ASSESSING DIRECTIONAL EFFECTS IN SPATIAL DATA

A variable is measured at two locations separated by a given distance. Are the values more similar to each other if the locations are orient ed in one direction than another? This question has application to stu dies of human genetics, epidemics, and acid rain. One obvious analytic approach, regression on latitude and longitude, fails when data are n on-directional (isotropic) but spatially autocorrelated. Moreover, alt hough non-zero slope implies similarity between neighbours, the conver se is not true. IDIFF, a statistic derived from Moran's coefficient of spatial autocorrelation, is developed to detect general directional e ffects that apply to the collection of data points. Simulations sugges t that, when data have isotropic spatial autocorrelation but are incor rectly assumed to be independent, IDIFF will at worst reject too littl e. IDIFF has good power to distinguish epidemics that spread non-direc tionally from those that spread in a favoured direction.