disclose how the top 1,000 results are selected, thus compromising the transparency and reproducibility of search results (Levay et al 2016).

The same supplement names OpenAlex (OurResearch, launched early 2022) as a free citation index alongside Web of Science, Scopus, and Scholar. Use OpenAlex or a licensed citation index when you need a transparent cited-by list; do not treat any of them as a peer-review filter.

Open indexes when you have no subscription

PubMed Central hosts free full-text biomedical articles. DOAJ lists quality open-access journals. arXiv is a preprint server for physics, mathematics, and computer science — not a journal. For legal routes around a paywall (resolver, interlibrary loan, author manuscript, repository), use How to access academic journals and Open access research. Do not use a pirated PDF site.

Build a search from the research question

The PubMed User Guide starts where every good search starts — concepts, then terms, then Search:

PubMed User Guide, How do I search PubMed?

Identify the key concepts for your search.

Enter the terms (or key concepts) in the search box.

Press the Enter key or click Search.

Cornell University Library's search-strategy page states why that sequence cannot be a Google sentence. Purdue University Libraries' search-strategy worksheet adds the step students skip: write 3–4 synonyms for each concept, because authors and databases tag the same study with different words, then pick the discipline that would bother to publish on the question — that choice is how you pick PubMed versus ERIC versus MLA.

Cornell University Library, Planning Your Search Strategy

In most cases, searching databases is different than searching Google. Typically, this is because databases are enabled to do powerful, research keyword searches across peer-reviewed literature, whereas Google performs natural language searching across web pages and freely-available documents.

Phrase your topic as a question, and then write it out in a sentence (make sure your topic is specific enough)
Break it into the important concepts
Be sure to think of as many synonyms or alternate terms as you can (use a thesaurus!)
Connect your concepts with AND and your similar terms with OR (use not to exclude common meanings you don't want)

Boolean operators (official wording)

MIT Libraries' Boolean operators page and the PubMed User Guide agree on the three operators. PubMed requires them in uppercase and applies AND between bare concepts:

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.

MIT Libraries official Boolean examples

AND (narrow)

cloning AND humans AND ethics

OR (synonyms)

cloning OR genetics OR reproduction

NOT (exclude an implied concept)

cloning NOT sheep

Implied AND in Google (MIT warning)

"college students" AND "test anxiety"

Nested OR

(ethic* OR moral*) AND (bioengineering OR cloning)

MIT's warning about implied AND is the usual Google failure: college students test anxiety becomes four separate words scattered across a record. Quoted phrases fix that. NOT is easy to overuse — PubMed will drop every record that mentions the excluded term, including reviews that mention it only to set it aside. Cochrane Chapter 4 says the same thing for systematic searches: avoid NOT unless you have checked what it deletes.

Phrases, truncation, and wildcards

Harvard's official truncation example (HOLLIS, 14 September 2026) is child* (child, children, childhood, childbirth). The official single-character wildcard is wom?n (woman, women). MIT's truncation page adds that the symbol itself changes by database (*, !, ?, or #) — read that database's help screen before you copy a HOLLIS string into PubMed.

Official PubMed wildcard rules

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. For example,"vaccin* schedul*" will find results that include "vaccine schedule", "vaccine scheduling", "vaccination schedules", and more variations of this phrase.

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

Wildcards turn off MeSH mapping

Wildcards turn off Automatic Term Mapping (ATM) and the process that includes the MeSH term and any specific terms indented under that term in the MeSH hierarchy.

Proximity searching in PubMed

When a quoted phrase is too tight and a bare AND is too loose, PubMed's official proximity syntax puts the terms in one field within N words of each other, in any order. It is available only in Title, Title/Abstract, and Affiliation. It does not accept wildcards, and it turns Automatic Term Mapping off for the quoted terms.

PubMed User Guide, Proximity searching

You can use proximity searching to search for multiple terms appearing in any order within a specified distance of one another in the [Title], [Title/Abstract], or [Affiliation] fields.

Official format

"search terms"[field:~N]

Official combine example

"hip pain"[Title:~4] AND stretching

Phrase not in the phrase index (official)

"cognitive impairment in multiple sclerosis"[tiab:~0]

Proximity searching is not compatible with wildcards (). If the double quoted terms in a proximity search include a wildcard (), the proximity operator will be ignored.

Keywords versus controlled vocabulary

Keywords are the words you thought of. Controlled vocabulary is the word the database assigned. ACRL lists both as required search languages. In PubMed the vocabulary is MeSH. The National Library of Medicine course Topic Searching in PubMed: Using the Medical Subject Headings (last reviewed 5 June 2024) is the official walkthrough. In ERIC it is the ERIC Thesaurus. Open one relevant record, copy its subject headings, and search those headings — that is pearl growing without a citation index.

National Library of Medicine, Topic Searching in PubMed (MeSH)

PubMed maps your search terms to Medical Subject Headings (MeSH), the vocabulary NLM uses to describe articles, to augment your search with additional terms.

