We have 3 areas, with varying rate ratios for smoking. After standardization, the SMR is 5 and this is equal to the crude rate ratio of 5 (in the Total column). Therefore, there is no confounding using the individual level data. The ecological regression analysis, using % smokers and the cancer rate, shows an inverse relationship (intercept of 9 and slope of -3), and this is due to the differences in rate differences between groups.