The Extended Cohort-Component Method and Applications to the U.S. and China Yi Zeng, Kenneth C Land, Danan Gu, ... Chapter 6 Household and Living Arrangement Projections at the Small Area Level 6.1 Basic Concepts to Apply the ...
Age-Period-Cohort Analysis: New Models, Methods, and Empirical Applications is based on a decade of the authorsâ€™ collaborative work in age-period-cohort (APC) analysis.
The 14 papers in the volume are arranged in sections that: 1) describe the organizational and political context of applied forecasting; 2) review the state-of-the-art for many forecasting models and methods; and 3) discuss issues of predictability, the implications of forecast errors, and model construction, linkage and verification. Most of the aspects covered are in the field of social phenomena, eg: econometrics, aggregate economic-demographic models, population projects, health status, and technological innovation and diffusion. The papers are abstracted separately with the original journal citations. -M.A.Bass
In this paper, we consider the following question for the analysis of data obtained in longitudinal panel designs: How should repeated-measures data be modeled and interpreted when the outcome or dependent variable is dichotomous and the objective is to determine whether the within-person rate of change over time varies across levels of one or more between-person factors? Standard approaches address this issue by means of generalized estimating equations or generalized linear mixed models with logistic links. Using an empirical example and simulated data, we show (1) that cross-level product terms from these models can produce misleading results with respect to whether the within-person rate of change varies across levels of a dichotomous between-person factor; and (2) that subgroup differences in the rate of change should be assessed on an additive scale (using group differences in the effects of predictors on the probability of disease) rather than on a multiplicative scale (using group differences in the effects of predictors on the odds of disease). Because usual approaches do not provide a significance test for whether the rate of additive change varies across levels of a between-person factor, sample differences in the rate of additive change may be due to sampling error. We illustrate how standard software can be used to estimate and test whether additive changes vary across levels of a between-person factor.
This study examines and further develops the classic Strehler-Mildvan (SM) general theory of mortality and aging. Three predictions from the SM theory are tested by examining the age dependence of mortality patterns for 42 countries (including developed and developing countries) over the period 1955-2003. By applying finite mixture regression models, principal component analysis, and random-effects panel regression models, we find that (1) the negative correlation between the initial adulthood mortality rate and the rate of increase in mortality with age derived in the SM theory exists but is not constant; (2) within the SM framework, the implied age of expected zero vitality (expected maximum survival age) also is variable over time; (3) longevity trajectories are not homogeneous among the countries; (4) Central American and Southeast Asian countries have higher expected age of zero vitality than other countries in spite of relatively disadvantageous national ecological systems; (5) within the group of Central American and Southeast Asian countries, a more disadvantageous national ecological system is associated with a higher expected age of zero vitality; and (6) larger agricultural and food productivities, higher labor participation rates, higher percentages of population living in urban areas, and larger GDP per capita and GDP per unit of energy use are important beneficial national ecological system factors that can promote survival. These findings indicate that the SM theory needs to be generalized to incorporate heterogeneity among human populations.
The Gln(27)Glu polymorphism but not the Arg(16)Gly polymorphism of the beta2-adrenergic receptor (ADRB2) gene appears to be associated with a broad range of aging-associated phenotypes, including cancers at different sites, myocardial infarction (MI), intermittent claudication (IC), and overall/healthy longevity in the Framingham Heart Study Offspring cohort. The Gln(27)Gln genotype increases risks of cancer, MI and IC, whereas the Glu(27) allele or, equivalently, the Gly(16)Glu(27) haplotype tends to be protective against these diseases. Genetic associations with longevity are of opposite nature at young-old and oldest-old ages highlighting the phenomenon of antagonistic pleiotropy. The mechanism of antagonistic pleiotropy is associated with an evolutionary-driven advantage of carriers of a derived Gln(27) allele at younger ages and their survival disadvantage at older ages as a result of increased risks of cancer, MI and IC. The ADRB2 gene can play an important systemic role in healthy aging in evolutionary context that warrants exploration in other populations.
Multiple functions of the beta2-adrenergic receptor (ADRB2) and angiotensin-converting enzyme (ACE) genes warrant studies of their associations with aging-related phenotypes. We focus on multimarker analyses and analyses of the effects of compound genotypes of two polymorphisms in the ADRB2 gene, rs1042713 and rs1042714, and 11 polymorphisms of the ACE gene, on the risk of such an aging-associated phenotype as myocardial infarction (MI). We used the data from a genotyped sample of the Framingham Heart Study Offspring (FHSO) cohort (n = 1500) followed for about 36 years with six examinations. The ADRB2 rs1042714 (C-->G) polymorphism and two moderately correlated (r(2) = 0.77) ACE polymorphisms, rs4363 (A-->G) and rs12449782 (A-->G), were significantly associated with risks of MI in this aging cohort in multimarker models. Predominantly linked ACE genotypes exhibited opposite effects on MI risks, e.g., the AA (rs12449782) genotype had a detrimental effect, whereas the predominantly linked AA (rs4363) genotype exhibited a protective effect. This trade-off occurs as a result of the opposite effects of rare compound genotypes of the ACE polymorphisms with a single dose of the AG heterozygote. This genetic trade-off is further augmented by the selective modulating effect of the rs1042714 ADRB2 polymorphism. The associations were not altered by adjustment for common MI risk factors. The results suggest that effects of single specific genetic variants of the ADRB2 and ACE genes on MI can be readily altered by gene-gene or/and gene-environmental interactions, especially in large heterogeneous samples. Multimarker genetic analyses should benefit studies of complex aging-associated phenotypes.
