The Missing Link: Data Capture Technology and the Making of a Longitudinal U.S. Census Infrastructure

Abstract

Social scientists’ use of linked decennial census data has grown extensively over the past two decades. For U.S. census data before 1950, a large body of linked data has been made available within the past few years. The 2000 and 2010 decennial data have been linked to one another and back to the 1940 Census. For the censuses in between these years—from 1950 to 1990—no linked data are available yet. This paper focuses on the technological advances in data capture that have enabled this centuries-long longitudinal data infrastructure to flourish while still leaving a sizeable “missing link” from 1950 to 1990. We will describe the development of modern technology to capture handwritten data at the Census Bureau, and ongoing efforts to digitize remaining information from and create linkages between the censuses of 1950 to 1990.