Academic databases compared: Web of Science, Scopus & Google Scholar

Serious work on a bachelor’s or master’s thesis rarely starts in Google. The three big research databases — Web of Science, Scopus and Google Scholar — shape which sources you find at all, how clean your literature review becomes, and how much junk ends up in your Zotero folder. This comparison shows you when each one is genuinely the better tool, and how to combine the three sensibly.

The big three at a glance

Web of Science and Scopus are selective, curated databases — a journal is admitted only if it meets defined quality criteria. Google Scholar is a search engine across practically everything on the web that looks like academic work: journals, preprints, repositories, conference proceedings, dissertations, patents, sometimes even blog posts. Those three models determine where you find what.

Web of Science — the strict library

The Web of Science Core Collection is the oldest of the three, running since the 1960s, and is the source of the Journal Impact Factor (JIF) and the Journal Citation Reports. If your department says “peer-reviewed only, indexed in Web of Science”, this is why.

Strengths

• Very strict inclusion criteria. A journal must publish regularly, be peer-reviewed and show an international author profile.
• First-class citation analysis: cited reference search, times cited, Hirsch index for authors.
• Stable, long backfiles — well suited for historical bibliometric work.
• Unique author IDs (ResearcherID, now linked with ORCID).

Weaknesses

• Anglophone and STEM-heavy. Humanities and social sciences, especially non-English ones, are under-represented.
• “Strict” also means: little grey literature, no preprints, almost no conference contributions outside the major series.
• Licence required — without a university login you only see titles and abstracts.

When to use it: natural sciences, medicine, engineering, economics with a quantitative focus. Whenever you need a clean, citable hit space and citation counts matter for your argument.

Scopus — the broadest index

Scopus is Elsevier’s answer to Web of Science, live since 2004. It indexes more journals than Web of Science (over 27,000 vs. ~22,000), is broader in scope, and its inclusion criteria are slightly more permissive — though the quality bar is still well above Google Scholar.

Strengths

• Largest curated coverage, better than Web of Science in social sciences and humanities.
• Good filters (subject area, document type, language, country).
• SJR (SCImago Journal Rank) and CiteScore as quality metrics — an alternative to JIF.
• Author profiles with Scopus ID, affiliation history and h-index.
• Solid API for systematic reviews.

Weaknesses

• Also licence-bound.
• Shallower backfiles than Web of Science — pre-1996 coverage is weaker.
• Small or new German-language journals can show gaps.

When to use it: when you want to cover a broad topic systematically, especially in social sciences, public health, education or nursing. Also first choice for systematic literature reviews.

Google Scholar — the free full-text machine

Google Scholar has been around since 2004, is free, and indexes anything Google’s crawler recognises as scholarly. That’s a huge pool — but an unfiltered one.

Strengths

• Maximum reach: you find things here that never appear in Scopus or Web of Science — grey literature, preprints (arXiv, SSRN, bioRxiv), working papers, dissertations, reports from NGOs and government bodies.
• Full-text search, not just title and abstract — sometimes Scholar finds a hit because your term appears on page 47 of the PDF.
• “Cited by” works similarly to Web of Science, with less depth but a wider base.
• No login required, accessible anywhere.

Weaknesses

• No quality control. Predatory journals, copy-pasted student papers, old lecture PDFs — they all show up in the result space.
• Ranking algorithm is opaque and not reproducible. A paper sits at position 3 today and 27 tomorrow.
• Search syntax is much weaker than Web of Science or Scopus. Boolean operators work, but filters like “reviews only”, “since 2020 only”, “German only” are fuzzy.
• Citation counts are displayed, but they double-count or estimate — not citable in a methods section.

When to use it: for quick exploration when you don’t yet know what a field calls itself. For grey literature (reports, working papers, dissertations). For full-text hits when you’re chasing a specific quote whose source you’ve lost. For free preprint versions (on arXiv) of an article whose publisher version is behind a paywall.

