Nearest Neighbour Index
Also known as: Clark-Evans Index, Nearest Neighbour Analysis, NNI
The nearest neighbour index, introduced by Clark and Evans in 1954, is a simple summary statistic that quantifies whether a set of points is clustered, randomly scattered, or evenly dispersed across an area. It compares the average distance from each point to its closest neighbour with the average distance that would be expected if the same number of points were placed completely at random. The ratio of observed to expected distance, together with a significance test, gives a single interpretable number that has become a staple of point-pattern analysis in geography and ecology.
Key highlights
- Produces one easily interpreted number anchored at 1 for randomness, with intuitive direction.
- Comes with an analytic expected value and significance test, requiring no simulation in the basic case.
- Computationally trivial and applicable to any set of point coordinates.
- Long-established and widely understood across geography, ecology, and the spatial sciences.
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 the nearest neighbour index for a quick, interpretable first characterization of a point pattern — settlements, retail outlets, trees, nests, crime locations — when you want a single number summarizing clustering versus dispersion and a test against randomness. It suits exploratory analysis and comparison across patterns of similar extent. It is less appropriate when the pattern shows structure at multiple scales (where Ripley's K or pair-correlation functions are better), when the study-area boundary is ill-defined or strongly affects the result through edge effects, or when the underlying intensity is markedly non-uniform, since the index assumes a single global density.
Strengths & limitations
- Produces one easily interpreted number anchored at 1 for randomness, with intuitive direction.
- Comes with an analytic expected value and significance test, requiring no simulation in the basic case.
- Computationally trivial and applicable to any set of point coordinates.
- Long-established and widely understood across geography, ecology, and the spatial sciences.
- Considers only the single nearest neighbour, so it summarizes one scale and can miss multi-scale structure.
- Highly sensitive to the definition and area of the study region and to edge effects near the boundary.
- Assumes a homogeneous (constant-intensity) process, which is often violated by real geographic patterns.
- Two very different patterns can share the same index value, so it is a coarse descriptor.
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 do the values of the nearest neighbour index mean?
The index R is the ratio of the observed mean nearest-neighbour distance to the distance expected under complete spatial randomness. R less than 1 means points are closer together than random, indicating clustering; R equal to 1 means the pattern is indistinguishable from random; and R greater than 1 means points are more evenly spaced than random, indicating dispersion. The theoretical maximum is about 2.149, achieved by a perfectly regular hexagonal arrangement.
Why are edge effects a problem for the index?
Points near the boundary of the study area may have their true nearest neighbour lying outside the studied region, so the measured nearest-neighbour distance is artificially inflated. This biases the index toward apparent dispersion. Edge-correction methods — buffering, toroidal wrapping, or analytical corrections — reduce this bias, and because the expected distance depends on the area A, defining the study region carefully is essential to a meaningful result.
How does the nearest neighbour index compare with Ripley's K function?
The nearest neighbour index summarizes the pattern at a single scale — the distance to the one closest neighbour — and gives a single number. Ripley's K function instead counts neighbours within a continuum of distances, revealing whether clustering or dispersion occurs at different scales. K is more informative for complex patterns, while the Clark-Evans index is simpler and quicker; the two are complementary, and analysts often report both.
Sources
- 1.Clark, P. J., & Evans, F. C. (1954). Distance to nearest neighbor as a measure of spatial relationships in populations. Ecology, 35(4), 445–453.
You have read it. What now?
Cite this page
ScholarGate. (2026, June 22). Nearest Neighbour Index. ScholarGate. https://scholargate.app/human-geography/nearest-neighbour-index