TY - JOUR
AU - Griffith, Daniel A.
AB -  “ I should like to acknowledge the helpful comments of Dr. L. Curry of the Geography Department at the University of Toronto, and Dr. E. Sheppard of the Geo aphy Department at the University of Minnesota on an earlier version of this paper. Of course, I g n e am responsible for any errors or omissions. Daniel A. Griffith is professor of geogrcipliy, Ryerson Polytechnical Institute. C~16-7363/78/0778-0296$00.50/0 0 1978 Ohio State University Press GEOGRAPHICAL ANALYSIS, X, no. 3 (July 1978) vol. Research Notes and Comments ring. This method utilizes the matrix model Y = py+ t + E, where (1) Y = n x 1 vector of measures for some characteristic py= n X 1 mean-score vector t = n x 1 vector of differences between the mean-score vector entry and the mean of region i( j =1,2,. . .,M ) to which areal unit k ( k = 1,2,. . .,n) belongs E = n x 1 vector whose entries are normally distributed, random, uncorrelated error terms having a mean of zero and a constant variance. Draper and Smith [ 1 3 ]have shown that equation (1) may be rewritten as the following regression equation: Y=XP+E, (2) 
TI - A Spatially Adjusted ANOVA Model
JF - Geographical Analysis
DO - 10.1111/j.1538-4632.1978.tb00661.x
DA - 1978-07-01
UR - https://www.deepdyve.com/lp/wiley/a-spatially-adjusted-anova-model-SjGzaTWxve
SP - 296
VL - 10
IS - 3
DP - DeepDyve
ER -