THE PAST IS PROLOGUE:

Abstract

Mahalanobis provided an example of how to use statistics to enlighten and inform government policy makers. His pioneering work was used by the US Bureau of the Census to learn more about measurement errors in censuses and surveys. People have many misconceptions about censuses, among them who is to be counted and where. Errors in the census do occur, among them errors in coverage. Over the years, the US Bureau of the Census has developed statistical techniques, including sampling in the census, to increase accuracy and reduce response burden. A root-mean-square-error model was developed to estimate the joint effects of variance and bias in the census. The model is used in this paper to look at the joint effects of response variance, adjustment of the bias caused by the undercount, and the use of sampling for follow-up.