PURPOSE: Health of the general population is improving along a number of major health dimensions. Using a cumulative deficits approach, we investigated whether such improvements were evident at the level of minor health traits. METHODS: We selected 37 small-effect traits consistently measured in the 9th (performed in 1964) and 14th (1974) Framingham Heart and 5th (1991-1995) Offspring Study exams to construct indices of cumulative deficits (DIs). RESULTS: We identified deficits-specific DIs characterizing health dimensions associated with no health changes (DI(NHC)), health worsening (DI(WRS)), and health improving (DI(IMP)) between the 1960s and 1990s. The risks of death attributable to the DI(NHC) dominate within shorter time horizons. For longer time horizons, both the DI(NHC) and DI(IMP) provide the same contribution to the risks of death. The mortality risks associated with the DI(WRS) are the weakest and least significant. CONCLUSIONS: The analyses show that the cumulative deficits approach might be an efficient tool for analyzing the effects of a large number of health characteristics for which the individual effects are small, inconsistent, or non-significant. They show favorable trends such that health of the Framingham studies participants either did not change or improved over time for the most serious small-effect traits.
OBJECTIVES: To compare how well frailty measures based on a phenotypic frailty approach proposed in the Cardiovascular Health Study (CHS) and a cumulative deficits approach predict mortality. DESIGN: Cohort study. SETTING: The main cohort of the CHS. PARTICIPANTS: Four thousand seven hundred twenty-one individuals. MEASUREMENTS: A phenotypic frailty index (PFI) was defined in the same way as proposed in the CHS: assessing weight loss, exhaustion, low physical activity, slowness, and poor grip strength. A cumulative deficit index (DI) was defined based on 48 elderly deficits (signs, symptoms, impairments, diseases) included in the index, with equal weights. RESULTS: Of the 1,073 frailest individuals with the lowest survival, the PFI, categorized as proposed in the CHS into robust, prefrail, and frail categories, underestimated the risk of death for 720 persons, whereas the DI categorized into the same three frailty categories underestimated the mortality risk for 134 persons. The higher power of the DI for discriminating frail individuals in their susceptibility to death also followed from comparison of quasi-instantaneous values of both indices. The three-level DI identified 219 individuals as frail of 361 individuals identified as frail according to the three-level PFI. CONCLUSION: The DI can more precisely evaluate chances of death because it assesses a broader spectrum of disorders than the PFI. Both indices appear to be frailty related. Integration of both approaches is highly promising for increasing the precision of discrimination of the risk of death and especially for identification of the most vulnerable elderly people.
OBJECTIVES: To investigate the relationship between body mass index (BMI) and 9-year mortality in older (> or = 65) Americans with and without disability. DESIGN: Cohort study. SETTING: The unique disability-focused National Long Term Care Survey (NLTCS) data that assessed the health and well-being of older individuals in 1994 were analyzed. PARTICIPANTS: Four thousand seven hundred ninety-one individuals in the 1994 survey. MEASUREMENTS: BMI (kg/m2) was calculated from self- or proxy reports of height and weight. The analysis was adjusted for 1-year change in BMI and demographic and health-related factors, as well as reports by proxies, and death occurring during the first 2 years after the interview. RESULTS: The relative risk of death as a function of BMI formed a nonsymmetric U-shaped pattern, with larger risks associated with lower BMI (
Based on unique data from the largest-ever sample of the Chinese oldest-old aged 80 and older, our multivariate logistic regression analyses show that either receiving adequate medical service during sickness in childhood or never/rarely suffering from serious illness during childhood significantly reduces the risk of being ADL (activities of daily living) impaired, being cognitively impaired, and self-reporting poor health by 18%-33% at the oldest-old ages. Estimates of effects for five other indicators of childhood conditions are similarly positive but mostly not statistically significant. Multivariate survival analysis shows that better childhood socioeconomic conditions in general tend to reduce the four-year period mortality risk among the oldest-old. But after additional controls for 14 covariates are put into the model, the effects are not statistically significant, thus suggesting that most of the effects of childhood conditions on oldest-old mortality are indirect-at least to the point of affecting current health status at the oldest-old ages, which itself is strongly associated with mortality. While acknowledging limitations of the present analyses due to a lack of information on childhood illness, the oldest-olds'recollection errors, and other data problems, we conclude, based on this and other studies, that policies that enhance childhood health care and children's socioeconomic well-being can have large and long-lasting benefits up to the oldest-old ages.
Background: We employ an approach based on the elaborated frailty index (FI), which is capable of taking into account variables with mild effect on the aging, health and survival outcomes, and investigate the connections between the FI, chronological age and the aging-associated outcomes in the elderly. Methods: Cross-sectional analysis of pooled data from the National Long Term Care Survey (NLTCS) assessing health and functioning of the U.S. elderly in 1982, 1984, 1989, 1994, and 1999. Results: Distributions of frequency, residual life span, mortality rate, and relative risk of death are remarkably similar over age and FI. Coefficients of correlation between FI and age are low both for males (0.127, p < .01) and females (0.221, p < .01). The FI-specific age patterns show deceleration at advanced ages. The FI can provide order of magnitude better resolution in estimating mean remaining life span compared to age. Males have smaller FI than females while males' mortality risks are higher. For short-time horizons, the FI and age are largely independently associated with mortality risks. Conclusions: The FI: (i) can be considered as an adequate sex-specific indicator of the aging-associated processes in the elderly, (ii) can characterize these processes independently of age, and (iii) is a better characteristic of the aging phenotype than chronological age. Â© 2006 Elsevier Ireland Ltd. All rights reserved.
