Child survivorship estimation: methods and data analysis

"The past 20 years have seen extensive elaboration, refinement, and application of the original Brass method for estimating infant and child mortality from child survivorship data. This experience has confirmed the overall usefulness of the methods beyond question, but it has also shown that...estimates must be analyzed in relation to other relevant information before useful conclusions about the level and trend of mortality can be drawn.... This article aims to illustrate the importance of data analysis through a series of examples, including data for the Eastern Malaysian state of Sarawak, Mexico, Thailand, and Indonesia. Specific maneuvers include plotting completed parity distributions and 'time-plotting' mean numbers of children ever born from successive censuses. A substantive conclusion of general interest is that data for older women are not so widely defective as generally supposed."

Analyses of public use decennial census data with multiply imputed industry and occupation codes

"This paper gives a brief introduction to multiple imputation for handling non-response in surveys. We then describe a recently completed project in which multiple imputation was used to recalibrate industry and occupation codes in 1970 U.S. census public use samples to the 1980 standard. Using analyses of data from the project, we examine the utility of analysing a large data set having imputed values compared with analysing a small data set having true values, and we provide examples of the amount by which variability is underestimated by using just one imputation rather than multiple imputations."

[Confronting the statistics on intra-European migration: toward greater harmonization?]

"Each international migration stream can be enumerated twice: at the time of departure from the country of emigration and at the time of arrival in the country of immigration. In practice, errors in international migration statistics are a major obstacle to the accurate analysis of migration flows. In connection with the process of harmonization set in train by Eurostat several years ago, a method of correcting defective migration data is proposed here, which makes use of this double enumeration. The method is applied to available data in a preliminary analysis. The correction factors presented here through this mathematical approach should turn out to be useful in more than one respect: to illuminate substantial problems of incompatibility of migration data and to follow the progress of the harmonization of those data." (SUMMARY IN ENG)

A study of mortality in China in the 1980s

"This article offers a detailed discussion of Chinese mortality as confirmed by data from the 1982 and 1990 censuses. It points out that, in measuring the actual mortality in China in the 1980s, attention should be paid not only to infant mortality but also to failures in reporting deaths of old people. Based on age-differentiated population as reported by the two censuses, the author computes the population's average life span and re-evaluates the error in age-differentiated mortality as calculated by using mortality data from the censuses."

Demographic event history analysis: a selective review

This is a selective review of the literature concerning demographic event history analysis. "We have attempted to emphasize work that we consider to be particularly important or innovative, to note some of the difficulties that may arise with the use of event history analysis, and to point to several substantive areas where research is still poorly developed."

[New perspectives in life-event history analysis]

"In the last decade, life-event history analysis, also called failure time data analysis or survival analysis, has been widely adopted by demographers. This methodology in demography allows [us] to overcome major hurdles especially when analyzing longitudinal survey data. This paper describes the new perspectives opened to research in that field, and is illustrated by new results and examples of research projects. The authors concentrate on four issues: the analysis of incorrect and imperfect data, the analysis based on more complex data and lastly the study of interaction between phenomena." (SUMMARY IN ENG AND SPA)

Alternative dual system network estimators

"When there are two or more data systems and none of them enumerates the population at an acceptable completeness level, concern about the bias due to underenumeration suggests an estimator which makes joint use of data compiled by the combined imperfect data systems. Conventional dual system estimators are based on the existence of two separate data collection systems. Dual system network estimators assume a main survey and a follow-up quality check survey. The main survey adopts a multiplicity counting rule that combines two mutually exclusive partial counting rules." Three dual system network estimators are presented. "One was previously proposed by Sirken (1979) and is the natural analogue of the conventional dual system estimator. The two other estimators are proposed as potential improvements, although neither of them is the natural analogue of the conventional dual system estimator. The design effects of the three estimators are compared analytically and empirically with one another, and with those of the single system conventional and network estimators." (summary in FRE)

The aging of the population and the financing of local governments

"This paper explores the residential property tax burdens faced by elderly homeowners in the United States. The results demonstrate that the use of annual data in income and property tax liabilities to calculate tax burdens overstates the regressivity of the property tax. The magnitude of the bias created by using annual income is determined by comparing tax burdens calculated from annual data with burdens calculated from tax liabilities and income data covering a 12-year period. The results indicate that while the residential property tax on the non-elderly is approximately proportional, the pattern of burdens on the elderly remains regressive."

How the 1990 Post Enumeration Survey transformed the census adjustment controversy

The author briefly examines the controversy surrounding the 1980 U.S. census undercount, including a lawsuit entered against the U.S. Bureau of the Census by the City of New York. He then describes the Post Enumeration Survey (PES), introduced after the 1990 census "to measure census coverage (i.e. undercounts) by place and by race and ethnicity, which had not previously been possible.... Results of the 1990 PES show that previous claims about the harmful effects of differential undercounts were exaggerated, but they also [show] that the undercounts and their effects are still there."

