Collaboration Distance and Erdős Number Analysis
Also known as: Erdős Number Analysis, Co-Authorship Distance, Collaboration Geodesic Analysis, Scientific Small-World Analysis
Collaboration distance analysis measures how closely connected scientists are through chains of co-authorship. Two researchers who have written a paper together are at distance 1; if they share a co-author but never wrote together, distance 2; and so on. The most famous instance is the Erdős number, the collaboration distance to the prolific mathematician Paul Erdős, popularized by the Erdős Number Project and analyzed by Rodrigo de Castro and Jerrold Grossman. M. E. J. Newman's landmark 2001 PNAS study generalized this idea, constructing large co-authorship networks across physics, biomedicine, and computer science and showing that they are 'small worlds': despite millions of authors, typical shortest paths are short and local clustering is high. Collaboration distance analysis thus characterizes the connectivity and reach of scientific communities through the geometry of their co-authorship graphs.
Key highlights
- Reduces complex collaboration structure to interpretable distances and reveals the small-world character of scientific communities.
- Identifies connecting hubs and the giant component that integrate a field, and quantifies how short the chains between researchers are.
- Built on standard, well-understood graph algorithms (shortest paths, clustering coefficient) with clear definitions.
- Enables comparison of cohesion and integration across fields and over time using consistent network metrics.
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 collaboration distance and Erdős number analysis when you want to understand the connectivity, reach, and structure of a research community from its co-authorship records: measuring how integrated a field is, identifying connecting hubs and brokers, comparing the cohesion of different disciplines, or studying how the giant connected component and typical distances evolve as a field grows. It is appropriate when you can construct a reliable co-authorship graph with good author disambiguation. It is less suitable when authorship data are noisy or names cannot be disambiguated (which corrupts the graph), when co-authorship is a poor proxy for the collaboration you care about (informal or non-publishing collaboration is invisible), or when you need to attribute credit or productivity rather than describe connectivity — distance metrics describe topology, not contribution.
Strengths & limitations
- Reduces complex collaboration structure to interpretable distances and reveals the small-world character of scientific communities.
- Identifies connecting hubs and the giant component that integrate a field, and quantifies how short the chains between researchers are.
- Built on standard, well-understood graph algorithms (shortest paths, clustering coefficient) with clear definitions.
- Enables comparison of cohesion and integration across fields and over time using consistent network metrics.
- Highly sensitive to author-name disambiguation; merging distinct authors or splitting one author badly distorts distances and clustering.
- Co-authorship is an imperfect proxy for collaboration, missing informal, advisory, or non-publishing interactions.
- Database coverage and field boundaries shape the graph, so distances depend on which papers and venues are included.
- Aggregate metrics describe topology but say nothing about the quality, intensity, or credit of any collaboration.
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
What exactly is an Erdős number?
It is the collaboration distance from a person to the mathematician Paul Erdős in the co-authorship graph. Erdős has number 0; anyone who co-authored a paper with him has number 1; anyone who co-authored with one of those people (but not Erdős himself) has number 2; and so on. Someone with no co-authorship chain to Erdős has an undefined or infinite number. It is the single-source shortest-path distance to a fixed anchor, and the same computation works for any chosen reference author.
What does it mean that scientific collaboration networks are 'small worlds'?
Newman showed these networks combine two properties: the average shortest path between authors is small — growing only logarithmically with the number of authors, so most researchers are separated by just a few co-authorship steps — while the clustering coefficient is high, meaning your collaborators tend to collaborate with each other. Random graphs have short paths but low clustering; regular lattices have high clustering but long paths. Collaboration networks have both, which is the signature of a small world and shapes how ideas spread.
Why is author-name disambiguation so important here?
Every distance and metric is computed on the co-authorship graph, so errors in identifying who is who propagate directly into the results. If two different scientists share a name and are merged into one node, the graph gains false links and artificially short paths; if one scientist's name variants are split into several nodes, real connections vanish and distances inflate. Newman and subsequent work treat disambiguation as a central methodological concern, because an unclean author list can make a network look more or less of a small world than it really is.
Sources
- 1.Newman, M. E. J. (2001). The structure of scientific collaboration networks. Proceedings of the National Academy of Sciences, 98(2), 404-409.
- 2.De Castro, R., & Grossman, J. W. (1999). Famous trails to Paul Erdős. The Mathematical Intelligencer, 21(3), 51-63.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 23). Collaboration Distance and Erdős Number Analysis. ScholarGate. https://scholargate.app/bibliometrics/collaboration-distance-analysis