Background: An index of age-associated health/well-being disorders (deficits), called the "frailty index" (FI), appears to be a promising characteristic to capture dynamic variability in aging manifestations among age-peers. In this study we provide further support toward this view focusing on the analysis of the FI age patterns in the participants of the National Long Term Care Survey (NLTCS). Methods: The NLTCS assessed health and functioning of the U.S. elderly in 1982, 1984, 1989, 1994, and 1999. Detailed information for our sample was assessed from about 26,700 interviews. The individual FI is defined as a proportion of health deficits for a given person. Results: The FI in the NLTCS exhibits accelerated age patterns. The acceleration is larger for elderly who, at younger ages, had a lower FI (low FI group) than for those who showed a higher FI at younger ages (high FI group). Age-patterns for low and high FI groups tend to converge at advanced ages. The rate of deficit accumulation is sex-sensitive. Conclusions: The accelerated FI age patterns suggest that FI can be considered as a systemic measure of aging process. Convergence of the (sex-specific) FI age patterns for low and high FI groups by extreme ages might reflect the limit of the FI-specific (or systemic) age as well as the limit of adaptation capacity in aging individuals. Â© 2006 Elsevier Ireland Ltd. All rights reserved.
The classic headship-rate method for demographic projections of households is not linked to demographic rates, projects a few household types without size, and does not deal with household members other than heads. By comparison, the ProFamy method uses demographic rates as input and projects more detailed household types, sizes, and living arrangements for all members of the population. Tests of projections from 1990 to 2000 using ProFamy and based on observed U.S. demographic rates before 1991 show that discrepancies between our projections and census observations in 2000 are reasonably small, validating the new method. Using data from national surveys and vital statistics, census microfiles, and the ProFamy method, we prepare projections of U.S. households from 2000 to 2050. Medium projections as well as projections based on smaller and larger family scenarios with corresponding combinations of assumptions of marriage/union formation and dissolution, fertility, mortality, and international migration are performed to analyze future trends of U.S. households and their possible higher and lower bounds, as well as enormous racial differentials. To our knowledge, the household projections reported in this article are the first to have found empirical evidence of family household momentum and to have provided informative low and high bounds of various indices of projected future households and living arrangements distributions based on possible changes in demographic parameters. Â© Springer 2006.
This study investigates an ill-posed problem (multicollinearity) in Hierarchical Linear Models from both the data and the model perspectives. We propose an intuitive, effective approach to diagnosing the presence of multicollinearity and its remedies in this class of models. A simulation study demonstrates the impacts of multicollinearity on coefficient estimates, associated standard errors, and variance components at various levels of multicollinearity for finite sample sizes typical in social science studies. We further investigate the role multicollinearity plays at each level for estimation of coefficient parameters in terms of shrinkage. Based on these analyses, we recommend a top-down method for assessing multicollinearity in HLMs that first examines the contextual predictors (Level-2 in a two-level model) and then the individual predictors (Level-1) and uses the results for data collection, research problem redefinition, model re-specification, variable selection and estimation of a final model.
OBJECTIVES: To better understand future home-based care needs and costs for disabled elders in China. METHOD: To further develop and apply the ProFamy extended cohort-component method and the most recent census and survey data. RESULTS: (a) Chinese disabled elders and the annual growth rate of the percentage of national gross domestic product (GDP) devoted to home-based care costs for disabled elders will increase much more rapidly than the growth of total elderly population; (b) home-based care needs and costs for disabled oldest-old aged 80+ will increase much faster than that for disabled young-old aged 65-79 after 2030; (c) disabled unmarried elders living alone and their home-based care costs increase substantially faster than those disabled unmarried elders living with children; (d) percent of rural disabled oldest-old will be substantially higher than that of rural population after 2030; (e) sensitivity analyses show that possible changes in mortality and elderly disability status are the major direct factors affecting home-based care needs and costs; (f) caregivers resources under the universal two-child policy will be substantially better than that under the rigorous fertility policy unchanged. DISCUSSION: We discuss policy recommendations concerning pathways to healthy aging with relatively reduced care costs, including reductions of the prevalence of disability, gender equality, the universal two-child policy and resources of caregivers, encouragements of rural-to-urban family migration and elder's residential proximity to their adult children, and remarriages of not-married elders.
Logistic regression analysis based on data from 822 Han Chinese oldest old aged 92+ demonstrated that interactions between carrying FOXO1A-266 or FOXO3-310 or FOXO3-292 and tea drinking at around age 60 or at present time were significantly associated with lower risk of cognitive disability at advanced ages. Associations between tea drinking and reduced cognitive disability were much stronger among carriers of the genotypes of FOXO1A-266 or FOXO3-310 or FOXO3-292 compared with noncarriers, and it was reconfirmed by analysis of three-way interactions across FOXO genotypes, tea drinking at around age 60, and at present time. Based on prior findings from animal and human cell models, we postulate that intake of tea compounds may activate FOXO gene expression, which in turn may positively affect cognitive function in the oldest old population. Our empirical findings imply that the health benefits of particular nutritional interventions, including tea drinking, may, in part, depend upon individual genetic profiles.