Worked search: vitamin C and the common cold

Do not invent a trendy social-media topic and pretend a database returned it. Use the example PubMed itself publishes. On the PubMed User Guide Boolean page, the official illustration is a two-concept biomedical question.

Official PubMed example, typed two ways

What you can type

vitamin c common cold

How PubMed translates it

vitamin c AND common cold

If you do not want automatic phrase parsing (official)

air AND bladder AND fistula

Run that search at pubmed.ncbi.nlm.nih.gov. Open Advanced and read Search Details — that pane shows the MeSH mapping. Then use the filters PubMed actually ships: Publication Date, Article Type (Review, Randomized Controlled Trial, Meta-Analysis), and Additional filters for species or age. If you need a narrative review of the same question, add Review as an article type instead of adding the word review to the box and hoping. The User Guide's own advice when the set is wrong:

PubMed User Guide, too many or too few citations

Too many

Replace general search terms with more specific ones (e.g., low back pain instead of back pain).

Include additional terms in your query.

Use the sidebar filters to restrict results by publication date, full text availability, article type, and more.

Education version of the same habit: search ERIC, tick Peer reviewed only only when the assignment requires journal articles, and leave it off when you also need IES reports and conference papers. The May 2024 ERIC Selection Policy states that ERIC indexes both peer-reviewed and non-peer-reviewed material, and that the peer-review flag on a journal article is a journal-level designation. Humanities version: search the MLA International Bibliography for the article, then the catalog for the book it is about.

Citation chaining from one real paper

McMaster University Library's citation-chaining handout (Erin Stienstra, 9 July 2020) is the clearest short definition this page found. Cochrane Chapter 4's technical supplement (September 2024 PDF) uses the same two directions under the names backwards citation searching and forwards citation searching:

McMaster University Library, Citation Chaining

Citation chaining is a research strategy that helps you find sources that cite each other.

Exclusively using citation chaining to locate resources can be limiting and can keep you from discovering other valuable research.

Cochrane Handbook 6.5 technical supplement, citation searching (September 2024)

Forwards citation searching identifies studies which cite a source study and backwards citation searching identifies studies cited by the source study.

Since researchers may selectively cite studies with positive results, forwards citation searching should be used only as an adjunct to other search methods in Cochrane Reviews.

Backward chaining is the reference list of a paper you already trust. Forward chaining is newer work that cites it — Google Scholar's Cited by, or cited-reference search in Web of Science, Scopus, or OpenAlex. McMaster names Web of Science, Scopus, and Google Scholar for the forward direction. Do not treat a high cite count as quality; it only means later writers mentioned the item.

Real pearl (PLOS Medicine citation line)

Ioannidis JPA (2005), official PLOS citation

Ioannidis JPA (2005) Why Most Published Research Findings Are False. PLoS Med 2(8): e124. https://doi.org/10.1371/journal.pmed.0020124

Open the article. PLOS Medicine is open access. The DOI is 10.1371/journal.pmed.0020124. The article page (opened 17 September 2026) still prints the 25 August 2022 correction.
Backward. The essay's own reference list is the prior conversation (Ioannidis cites earlier work on refutation, bias, and molecular epidemiology). Open two or three of those, not all of them.
Forward. In Google Scholar, search the exact title in quotes, then open Cited by. In Web of Science, use cited-reference search on Ioannidis JPA, 2005, PLOS MED.
Stop and evaluate. The same essay is a worked CRAAP case on How to evaluate sources. The correction record is 10.1371/journal.pmed.1004085 (published 25 August 2022). Finding the paper is not the same as agreeing with the title.

Failure modes students actually hit

The peer-review filter is not a truth serum

A database's Peer reviewed limiter usually means "this journal is peer reviewed," not "this item was refereed." ERIC's May 2024 Selection Policy says the same thing in official language: the flag is set at the journal level and applied to every record from that source. Editorials, letters, and book reviews in those journals still appear. Confirm the article type on the PDF, then confirm the journal in Ulrichsweb if your library subscribes — Ulrichsweb is the periodical directory librarians use for that check. The process of review itself is How peer review works.

Too many results, too few, or the wrong field

Too many. Follow PubMed's own list: swap a general term for a specific one, add a concept with AND, or use the sidebar filters. Search Title/Abstract instead of All Fields when the noise is in methods mentions. A date limiter is a way to see current approaches — that is a filter, not a rule that older work is worthless.
Too few. OR in synonyms, drop a concept, or change databases. An empty ERIC set does not mean medicine is empty; try PubMed.
Off-topic. Open one good record and steal its subject headings. That is often faster than adding NOT terms you will regret.
Wrong century. History and theory need older books. Clinical protocols and software papers usually do not. Recency is a property of the claim, which is the Currency letter on How to evaluate sources.
Wildcard ate the mapping. A PubMed * turns off Automatic Term Mapping and the MeSH explosion. Use it after you have seen Search Details, not as the first keystroke.

