Volume 64 - Number 4 - Winter 1998

N. Methods of Research and Analysis Including Models

Studies concerned with demographic methods and with methods from other disciplines that have been applied to demographic data as a whole. Includes mathematical demography and studies on methods of estimation and indirect estimation. Methodological studies and models concerned with one demographic variable, such as migration, are coded under the category concerned with that topic and cross-referenced to this heading. Studies on models used to investigate relationships between demographic variables and for the analysis of empirical data are also coded under this heading.

64:40733 Blanchet, Didier. Nonlinear demographic models and chaotic demo-dynamics. Population: An English Selection, Vol. 10, No. 1, 1998. 139-50 pp. Paris, France. In Eng. with sum. in Fre; Spa.
"Linear models derived from stable population theory still remain a basic reference, but it is clear that non linear models offer much more possibilities to account for actual demographic or demo-economic dynamics. On the other hand, there is one aspect of non linear models which has been probably overemphasized over the last [few] years, which is their ability to generate chaotic dynamics. Up to now, models which have been proposed have some difficulties in deriving such a behavior from plausible assumptions concerning demographic behavior or demo-economic relationships. We illustrate this point of view by discussing two examples, borrowed respectively from Day (1984) and Prskawetz/Feichtinger (1995)."
For the French version of this article, see 64:10759.
Correspondence: D. Blanchet, Ecole National de la Statistique et l'Administration Economique, 3 avenue Pierre Larousse, 92245 Malakoff Cedex, France. Location: Princeton University Library (SPR).

64:40734 Cambois, Emmanuelle; Robine, Jean-Marie; Brouard, Nicolas. Life expectancies applied to specific statuses: a history of indicators and calculation methods. [Les espérances de vie appliquées à des statuts spécifiques: historique des indicateurs et des méthodes de calcul.] Population, Vol. 53, No. 3, May-Jun 1998. 447-76 pp. Paris, France. In Fre. with sum. in Eng; Spa.
"Indicators of life expectancy applied to specific statuses, such as the state of health or professional status, were introduced at the end of the 1930s and are currently the object of renewed interest.... This article presents a historical survey of the methods of calculation and the indicators they have produced. These methods are organized into three groups: single-decrement life table methods, multiple-decrement life table methods, and increment-decrement life table methods. Each of these possesses its particular characteristics, strengths and weaknesses. The choice of method will depend on the aims pursued and the data that is available."
Correspondence: E. Cambois, Equipe INSERM, Démographie et Santé, Centre Val d'Aurelle, Parc Euromédecine, 34298 Montpellier Cedex 5, France. E-mail: cambois@valdorel.fnclcc.fr. Location: Princeton University Library (SPR).

64:40735 Courgeau, Daniel; Baccaïni, Brigitte. Multilevel analysis in the social sciences. Population: An English Selection, Vol. 10, No. 1, 1998. 39-71 pp. Paris, France. In Eng. with sum. in Fre; Spa.
"The multilevel approach can be used to study human behaviour taking into account not only individual characteristics but also the fact that these individuals belong to larger geographical units such as communes and regions. This article gives a detailed critical presentation of the aims and formulations of these models. Attention ranges from the most basic models, which introduce the many different levels in the form of individual and aggregated characteristics, to more complex models which operate with the random characteristics specific to each level, and culminates with multilevel event history models. The article concludes with a more general epistemological reflection on the contribution of these models."
For the French version of this article, see 64:10760.
Correspondence: D. Courgeau, Institut National d'Etudes Démographiques, 133 boulevard Davout, 75980 Paris Cedex 20, France. E-mail: courgeau@ined.fr. Location: Princeton University Library (SPR).