Â© 2015 Elsevier Ltd. Social scientists have recognized the importance of age-period-cohort (APC) models for half a century, but have spent much of this time mired in debates about the feasibility of APC methods. Recently, a new class of APC methods based on modern statistical knowledge has emerged, offering potential solutions. In 2009, Reither, Hauser and Yang used one of these new methods - hierarchical APC (HAPC) modeling - to study how birth cohorts may have contributed to the U.S. obesity epidemic. They found that recent birth cohorts experience higher odds of obesity than their predecessors, but that ubiquitous period-based changes are primarily responsible for the rising prevalence of obesity. Although these findings have been replicated elsewhere, recent commentaries by Bell and Jones call them into question - along with the new class of APC methods. Specifically, Bell and Jones claim that new APC methods do not adequately address model identification and suggest that "solid theory" is often sufficient to remove one of the three temporal dimensions from empirical consideration. They also present a series of simulation models that purportedly show how the HAPC models estimated by Reither etal. (2009) could have produced misleading results. However, these simulation models rest on assumptions that there were no period effects, and associations between period and cohort variables and the outcome were perfectly linear. Those are conditions under which APC models should never be used. Under more tenable assumptions, our own simulations show that HAPC methods perform well, both in recovering the main findings presented by Reither etal. (2009) and the results reported by Bell and Jones. We also respond to critiques about model identification and theoretically-imposed constraints, finding little pragmatic support for such arguments. We conclude by encouraging social scientists to move beyond the debates of the 1970s and toward a deeper appreciation for modern APC methodologies.
Social scientists have recognized the importance of age-period-cohort (APC) models for half a century, but have spent much of this time mired in debates about the feasibility of APC methods. Recently, a new class of APC methods based on modern statistical knowledge has emerged, offering potential solutions. In 2009, Reither, Hauser and Yang used one of these new methods - hierarchical APC (HAPC) modeling - to study how birth cohorts may have contributed to the U.S. obesity epidemic. They found that recent birth cohorts experience higher odds of obesity than their predecessors, but that ubiquitous period-based changes are primarily responsible for the rising prevalence of obesity. Although these findings have been replicated elsewhere, recent commentaries by Bell and Jones call them into question - along with the new class of APC methods. Specifically, Bell and Jones claim that new APC methods do not adequately address model identification and suggest that "solid theory" is often sufficient to remove one of the three temporal dimensions from empirical consideration. They also present a series of simulation models that purportedly show how the HAPC models estimated by Reither et al. (2009) could have produced misleading results. However, these simulation models rest on assumptions that there were no period effects, and associations between period and cohort variables and the outcome were perfectly linear. Those are conditions under which APC models should never be used. Under more tenable assumptions, our own simulations show that HAPC methods perform well, both in recovering the main findings presented by Reither et al. (2009) and the results reported by Bell and Jones. We also respond to critiques about model identification and theoretically-imposed constraints, finding little pragmatic support for such arguments. We conclude by encouraging social scientists to move beyond the debates of the 1970s and toward a deeper appreciation for modern APC methodologies.
This study aims to (1) explore perceptions of property crime at the neighborhood level and their correlates based on a random sample from Guangzhou, China and (2) assess the applicability of collective efficacy theory in contemporary urban China. Since the data used in this study are multilevel and the dependent variable is dichotomous, a generalized hierarchical linear model was used for analysis of the data. This study reveals that both community structural variables (residential stability and poverty) and community process variables (social ties, collective efficacy and semi-formal control) were found to affect individuals' perceptions of neighborhood property crime in Guangzhou. However, communities in Guangzhou are different from those in big cities in the US. This is evidenced by several findings in this study: (1) poorer communities in Guangzhou were not associated with lower levels of formal and informal control; (2) communities with higher levels of residential mobility were neither linked to higher levels of poverty nor disorganization; and (3) the correlation between residential stability and perceived neighborhood property crime was not mediated by community processes. Â© 2013 Springer Science+Business Media Dordrecht.
This article presents the core methodological ideas and empirical assessments of an extended cohort-component approach (known as the "ProFamy model"), and applications to simultaneously project household composition, living arrangements, and population sizes-gender structures at the subnational level in the United States. Comparisons of projections from 1990 to 2000 using this approach with census counts in 2000 for each of the 50 states and Washington, DC show that 68.0 %, 17.0 %, 11.2 %, and 3.8 % of the absolute percentage errors are
BACKGROUND: In genome-wide association studies (GWAS) of human life span, none of the genetic variants has reached the level of genome-wide statistical significance. The roles of such variants in life span regulation remain unclear. DATA AND METHOD: A biodemographic analyses was done of genetic regulation of life span using data on low-significance longevity alleles selected in the earlier GWAS of the original Framingham cohort. RESULTS: Age-specific survival curves considered as functions of the number of longevity alleles exhibit regularities known in demography as "rectangularization" of survival curves. The presence of such pattern confirms observations from experimental studies that regulation of life span involves genes responsible for stress resistance. CONCLUSION: Biodemographic analyses could provide important information about the properties of genes affecting phenotypic traits.
We address comments from the three discussants of our paper, paying particular attention to the properties of our model likely to be of interest in new applications to complex dynamic systems. Â© 2012 Elsevier B.V.