You found the citation and not the file

A catalog hit is not full text. Use the campus link resolver, then interlibrary loan, then the author's repository copy. Emailing the corresponding author is still normal. The legal options — and the ones to refuse — are on How to access academic journals. Do not stall a paper for a week waiting on a PDF you could have requested on day one.

The articles are unreadable

Start with a review article or a subject encyclopedia, then the empirical papers. Read the abstract, the last paragraph of the introduction, and the conclusion before the methods. That order is a reading tactic, not a citation shortcut — you still cite the study you actually used. For a literature-review assignment, pair this search with How to write a literature review.

You are searching only the open web

Google and Scholar will surface some journal articles. They will not replace a licensed subject database for a comprehensive undergraduate literature review, and they will not tell you cleanly which hits are refereed. Use both. Prioritize the database for the systematic pass; use Scholar for chaining and for titles you already know. Cochrane's supplement is blunt: Scholar cannot replace bibliographic databases, and the viewable list stops at 1,000.

Ask a librarian before you invent a third search string

One of ACRL's dispositions for this frame is to seek guidance from experts, including librarians. A 20-minute consultation usually costs less time than a third evening of off-topic hits. Bring the assignment sheet, the search you already ran, and one article that is close. Subject librarians can name the database, the heading, and the interlibrary-loan button.

Use chat, email, or a booked consultation — not only the walk-up desk.
Ask for the subject guide for your department; most libraries publish one.
Your instructor can name the two or three authors the assignment expects you to have found.
Save records to Zotero, Mendeley, EndNote, or RefWorks as you go. The folder habit is How to organize your research.

Frequently Asked Questions

How many scholarly sources do I need?: The assignment sheet wins. Common undergraduate ranges are about 5–10 sources for a short paper and 10–20 for a longer one; graduate work and theses run much higher. Count usable, cited sources, not every PDF you downloaded. A literature-review assignment is asking for coverage, not a lucky seven.
Can I use Google instead of a library database?: You can start there to learn the vocabulary. For the set you will cite, use a library catalog or a subject database (PubMed, ERIC, PsycINFO, MLA International Bibliography) and then Google Scholar for Cited by. Cornell Library's search-strategy page states the difference: databases run keyword searches across peer-reviewed literature; Google runs natural-language search across web pages. See How to use Google Scholar for the operators.
How do I know an article is peer-reviewed?: Check the article type on the PDF (research article versus letter or editorial), then the journal's site, then Ulrichsweb if your library has it. A database Peer reviewed filter is a journal-level flag — ERIC's May 2024 Selection Policy says so in those words. Preprints on arXiv or SSRN are not peer-reviewed until a journal says they are.
Should I use only sources from the last five years?: Only if the claim expires. A date limiter is a way to see current approaches, not a ban on older work. Keep foundational books and methods papers; refresh clinical, technology, and statistics claims.
What if my library does not subscribe to the database I need?: Ask a librarian for an equivalent index, a consortial login, or interlibrary loan. PubMed and ERIC remain free on their official sites. Document delivery exists for a single article you have already identified.
Why did PubMed return different results than Google Scholar for the same words?: PubMed maps terms to MeSH and requires uppercase AND, OR, and NOT. Scholar searches a wider mix of literature and does not use MeSH. The official PubMed example vitamin c common cold is translated as vitamin c AND common cold — Scholar has no such translation pane. Scholar also caps the viewable result list at 1,000 and does not disclose how those 1,000 are selected.
What is citation chaining, and is it enough by itself?: It is following references backward and Cited by forward from one good paper. McMaster Library warns that using only chaining hides work that never cited your pearl. Cochrane's Chapter 4 technical supplement (September 2024) adds that forwards citation searching should be an adjunct, not the whole search, because researchers may cite positive studies selectively. Pair chaining with a designed database search.
Do I need MEDLINE, Embase, and CENTRAL for a class paper?: No. Cochrane MECIR C24 names that trio as the minimum for a Cochrane intervention review. For an undergraduate paper, pick the one or two tools that match the question (PubMed or CINAHL for clinical nursing; ERIC for education; MLA International Bibliography for literature) and chain from one good hit. Use the Cochrane rule only when your methods section claims a systematic search.
What does PubMed proximity searching do that AND does not?: AND finds the words anywhere in the record. Proximity finds them in one field within N words of each other, in any order. The official format is "search terms"[field:~N]. PubMed allows it only in Title, Title/Abstract, and Affiliation. It ignores the operator if you put a wildcard inside the quotes.
I found Ioannidis 2005. Can I stop there?: No. Use it as a pearl: open two items from its reference list and a handful of later papers that cite it. Then evaluate the argument. The official citation is Ioannidis JPA (2005) Why Most Published Research Findings Are False. PLoS Med 2(8): e124. https://doi.org/10.1371/journal.pmed.0020124. The 25 August 2022 correction is 10.1371/journal.pmed.1004085.

Cite the sources you actually opened

Once the set is on your desk, format the ones you used. Pair this workflow with the Scholar operators page if you are chaining, or with the database guide if you are still inside an interface.

Generate a citation Google Scholar operators