Table 2 

comparison of classical regression and GWR results 

CHD 
Classical regression 
GWR 



O against E 
O = E+GP 
O against E 
O = E+GP 



Residual sum of squares 
0.032 
0.026 
0.026 
0.017 


Standard deviation 
0.010 
0.009 
0.009 
0.007 


Akaike Information Criterion 
2260.43 
2336.25 
2285.75 
2393.81 


Correlation coefficient 
0.299 
0.439 
0.427 
0.637 


Adjusted correlation coefficient 
0.295 
0.434 
0.387 
0.585 


Sum of squares 
0.0 
0.0 
0.0 
0.0 


Degrees of freedom 
2.00 
3.00 
328.09 
306.86 


Hypertension 
Classical regression 
GWR 



O against E 
O = E+GP 
O against E 
O = E+GP 



Residual sum of squares 
0.374 
0.250 
0.362 
0.241 


Standard deviation 
0.033 
0.027 
0.032 
0.026 


Akaike Information Criterion 
1400.41 
1539.57 
1403.50 
1543.67 


Correlation coefficient 
0.121 
0.412 
0.150 
0.432 


Adjusted correlation coefficient 
0.116 
0.407 
0.134 
0.421 


Sum of squares 
0.4 
0.2 
0.4 
0.2 


Degrees of freedom 
2.00 
3.00 
344.87 
344.04 


Stroke 
Classical regression 
GWR 



O against E 
O = E+GP 
O against E 
O = E+GP 



Residual sum of squares 
0.007 
0.006 
0.005 
0.003 


Standard deviation 
0.004 
0.004 
0.004 
0.003 


Akaike Information Criterion 
2807.25 
2873.16 
2838.92 
2932.93 


Correlation coefficient 
0.262 
0.392 
0.422 
0.621 


Adjusted correlation coefficient 
0.258 
0.387 
0.374 
0.561 


Classical regression 
GWR 



O against E 
O = E+GP 
O against E 
O = E+GP 



Sum of squares 
0.0 
0.0 
0.0 
0.0 


Degrees of freedom 
2.00 
3.00 
324.37 
302.87 


O against E: ratio of observed against expected prevalence O = E+GP: inclusion of GP supply as an additional independent variable 

Soljak et al. BMC Cardiovascular Disorders 2011 11:12 doi:10.1186/147122611112 