Miscounting the Spanish origin population in the United States: corrections to the 1970 census and their implications

"The purpose of this...paper is three-fold: (a) to identify the discrepancies in the 1970 census figures [for the Spanish-origin population in the United States], (b) to suggest alternative strategies for correcting the original figures and briefly report on an estimate of the possible 'correct' figures for 1970, and (c) to briefly summarize the implications of the corrected figures for past and present research." The revised figures lower the total Spanish-origin population by about one million and increase the relative importance of specific groups, such as Mexicans, Puerto Ricans, and Cubans. The results also indicate that the rate of growth for the Spanish-origin population from 1970 to 1980 was higher than previously estimated. (summary in FRE, SPA)

Age misreporting and its effects on adult mortality estimates in Latin America

"This article investigates whether misreporting of ages contributes to the apparently low mortality at older ages in Latin America. It compares the size of cohorts enumerated at two censuses, after allowance for intercensal deaths, in 10 intercensal periods in four countries. It finds evidence of very pervasive overstatement of age at advanced ages. Using an empirical age-reporting matrix for Costa Rica, it estimates the bias that such misstatement produces in measured adult mortality levels in that country."

The event history approach in demography

"In the present paper, we shall examine...changes in the field of demography, by following the progression from period analysis to cohort analysis. We shall discuss the hypotheses underlying each of these approaches and show how their distance from actual human behaviour could lead to mistaken or incomplete conclusions. We shall then show how event history analysis can solve these problems and we shall define the hypotheses implied by this new approach. Finally, we shall come back to the complimentarity of the different social sciences."

Homogeneity in Turkish Fertility Survey: implications for cluster size derived from homogeneity in basic survey variables

"This study examines several error statistics derived from the computed variance in basic survey variables of the Turkish Fertility Survey (1978). These are standard error, design factor, design effect and rate of homogeneity. [The] results are presented for the total sample as well as for subclasses. [It is demonstrated that] previous knowledge about the homogeneity and cluster size of main survey variables [will help] to design better surveys in the future. This process can be further extended for other survey variables by imputation." (summary in TUR)

Small area demographic forecasts

The author reviews the literature on small area forecasting with a focus on the work of Robert Schmitt. He finds that "the accuracy of small area forecasts has not increased appreciably over the past four decades, despite methodological progress, increased knowledge gained from the evaluation of forecasts, and more widely available and rich data. A comment made by Schmitt 40 years ago is still largely true today, 'No method is yet known for forecasting the population of small urban areas with a high degree of accuracy'...."

[Sex ratios in Turkey]

"The article [provides] examples of error in the censuses and surveys and questions the reliability of age specific sex ratios in the 1990 Census of Turkey...[and derives] sex ratio at birth from census and survey results as well as from hospital records. Finally the 1990 population is reconstructed by surviving birth cohorts [from] 1900 until 1990 using historical trends of birth rates and cohort, age and time specific survival rates using the East Family of Coale-Demeny Model Life Tables." (SUMMARY IN ENG)

[How consistent are personal data? Analysis of statements on the planning status of pregnancies]

The consistency of responses on the planning status of births is analyzed using data from a longitudinal fertility survey being carried out by the Demographic Institute of the Austrian Academy of Sciences. Interviews were conducted in 1978 and 1981 with a sample of two marriage cohorts; in both sets of interviews a question was asked concerning the reaction to a pregnancy at the time the woman heard about it. "After linking the data for the pregnancies leading to first or second births, a consistency index suggested by Ryder and Westoff...was applied which distinguishes between random and nonrandom consistency. For the first births a proportion of 54.4% identical answers yields a consistency index of 38.0...; for second births a proportion of 53.7% identical answers results in [a] consistency index of 30.1...." Differential consistency is also analyzed according to selected socioeconomic and demographic variables as well as a variable measuring the correspondence of husband's and wife's desired family sizes in the opinion of the husband. (summary in ENG)

The tragedy of the census

The author discusses the debate on how to reduce the undercount error in the 2000 U.S. census. He reviews the history of the census, and considers the impact of the "opposing ¿interest groups' represented in Congress and elsewhere debating methodological issues that can affect the distribution of census-allocated resources [who] tend to favor methodological choices that will increase their resources."

Measuring the quality and duration of contraceptive use: an overview of new approaches

This paper provides an overview of recently developed approaches to the measurement of contraceptive continuation and failure rates, using retrospective survey data rather than the data from clinical trials or program statistics which are more traditionally used. The approaches fall into 2 categories, 1 relying on retrospective contraceptive histories and the other on current status information. The 1st section of the paper discusses the need for new methodologies; conventional applications using clinic and acceptor data are described, some recent results from developed countries presented, and their limitations illustrated. In the 2nd section, each new approach is presented in terms of its data requirements, methods of calculation, and empirical applications. Finally, potential sources of bias and the ability of the approaches to accommodate them are discussed.

Estimation of adult male mortality in Turkey by the widowhood method

"In this paper, the [widowhood] method is applied to Turkish data to estimate adult male mortality. First, the method and its assumptions are briefly discussed. Secondly, the data source and its major characteristics are described and the method is applied; the resulting mortality measures estimated with the method are compared with findings from other sources, and the reliability of the results, the applicability of the method to the Turkish case, and possible biases risked with the application of the method are elaborated in the final section of the paper." (SUMMARY IN TUR)

Aspects of general linear modelling of migration

"This paper investigates the application of general linear modelling principles to analysing migration flows between areas. Particular attention is paid to specifying the form of the regression and error components, and the nature of departures from Poisson randomness. Extensions to take account of spatial and temporal correlation are discussed as well as constrained estimation. The issue of specification bears on the testing of migration theories, and assessing the role migration plays in job and housing markets: the direction and significance of the effects of economic variates on migration depends on the specification of the statistical model. The application is in the context of migration in London and South East England in the 1970s and 1980s."