Progress in unraveling the genetic origins of healthy aging is tempered, in part, by a lack of replication of effects, which is often considered a signature of false-positive findings. We convincingly demonstrate that the lack of genetic effects on an aging-related trait can be because of trade-offs in the gene action. We focus on the well-studied apolipoprotein E (APOE) e2/3/4 polymorphism and on lifespan and ages at onset of cardiovascular diseases (CVD) and cancer, using data on 3924 participants of the Framingham Heart Study Offspring cohort. Kaplan-Meier estimates show that the e4 allele carriers live shorter lives than the non-e4 allele carriers (log rank = 0.016). The adverse effect was attributed to the poor survival of the e4 homozygotes, whereas the effect of the common e3/4 genotype was insignificant. The e3/4 genotype, however, was antagonistically associated with onsets of those diseases predisposing to an earlier onset of CVD and a later onset of cancer compared to the non-e4 allele genotypes. This trade-off explains the lack of a significant effect of the e3/4 genotype on survival; adjustment for it in the Cox regression model makes the detrimental effect of the e4 allele highly significant (P = 0.002). This trade-off is likely caused by the lipid-metabolism-related (for CVD) and nonrelated (for cancer) mechanisms. An evolutionary rationale suggests that genetic trade-offs should not be an exception in studies of aging-related traits. Deeper insights into biological mechanisms mediating gene action are critical for understanding the genetic regulation of a healthy lifespan and for personalizing medical care.
The traditional sex morbidity-mortality paradox that females have worse health but better survival than males is based on studies of major health traits. We applied a cumulative deficits approach to study this paradox, selecting 34 minor health deficits consistently measured in the 9th (1964) and 14th (1974) Framingham Heart and 5th (1991-1995) Offspring Study exams focusing on the 55-78 age range. We constructed four deficit indices (DIs) using all 34 deficits as well as subsets of these deficits characterizing males' (DI(M)) and females' (DI(F)) health disadvantages, and no relative sex-disadvantages. The DI(34)-specific age patterns are sex-insensitive within the 55-74 age range. The DI(34), however, tends to selectively increase the risk of death for males. The DI(F)-associated health dimension supports the traditional morbidity paradox, whereas the DI(M)-associated dimension supports the inverse paradox, wherein males have worse health but better survival than females. The traditional paradox became less pronounced, whereas the inverse paradox became more pronounced from the 1960s to the 1990 s. The sex-specific excess in minor health deficits may vary according to particular set of deficits, thus providing evidence for traditional and inverse morbidity paradoxes. The time-trends suggest the presence of a strong exogenous effect modifier affecting the rate of health deterioration and mortality risk.
We evaluated the predictive potential for long-term (24-year) survival and longevity (85+ years) of an index of cumulative deficits (DI) and six physiological indices (pulse pressure, diastolic blood pressure, pulse rate, serum cholesterol, blood glucose, and hematocrit) measured in mid- to late life (44-88 years) for participants of the 9th and 14th Framingham Heart Study examinations. For all ages combined, the DI, pulse pressure, and blood glucose are the strongest determinants of both long-term survival and longevity, contributing cumulatively to their explanation. Diastolic blood pressure and hematocrit are less significant determinants of both of these outcomes. The pulse rate is more relevant to survival, whereas serum cholesterol is more relevant to longevity. Only the DI is a significant predictor of longevity and mortality for each 5-year age group ranging from 45 to 85 years. The DI appears to be a more important determinant of long-term risks of death and longevity than are the physiological indices.
We show that the observed changes in the period tempo of fertility are biased and derive a new formula for adjusting such bias. We present illustrative applications of our proposed method to the cases of the United States and Taiwan. We then describe the relevance of adjustments of observed period fertility tempo for evaluating family planning programs aiming at delaying and reducing births to slow down population growth in developing countries. The work reported in this article also can be regarded as an extension of Ryder's basic translation equation. The extension provides a set of formulas expressing relationships of quantum-tempo between cohorts and periods under specified assumptions.
Our sensitivity analysis shows that the adjusted TFR'(t) using the formula of Bongaarts and Feeney (1998), which assumes an invariant shape for the fertility schedule, usually does not differ significantly from an adjusted TFR"(t) that allows the shape of the fertility schedule to change at a constant annual rate. Because annual changes in the shape of the fertility schedules often are approximately constant except in abnormal conditions, the Bongaarts-Feeney (B-F) method is generally robust for producing reasonable estimates of the adjusted TFR'(t). The adjusted TFR'(t) neither represents any real cohort experiences from the past nor forecasts any future trend. It merely provides an improved reading of the period fertility measure, which reduces the tempo distortion.
Previous methodological research has shown that hidden heterogeneity in hazard rate regression modelsâ€”in the form of systematic differences between sample members in the risk or hazard of making a transition due to unobserved variables not accounted for by the measured covariatesâ€”can produce biased parameter estimates and erroneous inferences. However, few empirical applications of hazard regression do more than pay lip service to the complications of hidden heterogeneity. In part, this is due to the relative inaccessibility of the mathematical apparatus of continuous-time hazard regression methodology with flexible nonparametric specifications on the hidden heterogeneity. This article presents new methods for incorporating nonparametric specifications of hidden heterogeneity into hazard regressions by developing discrete-time Poisson rate/complementary log-log hazard regression models with nonparametric hidden heterogeneity that are analogous to the continuous-time models of Heckman and Singer. Maximum-likelihood estimators and associated hypothesis tests are described. An empirical application to data on criminal careers, which illustrates the utility of models that explicitly incorporate hidden heterogeneity, is presented.
