03 · What You Need to Know
Why Scopus and Web of Science Do Not Contain Exactly the Same Literature
Both Are Selective Multidisciplinary Citation Databases
Scopus and Web of Science are not universal repositories containing every scholarly publication. Both select content according to their own coverage policies and evaluation processes.
Scopus covers serial sources such as journals, book series, trade journals, and conference series, as well as non-serial books and conferences. Elsevier's Scopus Content Coverage Guide explains that journals constitute the largest portion of its content and that titles are selected according to Scopus content-selection policies.
Web of Science Core Collection is likewise selective. Clarivate currently describes it as covering journal, conference, and book content across more than 250 research areas. Its component citation indexes include the Science Citation Index Expanded, Social Sciences Citation Index, Arts & Humanities Citation Index, Emerging Sources Citation Index, Conference Proceedings Citation Index, and Book Citation Index.
The crucial point is that Elsevier and Clarivate operate separate selection systems. A source accepted by one is not automatically accepted by the other.
Think of Two Overlapping Circles, Not One Database Inside the Other
The relationship is easiest to understand as two overlapping sets.
Scopus-only coverage
Some sources or records may be covered by Scopus but not by the Web of Science collection you are searching.
Overlapping coverage
Many journals and publications are represented in both databases.
Web of Science-only coverage
Some sources or records may be covered by Web of Science but not by Scopus.
The practical consequence
Searching either database alone does not automatically reproduce the complete result set of the other.
The proportions of these areas are not fixed. They depend on what exactly you compare. A comparison of recent journal articles in one scientific specialty may show extensive overlap, while a comparison involving older literature, books, conference proceedings, regional publications, or another discipline may look different.
“Web of Science” Is Not a Single Undifferentiated Dataset
Comparisons become confusing because the term Web of Science can refer to the broader platform, while Web of Science Core Collection refers to a particular collection of citation indexes.
Within the Core Collection itself, coverage depends on which indexes are included. Clarivate identifies separate indexes for sciences, social sciences, arts and humanities, emerging sources, conference proceedings, books, and chemistry-related content.
Your institution's subscription depth can also affect what you can search. Clarivate explicitly notes that coverage available to a user depends partly on institutional subscription depth.
Watch Out
When someone says that a study compared "Scopus with Web of Science," check what Web of Science collection and timespan were actually searched. Comparing Scopus with one subset of Web of Science Core Collection is not necessarily equivalent to comparing it with every database available through the Web of Science platform.
Different Journal Selection Processes Produce Different Source Lists
Both databases evaluate sources, but they do so independently.
Clarivate states that journals entering Web of Science Core Collection are evaluated against a set of quality criteria, with additional impact criteria used for placement in its flagship citation indexes. Journals meeting the quality criteria may enter the Emerging Sources Citation Index, while journals meeting additional impact criteria may enter the Science Citation Index Expanded, Social Sciences Citation Index, or Arts & Humanities Citation Index.
Scopus uses its own content-selection process and coverage policy. Because the systems have separate criteria, evaluation procedures, and coverage decisions, their journal lists cannot be expected to match exactly.
Coverage is also dynamic. Sources can be added, re-evaluated, moved between collections, or discontinued. A statement that two databases overlapped by a particular percentage in a published study should therefore not be treated as an eternal property of either database.
A Journal Being in Both Databases Does Not Mean Every Record Will Match Perfectly
Source-level overlap and record-level equivalence are different questions.
Suppose a journal appears in both Scopus and Web of Science. That does not necessarily mean that every issue, document type, year, early-access item, conference-related paper, correction, editorial, or other record will appear identically in both systems.
Coverage dates may differ. Document classifications may differ. Records may enter the databases at different times. Metadata may also be structured differently.
Source overlap
The same journal, book series, or conference source is covered by both databases.
Record overlap
The same individual publication is retrievable in both databases for the period and content being examined.
A source-level comparison is therefore useful, but it does not by itself establish that two searches will retrieve identical publications.
Historical Coverage Can Change the Comparison Substantially
Publication year matters when comparing databases.
Web of Science Core Collection contains substantial historical citation coverage. Clarivate reports that the Science Citation Index Expanded and Social Sciences Citation Index extend back to 1900, while other indexes have different starting years. Its conference and book indexes also have their own coverage periods.
Scopus likewise contains extensive historical content, but its depth varies by source and content type. A comparison restricted to recent publications may therefore produce a different picture from one examining research published several decades ago.
If your research question depends heavily on historical scholarship, do not infer historical coverage from how well a database performs for current literature.
Conference and Book Coverage Can Also Differ
Journal articles dominate many database comparisons, but they are not the only scholarly outputs researchers may need.
Scopus indexes conference and book content in addition to journals. Web of Science Core Collection has dedicated Conference Proceedings Citation Index and Book Citation Index collections. The selection and coverage of these materials are not identical.
This can matter considerably in fields where conferences or books are central forms of scholarly communication. Computer science and some engineering fields, for example, may place greater weight on conference proceedings than many biomedical fields. Humanities and some social-science questions may depend substantially on books and book chapters.
If those materials matter to your research question, comparing journal counts alone tells you surprisingly little about which database offers the more appropriate coverage.
Different Coverage Produces Different Search Results
Suppose you enter what appears to be the same topic search into Scopus and Web of Science. Even before considering differences in search syntax, the databases are searching different underlying collections.
An article absent from Scopus cannot be retrieved by a Scopus search, however carefully designed. The same applies to an article absent from Web of Science.
Coverage is only part of the explanation, however. Search fields, indexing, document classifications, query syntax, stemming, database updates, and other platform behaviors can also affect retrieval. This is why the same search can produce different results in different databases even when their source coverage overlaps extensively.
Different Coverage Also Produces Different Citation Counts
A citation count is not simply a property permanently attached to an article. It is a count produced from the citing documents recognized within a particular citation system.
If Scopus indexes a citing document that Web of Science does not, that citation may contribute to the Scopus count without contributing to the corresponding Web of Science count. The reverse can also occur.
This is one reason the same paper, researcher, or journal may display different citation totals across Scopus, Web of Science, and Google Scholar.
The difference does not automatically mean that one database made an error. The systems may simply be counting citations from different document universes.
Which Database Is Larger Is Not the Most Useful Question
Database providers report large overall source and record counts, but comparing headline totals can be misleading. Counts may use different definitions, cover different document types, represent different time periods, or refer to different products and collections.
More importantly, researchers rarely need the database containing the largest number of records in the abstract. They need adequate coverage of the literature relevant to a particular question.
A database with fewer total records could provide stronger coverage of one research area. A larger database might offer broader international or interdisciplinary reach for another topic. A specialized database might outperform both for a particular discipline.
Instead of asking only which database is bigger, ask how well each database covers your actual field and publication types.
Coverage Studies Provide Useful Evidence, but Their Findings Are Contextual
Researchers have repeatedly compared Scopus and Web of Science and generally find substantial overlap alongside meaningful differences. Studies also show that the extent and consequences of those differences can vary by discipline and by the material being examined.
Such studies are valuable, but their percentages should be interpreted carefully. Coverage evolves, and a comparison conducted using one publication period, discipline, document type, or database configuration does not establish a permanent universal overlap rate.
For your own research, the most relevant question is not whether a published study once found a particular percentage of overlap. It is whether the databases adequately represent the literature required by your research question now.
