Network Text Analysis
Also known as: Text network analysis, Centering resonance analysis, Concept network analysis, Ağ Tabanlı Metin Analizi
Network text analysis represents the content of text not as counts of words or topics but as a network of concepts linked by their relationships, then applies social-network methods to reveal which ideas are central and how they connect. Centering resonance analysis (CRA), introduced by Corman and colleagues in 2002, is a leading variant that builds concept networks from the noun phrases that structure discourse.
Key highlights
- Captures relational meaning — how concepts connect — that word counts and topic proportions ignore.
- Brings the full social-network toolkit (centrality, community detection) to text, identifying pivotal concepts and bridges.
- Enables principled comparison of conceptual structure across authors, outlets, and time.
- Centering resonance analysis grounds concept selection in linguistic theory rather than arbitrary keyword lists.
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 text analysis when the relational structure of meaning matters — not just which concepts appear but how they are connected — and you want to compare that structure across texts, sources, or time. It suits studies of framing, shared meaning and influence, and the conceptual organization of discourse, and it pairs well with semantic-network and network-agenda-setting work. It assumes concepts and their links can be extracted reliably and that network structure is theoretically meaningful for your question. It is less appropriate when simple frequency or topic measures suffice, when texts are too short to form informative networks, or when the extraction of concepts and links is so noisy that the resulting graph is unreliable — in which case validation against human reading is essential.
Strengths & limitations
- Captures relational meaning — how concepts connect — that word counts and topic proportions ignore.
- Brings the full social-network toolkit (centrality, community detection) to text, identifying pivotal concepts and bridges.
- Enables principled comparison of conceptual structure across authors, outlets, and time.
- Centering resonance analysis grounds concept selection in linguistic theory rather than arbitrary keyword lists.
- Network quality depends on noisy concept-extraction and link-definition choices that strongly shape results.
- Defining what counts as a 'link' (window size, grammatical relation) is consequential and not standardized.
- Interpreting network metrics substantively requires care; high centrality is not automatically meaningful.
- Large concept networks can be hard to visualize and interpret without aggressive pruning that may distort structure.
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 does network text analysis differ from topic modeling?
Topic modeling groups words into latent themes and represents each document as a mixture of those themes; it answers what a corpus is about. Network text analysis instead represents concepts as nodes and their relationships as edges, then studies the structure — which concepts are central and how they connect. It answers how meaning is organized relationally. The two are complementary: topics describe thematic content, networks describe the connective structure among concepts.
What is centering resonance analysis?
Centering resonance analysis (CRA) is a network text method that builds concept networks from the noun phrases that, according to linguistic centering theory, structure and connect discourse. By focusing on these influential words rather than all terms, CRA produces networks meant to reflect the language people use to coordinate and influence one another. It then uses network centrality to identify the concepts that resonate across a text or corpus, enabling comparison of discursive structure across speakers and over time.
How are links between concepts defined?
It depends on the variant. Simple approaches link concepts that co-occur within a fixed window or the same sentence. More linguistically grounded approaches, like CRA, link concepts based on grammatical structure within and between sentences. The choice matters a great deal: wider windows create denser, less specific networks, while strict grammatical links create sparser, more precise ones. Because there is no single standard, the link definition should be justified and its sensitivity examined.
Sources
- 1.Corman, S. R., Kuhn, T., McPhee, R. D., & Dooley, K. J. (2002). Studying complex discursive systems: Centering resonance analysis of communication. Human Communication Research, 28(2), 157–206.
- 2.Wasserman, S., & Faust, K. (1994). Social Network Analysis: Methods and Applications. Cambridge: Cambridge University Press.ISBN 9780521387071
You have read it. What now?
Cite this page
ScholarGate. (2026, June 22). Network Text Analysis. ScholarGate. https://scholargate.app/communication/network-text-analysis