Premium
Spatial Autocorrelation and the Quantitative Revolution
Author(s) -
Haining Robert P.
Publication year - 2009
Publication title -
geographical analysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.773
H-Index - 65
eISSN - 1538-4632
pISSN - 0016-7363
DOI - 10.1111/j.1538-4632.2009.00763.x
Subject(s) - spatial analysis , autocorrelation , descriptive statistics , statistics , independent and identically distributed random variables , principal (computer security) , econometrics , spatial econometrics , statistical hypothesis testing , geography , computer science , data science , mathematics , random variable , operating system
This article reflects on the principal statistical methodological issues that quantitative geographers addressed in the 1950s and 1960s. The source of many of these methodological problems lies in the assumptions of classical statistics that data are independent and identically distributed. The correlation inherent in spatial data recorded at nearby locations and the irregular nature of spatial frameworks are contrary to these assumptions and impact upon the use of both descriptive and inferential statistics. Most of the wider developments in statistics at that time still fell short of meeting the needs of statistical geographers. What emerged were statistical tests for pattern, nonmodel based but capable of being adapted to an irregular areal framework and certain other factors considered important in any particular spatial analysis. This article notes that, while spatial autocorrelation was presented as a “problem” for geographers, it is also important to realize that its presence can represent an opportunity.