Abstract
The use of intercentroid distances to measure spatial separation between discrete zones, introduces substantial errors into the measurement of distances between zones. These errors can produce important distortions; for instance, in the performance of spatial models based on a power or exponential function of distance. The nature and source of the errors and the magnitude of their effect on the performance of spatial interaction models are established for different types of situations and for different degrees of complexity in the network along which distances are measured. Some suggestions are made on an approach to correcting these errors; a particular approximation including the self-distances in the calculation of interzonal distances is discussed, and it is shown that it reduces the magnitude of the errors to more acceptable levels.
Get full access to this article
View all access options for this article.
