Network-based Co-citation Analysis
Also known as: co-citation network analysis, bibliometric network co-citation, co-citation mapping, CCA network approach
Network-based co-citation analysis is a bibliometric technique that measures how often pairs of documents are cited together by later works, then models those relationships as a weighted network. Nodes represent documents (or authors or journals), edges represent co-citation frequency, and network algorithms identify clusters of intellectually related literature. It is widely used in systematic and scoping reviews to map the intellectual structure of a research field.
Key highlights
- Objectively maps the intellectual structure of a field from citation behavior rather than the researcher's subjective impression.
- Scales to very large corpora — thousands of publications — that would be impractical to review manually.
- Identifies foundational landmark documents and thematic clusters that anchor sub-fields.
- Temporal overlay reveals how research fronts emerge, merge, or decline over time.
- Reproducible: the same database query and software settings yield the same network.
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Use network-based co-citation analysis when you want to map the intellectual structure of a field — identifying foundational works, thematic clusters, and research fronts — based on a corpus of at least several hundred publications. It is well-suited to the bibliometric component of systematic, scoping, or integrative reviews in STEM, social sciences, and management. Prefer it over simple co-citation tabulation when the corpus exceeds roughly 500 publications, as the network view makes patterns visible that a table cannot convey. Do not use it as a substitute for reading and qualitatively assessing the content of key papers; it reveals structural relationships, not the quality or validity of findings. Avoid it when citation data are sparse (corpora under about 100 documents) or when the field's literature is mainly in books or grey literature not indexed in major databases.
Strengths & limitations
- Objectively maps the intellectual structure of a field from citation behavior rather than the researcher's subjective impression.
- Scales to very large corpora — thousands of publications — that would be impractical to review manually.
- Identifies foundational landmark documents and thematic clusters that anchor sub-fields.
- Temporal overlay reveals how research fronts emerge, merge, or decline over time.
- Reproducible: the same database query and software settings yield the same network.
- Co-citation frequency reflects how often documents are cited together, not whether the citing author's interpretation was accurate; spurious or critical citations inflate co-citation counts equally.
- Coverage depends entirely on the citation database used; conference papers, books, and grey literature are poorly indexed.
- Threshold and normalization choices can substantially alter cluster structure; results are sensitive to these analytic decisions.
- Highly cited older documents dominate the network regardless of their current relevance to the field.
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
How is co-citation analysis different from bibliographic coupling?
Co-citation analysis links documents that are cited together by later works — the relationship is retrospective, defined by the citing community. Bibliographic coupling links documents that share references — the relationship is prospective, defined at the time of publication. Co-citation networks tend to reflect established intellectual communities; bibliographic coupling tends to reflect current research fronts.
What is the minimum corpus size for a meaningful co-citation network?
A widely cited rule of thumb is at least 100–200 publications with enough reference overlap to produce a connected network; in practice, robust cluster structures typically emerge with 500 or more publications. With very small corpora, most pairs will have co-citation counts of 0 or 1, making the network sparse and uninformative.
Which software should I use?
VOSviewer is the most widely used tool for co-citation network visualization and is free; it handles author, document, and journal co-citation with built-in normalization. CiteSpace offers time-slice overlays and burst detection. The R package bibliometrix provides scripted, reproducible workflows. Gephi is more flexible but requires manual data preparation.
Should I report the network alongside my qualitative synthesis?
Yes. Best practice in mixed bibliometric-systematic reviews is to present the co-citation map as a structural overview — often in a figure with cluster labels — then read and qualitatively synthesize the content of the key documents within each cluster. The network guides what to read; the reading produces the substantive review.
Can co-citation analysis replace a full systematic review?
No. Co-citation analysis is a mapping and orientation tool, not a synthesis method. It does not assess study quality, extract findings, or evaluate evidence. It is best used as a complementary step within a larger review process to structure the literature before detailed reading.
Sources
- 1.Small, H. (1973). Co-citation in the scientific literature: A new measure of the relationship between two documents. Journal of the American Society for Information Science, 24(4), 265–269.
- 2.Chen, C. (2006). CiteSpace II: Detecting and visualizing emerging trends and transient patterns in scientific literature. Journal of the American Society for Information Science and Technology, 57(3), 359–377.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 3). Network-based Co-citation Analysis. ScholarGate. https://scholargate.app/scientometrics/network-based-co-citation-analysis