An increment-decrement stochastic-process life table model that continuously mixes measures of functional change is developed to represent age transitions among highly refined disability states interacting simultaneously with mortality. The model is applied to data from the National Long Term Care Surveys of elderly persons in the years 1982 to 1996 to produce active life expectancy estimates based on completed-cohort life tables. At ages 65 and 85, comparisons with extant period estimates for 1990 show that our active life expectancy estimates are larger for both males and females than are extant period estimates based on coarse disability states.
The status of women, which is relative and multidimensional, has an important bearing on any long-term reduction in fertility. In Indian society, where cohabitation and childbearing are socially sanctioned only after marriage, the length of the first-birth interval affects the completed family size by influencing the spacing and childbearing pattern of a family. This study examines the influence of certain aspects of the status of married women--education, employment, role in family decision making, and age at marriage--along with three socioeconomic variables--per capita income of the family, social position of the household, and the caste system--on the duration of the first-birth interval in an urban Hindu society of the north-east Indian state of Assam. The data were analysed by applying life table and hazard regression techniques. The results indicate that a female's age at marriage, education, current age, role in decision making, and the per capita income of the household are the main covariates that strongly influence the length of the first-birth interval of Hindu females of urban Assam. Of all the covariates studied, a female's education appears to be a key mediating factor, through its influence on her probability of employment outside the home and thereby an earned income and on her role in family decision making. Unlike other Indian communities, the effect of the caste system does not have a significant effect on first-birth timing in this urban Hindu society.
Few investigations of the social correlates of depressive symptomatology have addressed variation in the correlates across multiple dimensions of depression scales. We examined the relationships of selected social, clinical, and demographic correlates with four dimensions of the Center for Epidemiologic Studies-Depression (CES-D) scale in 3,401 community-dwelling elders in the Piedmont area of North Carolina. These correlates explained significant variation in somatic complaints and depressed affect; effects of chronic disability and recent negative events were particularly robust. Having a confidant explained reduced symptomatology for all four dimensions, but particularly for low positive affect and interpersonal problems. Positive affect was also buttressed by helping others. These patterns have particular relevance where treatment for depression is divorced from considerations of the social environment of the elderly patient.
This article reports empirical explorations of how well the predictive mean matching method for imputing missing data works for an often problematic variable - income -when income is used as an explanatory variable in a substantive regression model. It is found that the performance of the predictive mean method varies considerably with the predictive power of the imputation regression model and the percentage of cases with missing data on income. In comparisons of single-value with multiple-imputation methods, it also is found that the amount of bias and the loss of precision associated with single-value methods is considerably less than that associated with a weak imputation model. Situations in which using imputed data can lead to seriously biased estimates of regression coefficients (and related statistics) and situations in which the bias is so minimal as to be nonproblematic are identified.
This paper reports microlevel Tobit regression analyses of sociodemographic covariates of the life course accumulation of total household net worth data in eight waves of five distinct panels-spanning over 6 years from late 1984 through early 1991-of the Survey of Income and Program Participation (SIPP). It is found that the quadratic age-wealth relationship predicted by Modigliani's Life Cycle Hypothesis is evident in aggregate age-median wealth profiles as well as in the micro data for households with positive net worth. However, when adult status attainment variables are entered into the regression models either by themselves or in combination with marital/family status variables, the age of household head at which net worth begins to decline is far beyond the typical retirement age. In addition, the traditional criterion variables of sociological status attainment theory-educational attainment, occupational status, and earnings-are found to be positively associated with household net worth, although the net effect of occupational status generally is not statistically significant and the earnings effect is nonlinear. Further, consistent with status attainment theory, householder minority status (black, Hispanic) is negatively associated with the accumulation of net worth. It is found that both single male and single female householder status are negatively associated with the accumulation of household net worth (relative to married couple households) as is the size of the household (measured by the number of children under age 18 present). Separate logistic regression analyses show that households with zero and negative net worth are more likely than households with positive net worth to be black and have low earnings. Higher levels of educational and occupational status attainment reduce the probability of zero net worth but not the probability of negative net worth. Male- and female-headed households and households headed by Hispanics also are more likely to have zero net worth, but not negative net worth. The estimated sociodemographic covariate structures of household net worth are found to exhibit substantial stability across both waves and panels in the SIPP-although effects of the 1990-1991 recession are detectable in estimates for the 1990 panel. Possible applications of the estimated models in demographic projections of household net worth are suggested.
Specifications and moment properties of the univariate Poisson and negative binomial distributions are briefly reviewed and illustrated. Properties and limitations of the corresponding poisson and negative binomial (gamma mixtures of Poissons) regression models are described. It is shown how a misspecification of the mixing distribution of a mixed Poisson model to accommodate hidden heterogeneity ascribable to unobserved variablesâ€”although not affecting the consistency of maximum likelihood estimators of the Poisson mean rate parameter or its regression parameterizationâ€”can lead to inflated t ratios of regression coefficients and associated incorrect inferences. Then the recently developed semiparametric maximum likelihood estimator for regression models composed of arbitrary mixtures of Poisson processes is specified and further developed. It is concluded that the semiparametric mixed Poisson regression model adds considerable flexibility to Poisson-family regression models and provides opportunities for interpretation of empirical patterns not available in the conventional approaches.
