Estimating Worker Populations Using Capture-Recapture Methods. Preliminary Report.
NIOSH 1981:15 pages
The usefulness of various models for estimating worker populations through memory was evaluated. Five randomly selected faculty members were asked to list all faculty members in their department at Hunter College, past and present, and to give the year each was hired and the year last employed. Each individual scanned the other lists and added or corrected entries. Data was stratified by the year of last employment or by the number of years employed at the college as well as the year last employed. All results were compared with those known from employment records. The coverage of the faculty population estimated by those making the lists was 97 percent. True coverage was 76 percent when the lists were not stratified. Singly stratified estimates were more accurate and doubly stratified estimates were very close to true values with very narrow confidence intervals. The author concludes that it is reasonable to conduct a similar study at NIOSH based on doubly stratified data. Modification of a method of group recall that includes the capture/recapture method can be used to estimate worker populations.
NIOSH-Contract; Task-performance; Education; Group-dynamics; Occupational-psychology; Mathematical-models; Mental-processes; Information-retrieval-systems; Analytical-methods; Employees; Work-analysis;
NIOSH, U.S. Department of Health and Human Services, Task Order No. 81-2078, 15 pages