Mark E. Hill

Childhood conditions that predict survival to advanced ages among African-Americans

This paper investigates the social and economic circumstances of childhood that predict the probability of survival to age 85 among African-Americans. It uses a unique study design in which survivors are linked to their records in U.S. Censuses of 1900 and 1910. A control group of age and race-matched children is drawn from Public Use Samples for these censuses. It concludes that the factors most predictive of survival are farm background, having literate parents, and living in a two-parent household. Results support the interpretation that death risks are positively correlated over the life cycle.

African-american mortality at older ages: Results of a matching study

In this paper we investigate the quality of age reporting on death certificates of elderly African-Americans. We link a sample of death certificates of persons age 65+ in 1985 to records for the same individuals in U.S. censuses of 1900, 1910, and 1920 and to records of the Social Security Administration. The ages at death reported on death certificates are too young on average. Errors are greater for women than for men. Despite systematic underreporting of age at death, too many deaths are registered at ages 95+. This excess reflects an age distribution of deaths that declines steeply with age, so that the base for upward transfers into an age category is much larger than the base for transfers downward and out. When corrected ages at death are used to estimate age-specific death rates, African-American mortality rates increase substantially above age 85 and the racial “crossover” in mortality disappears. Uncertainty about white rates at ages 95+, however, prevents a decisive racial comparison at the very oldest ages.

Age reporting among white Americans aged 85+: results of a record linkage study.

This study investigates age reporting on the death certificates of older white Americans. We link a sample of death certificates for native-born whites aged 85+ in 1985 to Social Security Administration records and to records of the U.S. censuses of 1900, 1910, and 1920. When ages in these sources are compared, inconsistencies are found to be minimal, even beyond age 95. Results show little distortion and no systematic biases in the reported age distribution of deaths. To explore the effect of age misreporting on old-age mortality, we estimate “corrected” age-specific death rates by the extinct-generation method for the U.S. white cohort born in 1885. With few exceptions, corrected and uncorrected rates in single years differ by less than 3% and are not systematically biased. When we compare corrected rates with those for the same birth cohort in France, Japan, and Sweden, we find that white American mortality at older ages is exceptionally low.

The Accuracy of Age Reporting among Elderly African Americans Evidence of a Birth Registration Effect

This article expands on previous research that has documented disparities in age information for elderly African Americans. Drawing on a sample of death certificates for Maryland-born African Americans purportedly aged 65-79 years, the validity of age data on both death certificates and social security records is examined through comparison with birth dates provided on linked birth records. The commonly assumed relationship between birth registration and quality of age reporting also is investigated. When compared to birth records, age reporting on social security records is significantly more accurate than that on death records. Age agreement between matched death and social security records closely reflects age validity as determined from birth records. Findings based on logistic regression analysis support the hypothesized birth registration effect: Controlling for demographic characteristics, persons known to have birth certificates exhibited significantly greater age agreement on linked death certificates and social security records (odds ratio = 2.3).

Linking Death Certificates to Early Census Records: The African American Matched Records Sample

In this article we report record-linkage methods used to construct the African American Matched Records Sample (AAMRS).... We also identify those characteristics that contributed to successful linkage and draw lessons for future linkage efforts.... Our study was designed to link death certificates from 1980 and 1985 to early-life census records for the same individuals to assess the accuracy of age reporting on death certificates and improve mortality estimates for older African Americans. (EXCERPT)

MULTIVARIATE SURVIVORSHIP ANALYSIS USING TWO CROSS-SECTIONAL SAMPLES*

As an alternative to survival analysis with longitudinal data, I introduce a method that can be applied when one observes the same cohort in two cross-sectional samples collected at different points in time. The method allows for the estimation of log-probability survivorship models that estimate the influence of multiple time-invariant factors on survival over a time interval separating two samples. This approach can be used whenever the survival process can be adequately conceptualized as an irreversible single-decrement process (e.g., mortality, the transition to first marriage among a cohort of never-married individuals). Using data from the Integrated Public Use Microdata Series (Ruggles and Sobek 1997), I illustrate the multivariate method through an investigation of the effects of race, parity, and educational attainment on the survival of older women in the United States.

Mortality and Fertility Trends

This chapter presents new estimates of fertility and mortality for African Americans. The estimates are based upon the elaborate reconstructions of population and death described in Chapter 5. They provide a more secure basis for understanding the demographic circumstances of African Americans than is available from official sources. Even though our population reconstructions are similar to those of the U.S. Census Bureau for most ages in the most recent period, it should be recognized that official estimates of fertility and mortality made by the National Center for Heath Statistics do not correct for census omission or age misstatement. And no series, official or unofficial, is based upon the full range of data assessments and the comprehensive demographic framework that are introduced here.

Understanding the Sources of Age Misreporting

The volume of age misreporting among African Americans described in the previous chapter provides a serious challenge to demographic estimation. Age misreporting also poses a vexing problem for demographers who study populations in developing countries where vital registration is often deficient and where the demographic estimates must be based primarily on analysis of age distributions of populations or deaths (United Nations, 1982). There have been few attempts to identify the sources of age misreporting (Ewbank 1981). The record linkage described in the last chapter provides a rare opportunity to investigate the correlates of age misreporting. In addition, we have conducted a supplementary series of 40 in-depth interviews with elderly African Americans in Philadelphia to investigate concepts of age and of age accuracy. These individuals represent a “convenience sample” of persons drawn primarily from senior centers. We acknowledge the efforts of Averil Clarke and Cathleen Riddley, who conducted most of these interviews.

Childhood Conditions That Predict Survival to Advanced Ages

Studies of social and economic differentials in mortality typically relate circumstances at one moment in time to contemporary mortality risks. Literally hundreds of studies that date back more than a century show that, with rare exception, socially and economically disadvantaged groups suffer elevated risks of death (Williams 1990; Feinstein 1993). Such results are hardly surprising. Healthiness and longevity are nearly universal goals, and groups with more economic and social resources are better equipped to achieve these goals.

Ascertaining True Ages at Death Through a Matching Study

In Chapter 2, we identified major inconsistencies in national data on African Americans. Problems were shown to be more acute at older ages. An important clue about the source of errors is provided by a study by Kitagawa and Hauser of data for 1960, the middle of the period under investigation. Kitagawa and Hauser’s study matched records from death registration and from the 1960 U.S. Census (National Center for Health Statistics 1968). Large inconsistencies in age reporting were identified, especially among older people. Only 44.7 percent of nonwhite males and 36.9 percent of nonwhite females had the same age reported on the death certificate and census form. The study design did not permit inaccuracies to be traced to either system. It is however, plausible to assume that inaccuracies were greater on death certificates because ages there are always provided by someone other than the subject.