Socio-Demographic Differentials in Experiencing a Major Occupational Injury in the Prime Working Ages: Estimation Using Within-Survey and Cross-Survey Multiple Imputation of Injury Histories
Michael S. Rendall, RAND Corporation
Bonnie Ghosh-Dastidar, RAND Corporation
Margaret M. Weden, RAND Corporation
Zafar Nazarov, RAND Corporation
Sociodemographic differentials in ever experiencing a major workplace injury in the prime working ages (25 to 44) are estimated from left- and right-censored injury histories in the National Longitudinal Survey of Youth (NLSY79), and the Survey of Income and Program Participation (SIPP). Because the injury hazard is higher for individuals with previous injuries, the age-specific hazard for those with no previous injury since age 25 must first be estimated. Injury histories from age 25, however, are available for a fraction of the NLSY sample and for none in the SIPP sample (only the most recent injury is recorded). For unbiased incorporation of all NLSY and SIPP observations, injury histories are first multiply imputed within the NLSY from non-left-censored histories. Injury histories are then multiply imputed from this “completed” NLSY dataset to every SIPP individual. Efficiency and bias of NLSY-only and NLSY-SIPP estimation are compared to estimation that ignores injury history.
See paper
Presented in Session 122: Methodological Issues in Health and Mortality: Longitudinal studies