Error in the measurement of mortality: an application to the analysis of racial mortality disparity

Yu-Chieh Hsu, University of Chicago and NORC

This paper examines the nature of measurement error in the reporting of deaths in panel data sets, using the National Longitudinal Survey of Older Men (NLS-OM) as a case study. The NLS-OM collected socioeconomic data for men aged 45-59 in 1966 and in several subsequent years, and then also recorded deaths---going so far as to match with death certificate data collected in 1990. Panel data of this sort are extremely useful for examining the antecedents of mortality, e.g., studying racial differences in mortality rates. However, considerable care must be taken when analyzing such data; theoretical reasoning developed in this paper shows that the most likely forms of error in the measurement of mortality can bias estimates of the racial mortality gap. An examination of the 1990 data suggests that the match of the death certificates was less complete for blacks than for whites. In consequence, standard practice leads to an under-estimation of the black-white mortality gap. Importantly, there is now a new match of NLS-OM data to death records, and analysis of these new data confirms this finding.