Spatial Autocorrelation

Navigation:  Data Exploration and Spatial Statistics >

Spatial Autocorrelation

Previous pageReturn to chapter overviewNext page

The term spatial autocorrelation owes its origins to work in a related field, time series analysis (TSA), and in turn to the notion of correlation in univariate statistics. Since spatial autocorrelation follows these concepts so closely we will introduce the subject using these more familiar origins. Readers may also wish to refer to Cliff and Ord (1981), Goodchild (1986, CATMOG 47), or Fortin and Dale (2008). Throughout this discussion it is important to recognize that correlation may suggest, but does not imply, causation; it is extremely easy to make erroneous assumptions regarding cause-effect relationships from correlation data, but such results should be examined in great detail before suggesting any causation.