Douglas M. Flewelling, Ph.D.
Russell Weaver, Ph.D.
This paper statistically tests the accuracy and efficacy of areal interpolation functionality introduced in ESRI’s ArcGIS 10.1 Geostatistical Wizard. Esri has implemented a modified kriging-based technique for areal interpolation in the Geostatistical Wizard which is intended to simplify the overall process of interpolation. However, since kriging assumes data stationary, the feasibility of this method is questionable when applied to censusbased socioeconomic data. To test the accuracy and efficacy of kriging-based areal interpolation, this project tested the efficacy of esri’s methodology using nine variables representing continuous, count and ratio measurement scale data in the Combined Statistical Area of Kansas City, Kansas, Kansas City, Missouri, and Topeka, Kansas. The testing methodology employed both a Wilcoxon signed ranks test and a 90% confidence interval to determine whether kriging-based areal interpolation estimates generated by esri’s Geostatistical Wizard would reproduce results considered acceptable by the U.S. Census Bureau.
Carson, B. D. (2013). Testing Kriging-Based Areal Interpolation for Census-Based Socioeconomic Data (Master's thesis, University of Redlands). Retrieved from http://inspire.redlands.edu/gis_gradproj/194