Assortative Mixing

Version

v4 · 2026-09-28 · History

Prime #

1480

Domain group

Natural Sciences

Origin domain

Physics

Subdomain

Mixing Patterns → Physics

Also from

Systems Thinking & Cybernetics, Sociology & Anthropology, Biology & Ecology, Computer Science & Software Engineering

Aliases

Assortativity, Assortative matching

Related primes

Network, Correlation, Similarity Measure, Bias, Modularity

Core Idea

Assortative mixing is the network pattern in which adjacent nodes are more similar on a declared attribute than an appropriate edge-end baseline predicts. Disassortative mixing is the opposite pattern: edges connect unlike nodes more often than expected. Neutral mixing means that, conditional on the baseline and admissible edges, the attributes observed at the two ends of an edge have no residual association.

The abstraction does not require a human preference. A social friendship network can be assortative by age; a technological network can be disassortative by degree; a protein-interaction network can exhibit degree correlations; and an information network can be assortative by political label.

How would you explain it like I'm…

Alike Links to Alike

Watch which kids hold hands at recess. If the kids holding hands usually wear the same color shirt, more often than if partners were picked by pulling names out of a hat, that's assortative mixing. It only tells you who ends up connected to whom; it doesn't tell you why.

Like Links With Like

In a network, the dots are things like people or computers, and the lines are the connections between them. Assortative mixing is when connected dots are more alike than you'd expect by chance, for example friends tending to be the same age. The "by chance" part matters: if almost everyone in a group is ten years old, lots of ten-year-old pairs would happen anyway, so you compare with what random pairing would give. The opposite, disassortative mixing, is when connections join unlike things more than chance, like a few big hubs linking mostly to many small places. Finding the pattern doesn't explain its cause; it could be choice, who lives nearby, or just how the network was built.

Similarity Across Network Links

Assortative mixing is a network pattern in which nodes joined by an edge are more similar on some attribute than a baseline predicts. The baseline is key: take how often each attribute value shows up at the ends of edges, and ask what pairings you'd see if the two ends were matched independently. If like-with-like pairs appear more often than that, the network is assortative; if unlike pairs appear more often, it's disassortative; if neither, it's neutral. An important case is degree assortativity, where "degree" means how many connections a node has: in many social networks well-connected people link to other well-connected people, while in many technological networks hubs connect mostly to low-degree nodes. The pattern is measured with a coefficient r, where 0 means baseline and positive values mean assortative. But r is a measurement, not an explanation: choice, geography, group sizes, or how the data were sampled can all produce it.

 

Assortative mixing is the pattern in which adjacent nodes share a declared attribute more than an appropriate edge-end baseline predicts; disassortative mixing is the reverse, and neutral mixing means no residual association once the baseline is accounted for. For a categorical attribute, let e_ij be the fraction of edges joining a type-i end to a type-j end, with marginals a_i = Σ_j e_ij and b_j = Σ_i e_ij. Neutral mixing corresponds to e_ij = a_i b_j, and Newman's coefficient r = (Σ_i e_ii − Σ_i a_i b_i) / (1 − Σ_i a_i b_i) equals 1 for perfect like-with-like mixing and 0 at baseline; negative values indicate disassortativity, though the attainable minimum depends on the margins and need not reach −1. For scalar attributes, r is the Pearson correlation of the values at the two ends of an edge. Degree assortativity uses (remaining, or excess) degree as the attribute: positive values mean hubs link to hubs, negative values produce hub-and-periphery structure. The pattern appears in social, technological, biological, and information networks and doesn't require any agent to have a preference. The coefficient describes edge-end association only; homophily, opportunity, geography, group sizes, degree constraints, and edge-sampling rules can each generate it.

Tensions in Practice

• Same members and degrees, different partners

Broad Use

The abstraction travels literally across networks whose nodes and edges have unrelated meanings.

Social networks. Friendship, marriage, advice, work, support, and communication ties often connect people similar in age, education, religion, occupation, ethnicity, attitudes, or other attributes. McPherson, Smith-Lovin, and Cook synthesize this as homophily and emphasize both its breadth and the ecological contexts that shape who can meet. Network assortativity supplies a general measurement interface for these patterns without making all of them the result of deliberate choice.

Clarity

The fastest diagnostic is to compare three quantities: raw same-type edge share, expected same-type share under the endpoint marginals, and their normalized difference. Suppose an undirected network has two equally represented endpoint types, A and B. Of eight edges, three are A–A, three are B–B, and two are A–B. In the symmetric mixing matrix, (e_{AA}=⅜), (e_{BB}=⅜), and the cross-type edge contributes (e_{AB}=e_{BA}=⅛).

Manages Complexity

A network with millions of edges and a rich attribute table contains an enormous set of pairwise observations. The mixing matrix compresses them into a joint endpoint distribution; the coefficient compresses that distribution further into a signed scalar. This reduction makes networks with different sizes and category frequencies comparable under an explicit baseline.

Abstract Reasoning

Assortative mixing licenses several disciplined inferences.

Baseline inference. If (e_{ij}=a_i b_j) for all relevant type pairs, edge endpoints are neutral with respect to that attribute under the stated sampling model. A diagonal count above its marginal expectation supports assortativity; raw sameness alone does not.

Knowledge Transfer

Transfer begins by replacing the substrate while preserving the roles. In a friendship network, nodes are people, edges are friendships, and the attribute might be age. In the Internet, nodes are autonomous systems, edges are peering relations, and the attribute might be degree. In a protein network, nodes are proteins, edges are interactions, and the attribute might be functional class. In a citation network, nodes are documents, directed edges are citations, and the attribute might be field. Each case supplies the same network, endpoint attribute, joint mixing distribution, marginals, null baseline, and signed comparison.

Relationships to Other Abstractions

Current abstraction Assortative Mixing Prime

Parents (1) — outgoing typed relations

• Assortative Mixing presupposes Network Prime

  The accepted reference-grade review places Assortative Mixing under Network because the child instantiates or depends on the parent's broader structure while retaining its own constitutive identity.

Hierarchy path (1) — routes to 1 parentless root

• Assortative Mixing → presupposes Network