Estimating deaths for the United States in 1900 by cause, age, and sex.

MOST HISTORICAL COMPILATIONS of mortality statistics for the United States begin with 1900, the year for which the Bureau of the Census began the annual collection of mortality statistics. (Responsibility for national vital statistics was transferred to the Public Health Service in 1946.) In 1900, however, the registration of deaths by State governments was still imperfect in many parts of the country. Hence, the Census Bureau limited tabulations of the data to those areas in which registration was believed to be reasonably complete and accurate.These areas became known as the Death-Registration Area (DRA), which originally consisted of 10 States and the District of Columbia-called Death-Registration States (DRS)and a number of cities in nonregistration States. As registration improved, more States were added to the DRA. In 1933, the DRA finally included the whole of the continental United States (1). Two limitations have been, at the least, an inconvenience for researchers: (a) the published death rates for the DRS in 1900 represented only 20 million of the country's 76 million population and (b) that population was solely from the northeast and northcentral regions. The resulting bias in historical series purporting to represent the nation is one which researchers have had to pass over with little more than a reference. The best possible estimate of mortality in the rest of the country would require intensive study of local data and a large investment, and the need did not appear to justify these.