Geary’s contiguity ratio (Geary’s c)

Geary's contiguity ratio (i.e., Geary's c) is used to measure spatial autocorrelation in data with discrete spatial support. Calculation of Geary's c depends on the observed spatial data and a set of dyadic weights that are generally defined based on the proximity of observations to each other, but can be generalized to accommodate geographic distances among observations and the lengths of shared boundaries for the regions from which the data arise. Geary's c can be used to test hypotheses about spatial autocorrelation under both parametric and nonparametric assumptions. Also, components of Geary's c can be used to assess how autocorrelation may vary throughout the spatial domain. We demonstrate the application of Geary's c to assess spatial structure in vegetation data arising from a survey in Alaska, USA, using two different approaches for defining proximity.