Search Results

Author: Brown, James N.
Resulting in 1 citation.
1. Brown, James N.
Light, Audrey L.
Interpreting Panel Data on Job Tenure
Journal of Labor Economics 10,3 (July 1992): 219-257.
Cohort(s): Older Men, Young Men
Publisher: University of Chicago Press
Keyword(s): Data Quality/Consistency; Job Tenure; Panel Study of Income Dynamics (PSID); Variables, Independent - Covariate

This paper discusses the quality of job tenure data within two longitudinal data sets, the PSID and the NLS (Older Men and Young Men). It focuses on two major issues facing researchers using these data sources, namely: (1) the ability to identify distinct jobs within each data set; and (2) the choice of how and whether to clean the data to ensure that tenure and time increment by the same amount within jobs. The following four problems are highlighted: (1) The use of error- ridden tenure data to identify job changes may seriously reduce the reliability of estimates that require precise information about job changes. Estimates of the effect of job change on wage growth and the effects of many covariates on the probability of job separation appear especially sensitive to this problem. (2) Researchers who must identify job changes from tenure responses are likely to overstate the precision of their estimates, because standard errors (as conventionally computed) do not account for the effect of partition error on parameter estimates. (3) The failure to use internally consistent tenure sequences can lead to misleading conclusions about the slope of wage-tenure profiles. It does not appear to matter how internal consistency is imposed, as long as it is done. (4) The inclusion of jobs that contain unusually inconsistent tenure responses can alter the results in certain applications. Methods for utilizing these data are offered and recommendations for improvements in the PSID and NLS survey instruments suggested.
Bibliography Citation
Brown, James N. and Audrey L. Light. "Interpreting Panel Data on Job Tenure." Journal of Labor Economics 10,3 (July 1992): 219-257.