There are few studies of the interrelationships among breastfeeding, child spacing, and child mortality in traditional societies that incorporate extensive controls for social and demographic characteristics of the mother and child. In this paper, we investigate the impact of breastfeeding and the length of the preceding birth interval on early child mortality (defined as a death in the first two years of life) using data from a traditional society of India. Multivariate hazards models are used to analyze the data. Most prior analyses related the impact of breastfeeding duration to the duration of child survivability by taking breastfeeding as a fixed covariate. The present study has a methodological focus in the sense that breastfeeding information from retrospective survey data is treated as a time-dependent covariate both as a status variate as well as a duration--with empirical findings compared across the two specifications. The effects of postpartum amenorrhoea and various other demographic and socioeconomic characteristics of mother and child are also studied. The results suggest that breastfeeding duration has a strong impact in reducing the relative risk of early child mortality; but it does not explain the effect of the length of the preceding birth interval on early child mortality.
The traditional preference for sons may be the main hindrance to India's current population policy of two children per family. In this study, the effects of various sociodemographic covariates, particularly sex preference, on the length of the third birth interval are examined for the scheduled caste population in Assam, India. Life table and hazards regression techniques are applied to retrospective sample data. The analysis shows that couples having two surviving sons are less likely to have a third child than those without a surviving son and those with only one surviving son. Age at first marriage, length of preceding birth intervals, age of mother, and household income have strong effects on the length of the third birth interval.
A fundamental limitation of current multistate life table methodology-evident in recent estimates of active life expectancy for the elderly-is the inability to estimate tables from data on small longitudinal panels in the presence of multiple covariates (such as sex, race, and socioeconomic status). This paper presents an approach to such an estimation based on an isomorphism between the structure of the stochastic model underlying a conventional specification of the increment-decrement life table and that of Markov panel regression models for simple state spaces. We argue that Markov panel regression procedures can be used to provide smoothed or graduated group-specific estimates of transition probabilities that are more stable across short age intervals than those computed directly from sample data. We then join these estimates with increment-decrement life table methods to compute group-specific total, active, and dependent life expectancy estimates. To illustrate the methods, we describe an empirical application to the estimation of such life expectancies specific to sex, race, and education (years of school completed) for a longitudinal panel of elderly persons. We find that education extends both total life expectancy and active life expectancy. Education thus may serve as a powerful social protective mechanism delaying the onset of health problems at older ages.
This paper investigates the effects of continued breast-feeding after resumption of menses on fertility, using data from two retrospective surveys in India and single decrement life table and multivariate time-dependent hazards analyses. Breast-feeding even after the return of menses is found to be associated with longer birth intervals. The interaction of breast-feeding duration after resumption of menses and postpartum amenorrhoea has a significant effect on the risk of conception after return of menses.
The length of the first birth interval is one of the strongest and most persistent factors affecting fertility in noncontracepting populations, with longer intervals usually associated with lower fertility. Compared to Western society, the average length of the first birth interval is much longer in traditional Indian society. Yet Indian fertility rates are higher because of either ineffective family planning procedures or deliberate nonuse of birth control and because of the high proportion of the population that is married. Here, we examine the effects of various sociodemographic covariates (with an emphasis on the role of age at marriage) on the length of the first birth interval for two states of India: Assam and Uttar Pradesh. Life table and multivariate hazards modeling techniques are applied to the data. Covariates such as age at marriage, present age of mother, female's occupation, family income, and place of residence have strong effects on the variation of the length of the first birth interval. For each subgroup of females (classified according to different levels of the covariates), the median length of the first birth interval for the Assam (Bengali-speaking) sample is shorter than that of the Uttar Pradesh (Hindi-speaking) sample.
BACKGROUND AND METHODS: Persons of low socioeconomic status are known to have reduced life expectancy. In a study of the relation of socioeconomic status to disability-free or active life expectancy among older persons, we analyzed prospectively gathered data on 2219 blacks and 1838 whites who were 65 years of age or older in the Piedmont region of North Carolina. We defined disability as the inability to perform independently one or more basic functional activities such as walking, bathing, dressing, eating, and using the toilet. For subgroups defined by sex, race, and education, statistical models were used to estimate, for persons at each year of age, the probability of transition from not being disabled or being disabled at base line to not being disabled, being disabled, or having died one year later. These transition probabilities were then entered into increment-decrement life tables to generate estimates of total, active, and disabled life expectancy (with total life expectancy equal to active life expectancy plus disabled life expectancy). RESULTS: Sixty-five-year-old black men had a lower total life expectancy (11.4 years) and active life expectancy (10 years) than white men (total life expectancy, 12.6 years; active life expectancy, 11.2 years), although the differences were reduced after we controlled for education. The estimates for 65-year-old black women (total life expectancy, 18.7 years; active life expectancy, 15.9 years) were similar to those for white women. Black men and women 75 years old and older had higher values for total life expectancy and active life expectancy than whites, and the differences were larger after stratification for education. Education had a substantially stronger relation to total life expectancy and active life expectancy than did race. At the age of 65, those with 12 or more years of education had an active life expectancy that was 2.4 to 3.9 years longer than the values for those with less education in all the four subgroups defined by sex and race. Overall, the subgroups with longer total life expectancy and active life expectancy also lived more years with a disability. CONCLUSIONS: Among older blacks and whites, the level of education, a measure of socioeconomic status, has a greater effect than race on total life expectancy and active life expectancy.
There is considerable variation in the length of the postpartum amenorrhea during which breastfeeding suppresses fertility, both within and between societies. In this paper, we investigate the association between breastfeeding and the resumption of menses and the impact of various biological and social covariates thereon, using data from two retrospective surveys in India. We use both univariate life table and multivariate time-dependent hazards techniques to analyze the data. Most prior investigations related the impact of breastfeeding to postpartum amenorrhea by taking duration of breastfeeding as a fixed covariate. However, breastfeeding beyond the resumption of menstruation cannot affect the duration of menses. Accordingly, the present study has a methodological focus in the sense that breastfeeding is treated as a time-dependent covariate. We found that breastfeeding, age of mother at child's birth, social status, level of income, religion and caste (subcaste), and residential status have significant effects on return of menses in Indian traditional society.
