A Note on the Census Survival Ratio Method of Estimating Net Migration

Abstract

In the census survival ratio method, one of the common assumptions is that the proportion which the enumerated population in any age-sex group bears to the actual population is the same at each census for each state as for the nation. The significance of this assumption is examined and it is established that such an assumption is not necessary. It is therefore replaced by a less stringent assumption, namely that the ratio of the degree of enumeration in a state (i.e. the proportion which the enumerated population in any age-sex group bears to the true population) to that of the nation is the same for the same cohort. In the light of the change in assumption, it is shown that much of the criticism of the method made by Daniel O. Price is untenable. Though his main conclusion (i.e. small relative differences in estimates of net migration should be interpreted with extreme caution) is correct, the errors in migration rates estimated by the census survival ratio method are probably much smaller than his analysis suggests.