The Use of Historical Census Data for Mortality and Fertility Research
NBER Historical Working Paper No. 31
This paper illustrates the application of indirect techniques of fertility and mortality estimation to historical census data, both in published form and as micro census samples derived from the original enumerators' manuscripts. There are many instances in which census data exist but adequate vital registration data do not, such as in the United States prior to 1933, when the Birth and Death Registration Areas finally covered the entire nation. Since the United States has taken decennial censuses since 1790, and since all the original population schedules except those for 1890 have been preserved, it is possible to apply these indirect methods. For example, the censuses of 1900 and 1910 asked questions on children ever born, children surviving, and duration of current marriage, but this information was never tabulated or used for 1900 and only partly tabulated for 1910. The Public Use Samples of the 1900 and 1910 censuses make possible the utilization of those data to estimate levels, differentials, and even recent trends in childhood mortality. Application of own-children methods to samples of the censuses since 1850 permits estimation of age-specific overall and marital fertility rates. Finally, the use of the 1900 Public Use Sample in conjunction with published data on parity from the 1910 census (or tabulations from the 1910 Public Use Sample) allows application of the two-census, parity increment method of birth rate estimation.