Variability in Life-Table Estimates Based on the National Health Interview Survey Linked Mortality Files

• Activity Number: 284
• Type: Contributed
• Date/Time: Tuesday, August 4, 2009 : 8:30 AM to 10:20 AM
• Sponsor: Section on Survey Research Methods
• Author(s): Van L. Parsons*+ and Nathaniel Schenker and Kimberly Lochner and Gloria Wheatcroft and Elsie R. Pamuk
• Companies: National Center for Health Statistics and National Center for Health Statistics and National Center for Health Statistics and National Center for Health Statistics and National Center for Health Statistics
• Keywords: Balanced Repeated Replication ; Variance Estimation

Abstract

This study investigates methods for assessing variability of estimated life-table functions when data are longitudinal and involve a complex survey. Traditional methods for estimating such variability, developed by Chiang, are appropriate when the observations are independent. With longitudinal data and a complex survey design, however, there can be two types of dependence: (1) within age groups, due to, e.g., clustering in the survey; and (2) across age groups, due to observations for a single individual contributing information for more than one age group. Chiang's method is compared to balanced repeated replication in the context of a study of disparities in life expectancy, for which data are obtained through linkage of respondents in the National Health Interview Survey to death records. Differences in the two methods due to the two types of dependence discussed above are explored.