64:40736 Courgeau, Daniel. New methodological approaches in the social sciences. Population: An English Selection, Vol. 10, No. 1, 1998. 240 pp. Institut National d'Etudes Démographiques [INED]: Paris, France. In Eng. with sum. in Fre; Spa.
This special issue of the English-language version of the INED review Population contains eight papers introducing some of the methodological developments that might have relevance to the social sciences.
Selected items will be cited in this or subsequent issues of Population Index.
Correspondence: Institut National d'Etudes Démographiques, 133 boulevard Davout, 75980 Paris Cedex 20, France. E-mail: ined@ined.fr. Location: Princeton University Library (SPR).

64:40737 Ghilagaber, Gebrenegus. Simultaneous log-linear modeling of multiple risks. Department of Statistics Research Report, No. 98-2, 1998. 22 pp. Uppsala Universitet, Department of Statistics: Uppsala, Sweden. In Eng.
"A common approach to the analysis of survival data with multiple causes of failure is to model, separately, each cause-specific hazard rate and estimate the corresponding parameters by single-cause approaches. The literature also contains an approach where two or more cause-specific hazard rates are modeled simultaneously and the parameters are estimated from the same model. In the present paper we provide analytical justification to demonstrate that these two approaches lead to the same estimates of hazard rates. More importantly, we demonstrate that the current approach of simultaneous modeling is based on the assumption of independence among the causes of failure in which case the crude and net hazard-rates are equal. We illustrate the issues through analysis of a data set on family-related events experienced by single men. A log-linear model with a piece-wise constant baseline hazard is used as a model underlying the process of family dynamics." The data are from the 1985 Mail Survey of Swedish Men and concern hazard rates of cohabitation, marriage, and first preunion birth.
Correspondence: Uppsala University, Department of Statistics, P.O. Box 513, 75 120 Uppsala, Sweden. Location: Princeton University Library (SPR).

64:40738 Guérin-Pace, France. Textual statistics: an exploratory tool for the social sciences. Population: An English Selection, Vol. 10, No. 1, 1998. 73-95 pp. Paris, France. In Eng. with sum. in Fre; Spa.
"The value of textual statistics is now widely acknowledged in many branches of the social sciences.... The statistical processing may be of responses to open questions, interviews, texts and even individual itineraries. The procedures are numerous and can be adapted to different kinds of material. The aim in the present article is to give a broad overview of the various applications of these methods, ranging from the most traditional to the most recent, and focusing on both the methodology used and the results obtained."
For the French version of this article, see 64:10762.
Correspondence: F. Guérin-Pace, Institut National d'Etudes Démographiques, 133 boulevard Davout, 75980 Paris Cedex 20, France. E-mail: guerin@ined.fr. Location: Princeton University Library (SPR).

64:40739 Inaba, Hisashi. Some developments in mathematical demography. Jinkogaku Kenkyu/Journal of Population Studies, No. 21, Nov 1997. 7-17 pp. Tokyo, Japan. In Jpn. with sum. in Eng.
This is a general review of developments in the application of mathematical models to the study of demography from the work of Alfred Lotka in stable population theory and life table analysis up to the present day. The author suggests that the two-sex problem remains a major challenge in mathematical demography. The need for more interdisciplinary studies in this field is also recommended, as is a greater concern with the applicability of the developments in mathematical demography to the problems of the real world.
Correspondence: H. Inaba, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113, Japan. Location: Princeton University Library (SPR).

64:40740 Jones, Michael E.; Swerdlow, Anthony J. Bias in the standardized mortality ratio when using general population rates to estimate expected number of deaths. American Journal of Epidemiology, Vol. 148, No. 10, Nov 15, 1998. 1,012-7 pp. Baltimore, Maryland. In Eng.
"Cohort studies often compare the observed number of cases arising in a group under investigation with the number expected to occur on the basis of general population rates. The general population is taken to represent unexposed persons, but it is almost inevitably biased in that it comprises all types of people including exposed ones. To identify circumstances when this bias matters, the authors modeled its effect in relation to the size of the observed standardized mortality ratio (SMR) and the prevalence of exposed individuals in the general population. The authors found that bias may be a major problem, causing substantial underestimation of the true relative risk, when either the prevalence of exposure in the general population or the SMR are large.... Cohort studies of people with common diseases or exposures, or that find large SMRs, when using general population expectations, need to consider the extent of bias from this source."
Correspondence: M. E. Jones, London School of Hygiene and Tropical Medicine, Department of Epidemiology and Population Health, Epidemiological Monitoring Unit, Keppel Street, London WC1E 7HT, England. Location: Princeton University Library (SZ).

