Testing regression residuals for spatial autocorrelation using SAS: a technical note

Spatial autocorrelation of regression residuals is a violation of an assumption of the general linear model. Yet despite the widespread use of spatial data in applied work, many researchers fail to examine residuals for spatial autocorrelation, or they rely on subjective interpretation of residual maps. In part, this problem stems from a lack of easy-to-use programs that can be incorporated into existing research procedures. In this note we present a short program for testing spatial association among regression residuals. Based on Moran's I statistic, the program employs SAS's PROC MATRIX language, and can easily augment regression analyses run using the SAS package. -Authors