This article demonstrates that disabled life expectancies that are based on conventional multistate life-table methods are significantly underestimated because of the assumption of no changes in functional status between age x and death. We present a new method to correct the bias and apply it to data from a longitudinal survey of about 9,000 oldest-old Chinese aged 80-105 collected in 1998 and 2000. In our application, the age trajectories of disability (activities of daily living--ADL), status-specific death rates, and the probabilities of transitions between ADL states of the oldest-old were investigated for the first time in a developing country. In this article, we report estimates of bias-corrected disabled and active life expectancies of the Chinese oldest-old and demonstrate patterns of large differences associated with initial status, gender, and advances in ages. Using combined information on ADL disabilities and length of having been bedridden before dying, we analyze gender and age patterns of the extent of morbidity before dying among the oldest-old and their implications for debates on the hypothesis of compression of morbidity.
Recent Supreme Court decisions have signaled the need for sound empirical studies of the secondary effects of adult businesses on the surrounding areas for use in conjunction with local zoning restrictions. This study seeks to determine whether a relationship exists between adult erotic dance clubs and negative secondary effects in the form of increased numbers of crimes reported in the areas surrounding the adult businesses, in Charlotte, North Carolina. For each of 20 businesses, a control site (matched on the basis of demographic characteristics related to crime risk) is compared for crime events over the period of three years (1998-2000) using data on crime incidents reported to the police. We find that the presence of an adult nightclub does not increase the number of crime incidents reported in localized areas surrounding the club (defined by circular areas of 500- and 1,000-foot radii) as compared to the number of crime incidents reported in comparable localized areas that do not contain such an adult business. Indeed, the analyses imply the opposite, namely, that the nearby areas surrounding the adult business sites have smaller numbers of reported crime incidents than do corresponding areas surrounding the three control sites studied. These findings are interpreted in terms of the business mandates of profitability and continuity of existence of the businesses.
A better understanding of processes and mechanisms linking human aging with changes in health status and survival requires methods capable of analyzing new data that take into account knowledge about these processes accumulated in the field. In this paper, we describe an approach to analyses of longitudinal data based on the use of stochastic process models of human aging, health, and longevity which allows for incorporating state of the art advances in aging research into the model structure. In particular, the model incorporates the notions of resistance to stresses, adaptive capacity, and "optimal" (normal) physiological states. To capture the effects of exposure to persistent external disturbances, the notions of allostatic adaptation and allostatic load are introduced. These notions facilitate the description and explanation of deviations of individuals' physiological indices from their normal states, which increase the chances of disease development and death. The model provides a convenient conceptual framework for comprehensive systemic analyses of aging-related changes in humans using longitudinal data and linking these changes with genotyping profiles, morbidity, and mortality risks. The model is used for developing new statistical methods for analyzing longitudinal data on aging, health, and longevity.
Journal of the American Statistical Association
Â© Springer Science+Business Media Dordrecht 2015. â€œHow are the kids doingâ€ is a wellbeing question. Many adults might be able to answer the question for their own children or those in their immediate surroundings, and many children may provide information about their own wellbeing or those of other children. However, applied to large populations of children at the national and cross-nationallevels, questions of this genre are more challenging and have stimulated the rapid development of studies of child wellbeing in recent decades. This chapter reviews the objective and subjective approaches to measuring child wellbeing, describes the roots of this field in the social indicators movement of the 1960s and 1970s, outlines the conceptual and methodological development of child wellbeing research, and summarizes empirical findings from several major national and cross-nationalstudies of child wellbeing. The chapter concludes with future directions and needed conceptual and data developments to advance the global monitoring of child wellbeing.
Demography of aging is a subfield of demography that focuses on the older members of a population as well as the processes and consequences of population aging. Research in the demography of aging examines a number of topics, including the state and status of the older population, changes in the numbers, proportionate size, and composition of the older population, demographic forces of fertility, mortality, and migration that bring about these changes, and the effects of these changes on the social, economic, health, and personal well-being of the elderly. Major factors associated with population aging are reviewed. Â© 2008 Copyright Â© 2008 Elsevier Inc. All rights reserved.
This chapter reviews some major directions and findings from recent research on morbidity, disability, and mortality among adults and the elderly. One of the findings is that during the last two decades of the 20th century, the United States continued to exhibit shifts in the age distribution of both overall deaths and deaths from major degenerative diseases toward older ages. This is consistent with predictions from the fourth stage of the epidemiologic transition-the age of delayed degenerative diseases. Research on the dynamics of morbidity and mortality in medical demography shows that the age dependence of mortality risk can be substantially reduced or explained by taking into account several measures of physiological functioning. If the therapies have the effect of helping a larger fraction of elderly cohorts maintain their physiological parameters near the optimal values, this could reduce the age dependence of mortality and further raise life expectancy, which has implications for forecasts of the size and health status of the elderly population that are based on the medical demography model. Research in social demography, epidemiology, and medical sociology has greatly improved knowledge of how social, economic, and lifestyle/behavioral factors affect differentials in morbidity, disability, and mortality by sex, race/ethnicity, and socioeconomic status. Â© 2006 Elsevier Inc. All rights reserved.