This report presents two sets of national "census survival rates" designed for use in computing estimates of net migration by age, color, and sex, for various subgroups of the United States population, particularly geographic subdivisions. One set is based on the total population of the United States; the other is based on the native population, excluding the foreign born. The report also presents a set of rates for the Negro population. Rates of this kind have been used widely for the purpose of estimating net migration for States and local areas between censuses^/1 because of the advantages they have in this application over reported death statistics by age or over life table survival rates.² The sets of rates shown in this report have been prepared and are being made available in response to a number of requests for such figures for use in connection with studies of internal migration. They may be applied in estimating intercensal net migration for 1950 to 1960 for geographic areas (States, counties, cities, etc.), categories of residence (urban and rural, farm and nonfarm), and various other subgroups in the population (e.g., native, and foreign born by national origin; distribution by educational level).

Census survival rates are based on the population as enumerated by age in two successive censuses and represent the ratio of population in a given age group at the second census to the population in the same "cohort"³ at the earlier census. Such ratios are affected not only by intercensal mortality but also by changes in the census net undercounts^/4 for the cohort from the first to the second census. For this reason, the ratios may change irregularly from one age group to the next and may even have a value in excess of unity. The national census survival rates presented here, however, exclude the effect of both internal migration and net immigration. (Ideally, national census survival rates should relate to a completely closed population, that is, one not affected by migration; such a population can only be approximated for the United States, but little error is believed to result from this approximation.) Hence, when they are applied to the population of some local area in 1950, the difference between the expected survivors and the census population in 1960 represents net migration alone, since mortality and changing census errors have presumably been allowed for by the census survival rates.

The discussion in this report, is concerned solely with the use of national census survival rates for 1950 to 1960 in deriving estimates of amounts of net migration from census data for the 1950–60 decade. It is not concerned as such with the estimation of rates of net migration for this particular decade nor with the estimation of amounts or rates of net migration for use in preparing population projections for local areas. These are special problems requiring special solutions.⁵

__________
¹ See, particularly, Bogue, reference 1; Economic Research Service, U.S. Department of Agriculture, reference 2; Lee and others, reference 8; and Miller, reference 9. References are shown in bibliography on page 7.
² See Hamilton and Henderson, reference 5, and Lee and others, reference 8, pp. 25–27.
³ The term cohort as used here refers to a group of persons born in the same year or group of years (e.g., ages 20–24 in 1950 and ages 30–34 in 1960).
⁴ Census net undercounts comprise both net underenumeration and misreporting of age.
⁵ For example, the selection of the base of migration rates, for combination with alternative estimates of net migration, for various uses, is a matter for separate consideration. With specific regard to population projections, use of local census survival rates has been proposed for this purpose; these allow simultaneously for local variations in mortality, in the pattern of census net undercounts, and in the rate of net migration. See C. Horace Hamilton, "Practical and Mathematical Considerations in the Formulation and Selection of Migration Rates," Demography, Vol. II, 1965; C. Horace Hamilton, "Educational Selectivity of Net Migration from the South," Social Forces, Vol. 38, No. 1, October 1959, pp. 33–42, esp. 40–42; Ralph Thomlinson, "The Determination of a Base Population for Computing Migration Rates," Milbank Memorial Fund Quarterly, Vol. 40, No. 3, July 1962, pp. 355-366; C. Horace Hamilton and Josef Perry, "A Short Method for Projecting Population by Age From One Decennial Census to Another," Social Forces, Vol. 41, No. 2, Dec. 1962, pp. 163–170; Kripalani, reference 6; and Tarver, reference 14.