Answers are grounded in this paper and cite their sources.
SUMMARY Statistical methods analogous to analysis of variance are given for analysing hetero- geneity among standardized mortality ratios. Both the case in which an external standard reference population is to be used and the case in which such a standard must be estimated from the sample data are treated. As an example, grouping populations by latitude is shown to account for 93-9 per cent of melanoma standard mortality ratio heterogeneity among nine U.S. populations, and the between-group heterogeneity was highly statistically significant. THIs paper gives methods for analysing heterogeneity among standardized mortality ratios (SMRs). The methods are analogous to analysis of variance and can be applied to any epidemiologic hypothesis which partitions populations into groups. As an example we show that grouping nine U.S. populations into four northern, two middle latitude and three southern populations accounts for 93 9 per cent of the melanoma SMR heterogeneity and that the corresponding between group heterogeneity is highly statistically significant. The analysis of heterogeneity is based on an assumed product model which has been used by Kilpatrick (1962), Bjarnason et al. (1974), Gart (1974), Breslow and Day (1975) and Osborn (1975). In Section 2 we point out that the SMR is a good summary index of population mortality precisely when the product model holds, and we review methods for testing this model. Section 3 contains the methods needed for an analysis of heterogeneity based on the product model. For completeness we give methods both for the case in which an external standard reference population is to be used and for the case in which an internal standard is to be estimated from the observed populations.
No takes yet. Share an insight, caveat, or question.
Mitchell H. Gail (1978) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: