What makes Web of Science distinctive
Web of Science is operated by Clarivate, and its Core Collection is the part most bibliometric studies rely on. The Core Collection is made up of several indexes, including the Science Citation Index Expanded (SCIE), the Social Sciences Citation Index (SSCI), the Arts and Humanities Citation Index (AHCI), the Emerging Sources Citation Index (ESCI) and the conference proceedings indexes. Its strengths are a selective journal evaluation process and structured citation data that extends back much further than most alternatives.
One detail is frequently overlooked. Institutions subscribe to different combinations of indexes and different year ranges, so two researchers at two universities running the identical query can see different result counts. A reproducible methods section must therefore list the indexes and years that were actually available, and we confirm this with you before any data is collected. If your institution lacks an index that matters for your topic, for instance the social sciences index for a policy question, we discuss whether Scopus or an open source should fill the gap before the search strategy is fixed.
Writing a Web of Science query
Web of Science uses its own field tags. TS searches title, abstract, author keywords and Keywords Plus together; TI, AB and AK search those fields individually; PY restricts publication years and DT restricts document types. An illustrative query for a hypothetical topic:
TS=("digital literacy" OR "digital competenc*")
AND TS=(teacher* OR "pre-service")
AND PY=(2010-2025)
AND DT=(Article OR Review)
Because TS includes Keywords Plus, it sometimes retrieves papers that are only loosely related. We compare a TS search against a combination of TI, AB and AK, read a random sample from each, and choose the version that best balances recall and precision. All tested versions are logged with their hit counts.
Exporting and merging with Scopus
Citation-based analyses need full records including cited references, usually exported as Plain text or Tab delimited files. The maximum number of records per export depends on the content selected and can change, so we export in batches and verify the combined total against the original hit count.
Merging with Scopus is often worthwhile because it reduces single-source coverage bias, but it introduces technical complications that should be handled openly:
| Merging issue | How we handle it |
|---|---|
| Different field structures | Both sources are mapped to a common structure in Bibliometrix before merging |
| Records present in both databases | Deduplicated by DOI first, then by normalised title for records without a DOI |
| Different cited reference formats | Where harmonisation would lose information, co-citation is run on each source separately |
| Indexer keywords that differ in nature | Author keywords are preferred for thematic networks on merged data |
| Diverging citation counts | Each table states which database its citation counts come from |
How the project runs
Indexes and access
We confirm which indexes and year ranges your institution subscribes to and use that legitimate access wherever the licence requires it.
Query refinement
Field tags are tested and every version is logged with its hit count.
Batch export
Full records with cited references are exported in batches, combined and reconciled against the total.
Cleaning and merging
Authors, affiliations and keywords are harmonised with a thesaurus; Scopus records are merged and a PRISMA 2020 diagram is produced in the Advanced package.
Analysis and delivery
Bibliometrix and VOSviewer outputs, eight maps, a thematic map, three-field plot, trend topics and an English methods description.
Strengths for citation and longitudinal work
Structured, long-running citation data makes Web of Science especially suitable for co-citation, bibliographic coupling and historical citation analysis. The Web of Science Categories also provide a consistent way of describing the disciplinary profile of a literature, which is useful when a topic spans several fields.
Two caveats belong in most papers. The cited references field typically records only the first author of each cited work, so author co-citation reflects first authors. And Keywords Plus is generated automatically from the titles of cited references rather than chosen by the authors, so it can be broader than a paper's actual content. We usually build thematic networks from author keywords and use Keywords Plus as a cross-check.
Describing the data source in your paper
Name the Web of Science Core Collection along with the specific indexes and years, the exact query, the retrieval date, filters and record counts at each stage. If Scopus was merged in, describe the deduplication procedure and the number of overlapping records. Our technical report supplies every element. The argument for why these sources suit your question, and the discussion of their limitations, is yours to write, though we are happy to explain any technical step during the revision window.
A hypothetical record flow for a merged project might read: 1,640 records from Web of Science and 2,010 from Scopus were combined, 1,122 duplicates were removed, and 2,528 unique documents entered the analysis. Reporting the overlap explicitly is useful in itself, because it tells readers how much each database added beyond the other, and reviewers often ask for exactly that figure when two sources are used.
Data terms and integrity
Compare with our Scopus bibliometric analysis service, see pricing or place an order.