Since about 1960, there has been a rapidly increasing technical literature on the concept of spatial autocorrelation and its implications for statistical estimation in geography. Because this literature has tended, of necessity, to be theoretical and difficult to follow, many potential users have either shied away from it or misunderstood its usefulness and its limitations. This is unfortunate. Many empirical geographers could beneficially use the techniques developed in this area, especially in regression analysis with zonal data. Further, this literature identifies a number of fundamental conceptual issues in geographic research of which every geographer should be aware. It is with these concerns in mind that I have written this paper. The purpose of the paper is to provide a relatively non-technical introduction to spatial atuocorrelation and its implications for regression analysis.
[1]
P. J. Diggle,et al.
Exploratory and Explanatory Statistical Analysis of Spatial Data
,
1981
.
[2]
Ian Masser.
Theory and practice in regional science
,
1976
.
[3]
Tsoung-Chao Lee,et al.
The Theory and Practice of Econometrics
,
1981
.
[4]
Andries S. Brandsma,et al.
Further evidence on alternative procedures for testing of spatial autocorrelation among regression disturbances
,
1979
.
[5]
S. Arora,et al.
Alternative Approaches to Spatial Autocorrelation: An Improvement Over Current Practice
,
1977
.
[6]
K. Ord.
Estimation Methods for Models of Spatial Interaction
,
1975
.
[7]
L. Hordijk.
Spatial correlation in the disturbances of a linear interregional model
,
1974
.