Years on the NLS team: 1995-2000
Role: Provided the survey design and estimation procedures for the NLSY97 and PAY97.
NLSY97/PAY97 was one of the first large-scale social surveys to be designed on the basis of data collected the 1990 Decennial Census. It was also one of the first surveys to utilize tools and databases of modern geography, especially the Census Bureau's new TIGER (Topologically Integrated Geographic Encoding and Referencing) database, which I had a role in developing some years earlier. The NLSY estimation procedure employed a calibration of the survey weights to the new census population totals. The design incorporated an oversample of Black and Hispanic youths. To increase the precision of the survey statistics, we selected the primary and second stage sampling units with probability proportional to a weighted combination of the black, Hispanic, and Nonblack-NonHispanic populations. The overall complex sampling design incorporated three separate surveys: NLSY97, the Student Testing Program, and the Enlistment Testing Program. Importantly, the survey screener collected the ages of all residents of the household. These data enabled analysis of the age distribution of all persons in responding households, which in turn revealed that youths age 12 to 16 years had been substantially under-reported by screener respondents. I worked extensively with the NLSY technical advisory committee and with the Defense Advisory Committee to examine and verify the representativeness of the three surveys and to devise and implement appropriate methods of estimation.
I am delighted to have had a role in creating America's preeminent database for the study by economists, sociologists, and other scientists of the transition from school to work and from adolescence to adulthood. The NLSY data offer tremendous insight into these key social processes and they inform public policy and debate.