Statistics Glossary → Quadrat count test
Quadrat count testconcept
A chi-square test of whether points are distributed according to complete spatial randomness (CSR), carried out by dividing the study area into equal cells (quadrats) and comparing the number of points in each.
If points are scattered over an area completely at random and independently (a homogeneous Poisson process), the expected number of points in each cell is proportional to its area. The quadrat count test compares the observed cell counts with these expected counts using a chi-square statistic. If the variance of the counts is clearly larger than their mean, the points are clustered (some cells crowded, others empty); if it is clearly smaller, the points are regular, keeping their distance from one another.
The result depends strongly on cell size: with very small cells, expected counts are low and the chi-square approximation weakens; very large cells hide small-scale clustering. The test gives information at a single scale only and does not say where the clustering is; distance-based methods such as Ripley's K/L functions give more detailed information across several scales. And if the area itself is heterogeneous (some parts are a lake, say), clustering may stem from environmental conditions rather than the process itself.
An everyday example
The positions of 100 trees in a forest are mapped and the area is divided into 5 x 5 = 25 cells, so each cell is expected to hold 4 trees. If some cells contain 12-15 trees and many contain none, the chi-square test is significant, showing that the trees are clustered (for example because seeds fall near the parent tree). If every cell holds 3-5 trees, the pattern is random or regular.
In MerQur
In Point Pattern Analysis, the 'Quadrat counts (columns x rows; default 5 x 5)' option divides the window into cells and runs a chi-square goodness-of-fit test on the counts; with a polygon window, clipped cell areas are used.
Analyses where this concept appears
Point Pattern Analysis (Ripley K/L, G/F/J, CSR)