Abstract
By working out an (artificial) example, the principle is demonstrated that between-groups comparisons of rates may be entirely misleading, 1) if the comparisons are made on the basis of crude rates, 2) and if these rates are composed of specific rates which, by definition or by action of Nature, are different, and weighted together in unequal proportions. To overcome some of the difficulties, the technique of standardized is presented; namely that the specific rates within each group be weighted together with weights which are equal from group to group. Effects of selection of the weights are discussed, and the variance of the standardized rate is derived.