Search strategy: combining the three

In practice you don’t use one but all three in sequence. A sensible order for a bachelor’s thesis:

1. Scope in Google Scholar. Type your topic in plain words, scan the first two pages, jot down recurring terms, author names, key journals.
2. Sharpen the vocabulary. In Scopus or Web of Science, search with those subject keywords (Scopus calls them “Author Keywords” and “Indexed Keywords”). The hits are higher quality.
3. Snowball. From the five best hits: read the references (“Cited References” in Web of Science, “References” in Scopus). Who cites whom?
4. Cited-by chain. The other way around: who cited these key articles later? That surfaces more recent follow-up work.
5. Fill gaps with Scholar. If dissertations, reports or grey literature are still missing at the end, go back to Scholar.

This choreography saves you from the classic mistake: digging in Google Scholar for days, collecting fifteen mediocre hits and still not having an overview.

University access: Shibboleth, VPN, EZproxy

Web of Science and Scopus cost money — a lot of money. If your university has a licence (which any research-active institution does), use it properly:

• On campus with the university Wi-Fi you get access directly.
• From home via VPN (e.g. Cisco AnyConnect) or Shibboleth / SWITCH edu-ID (in Switzerland). On first login you pick your home institution.
• EZproxy is an older variant: a URL prefix is placed in front of the database address — your library usually provides a bookmarklet generator for it.

With proper access you see not only abstracts but full texts via your library’s “OpenURL” linker. If that fails: order interlibrary loan, usually arrives within three days.

Discipline-specific databases

The big three are generalist. Many disciplines have stronger specialist databases:

• Medicine, nursing: PubMed / MEDLINE (free), CINAHL, Cochrane Library for systematic reviews.
• Psychology: PsycINFO (APA), PSYNDEX (German-language).
• Economics: EconLit, Business Source Premier, RePEc.
• Law: Westlaw, LexisNexis, HeinOnline, Beck-online (DE/CH).
• Engineering / computer science: IEEE Xplore, ACM Digital Library, INSPEC.
• Humanities: JSTOR, MLA International Bibliography, Project Muse.
• Education: ERIC, FIS Bildung.
• Sciences: SciFinder (chemistry), Web of Science Core Collection.

In the methodology section of a systematic literature review, an honest list of the databases you searched belongs in. “I only used Google Scholar” is methodologically attackable.

Quality filters you should always set

Whichever database you land in — before you sift through results, apply these filters:

• Peer-reviewed only (Scopus: “Source type” → Journals; Web of Science automatically).
• Language (often English OR German for German-language theses).
• Time range: usually the last 10–15 years, plus the field’s classics before that.
• Document type: Article + Review, no editorials, no letters (unless that’s your field).
• Open Access, if you need full-text access and have no licence.

In Google Scholar these filters are largely absent — another reason to use Scholar only for scoping.

From the hit list into your Zotero

All three databases support direct export to reference managers. In Scopus click “Export → RIS” or “Add to Mendeley/Zotero”, in Web of Science “Export → Plain Text File / EndNote / RIS”, in Google Scholar the little quote-mark icon under each hit → “BibTeX” or “RIS”.

A browser extension (Zotero Connector, EndNote Capture) recognises most hit lists automatically and imports with one click. Maintain two fields on import: a topic tag and a stage tag (“read”, “full text”, “in progress”). That saves you hunting for sources later.

Hit quality is only half the battle

A cleanly researched source is just the foundation. What ends up in the finished text — the claim you anchor to a specific passage of a source — has to actually be there. This is where the most common errors in theses occur: the source is solid, the citation style correct, but the attributed claim isn’t actually in the original.

That’s where Acurio comes in: you upload your thesis and the matching PDFs from your Zotero, and Acurio checks claim by claim whether the statement actually appears in the source — including the page reference. Web of Science and Scopus tell you what’s citable. Acurio tells you whether your citations are correct.

Existence first: Before every citation is read against the full text, a prefilled example shows how fabricated and mismatched entries look. Check a sample citation →

Short checklist before your first research session

• Which databases has my university licensed? (Library website or info desk.)
• Have I set up VPN / Shibboleth?
• Which specialist database is the right one for my topic?
• Have I installed Zotero Connector in my browser?
• Am I logging my search strings in a spreadsheet (for the methods section of the review)?

With that you begin the next research session without getting lost on the first Google Scholar page — and with a clear picture of which database serves you best at which point.