64:40741 Lelièvre, Eva; Bonvalet, Catherine; Bry, Xavier. Event history analysis of groups: the findings of an on-going research project. Population: An English Selection, Vol. 10, No. 1, 1998. 11-38 pp. Paris, France. In Eng. with sum. in Fre; Spa.
"The event history approach...is applied using individual longitudinal data. Ideally, however, each individual itinerary would be situated in as broad a context as possible, and the analysis of individual demographic processes would take account of the close or competing events affecting the individual's contact circle. In event history modelling, a shift from the individual to their entourage, for both data collection and analysis, implies a reconsideration of the choice of entities for longitudinal monitoring. A compromise must be reached between conceptual operationality and analytical consistency, in terms of both theory and models: this article presents the formal developments, then the more applied results from this on-going research [using French data]."
For the French version of this article, see 64:10764.
Correspondence: E. Lelièvre, Institut National d'Etudes Démographiques, 133 boulevard Davout, 75980 Paris Cedex 20, France. E-mail: eva@inded.fr. Location: Princeton University Library (SPR).

64:40742 Lotka, Alfred J. Analytical theory of biological populations. Plenum Series on Demographic Methods and Population Analysis, ISBN 0-306-45927-2. LC 98-41063. 1998. xxxi, 220 pp. Plenum Press: New York, New York. In Eng.
This is an English translation of the work that Alfred Lotka published in the two-part Théorie Analytique des Associations Biologiques in 1934 and 1939. It represents Lotka's contributions in the field of demographic analysis. Among the concepts discussed are renewable and stable population theory, observed and intrinsic rates, relations between various demographic measures, and specific problems associated with reproduction and death.
Translated from the French by David P. Smith and Hélène Rossert.
Correspondence: Plenum Press, 233 Spring Street, New York, NY 10013-1578. Location: Princeton University Library (SPR).

64:40743 Mitchell, R.; Martin, D.; Foody, G. M. Unmixing aggregate data: estimating the social composition of enumeration districts. Environment and Planning A, Vol. 30, No. 11, Nov 1998. 1,929-41 pp. London, England. In Eng.
"In this paper the authors address the problem of interpreting and classifying aggregate data sources and draw parallels between tasks commonly encountered in image processing and census analysis. Both of these fields already have a range of standard classification tools which are applied in such situations, but these are hindered by the aggregate nature of the input data. An approach to `unmixing' aggregate data, and thus revealing the nature of the subunit variation masked by aggregation, is introduced. This approach has already shown considerable success in Earth Observation applications, and in this paper the authors present the adaptation and application of the approach to Census small area statistics data for Southampton, Hants, [in England] revealing something of the social composition of Southampton's enumeration districts. The unmixing technique utilises an artificial neural network."
Correspondence: R. Mitchell, City University, Social Statistics Research Unit, Northampton Square, London EC1V 0HB, England. E-mail: rm@ssru.city.ac.uk. Location: Princeton University Library (UES).

64:40744 Otsuka, Tomomi. Evaluation and application of the Ewbank model. Jinkogaku Kenkyu/Journal of Population Studies, No. 22, May 1998. 9-23 pp. Tokyo, Japan. In Jpn. with sum. in Eng.
"The purpose of this paper is to evaluate the four-parameter logit model developed by Ewbank and others, to ascertain the validity and the applicability of this model, and finally to look for a more rational way of estimating the four parameters." The focus is on the application of such models to the analysis of mortality data. The Ewbank model is applied to Japanese life tables. "As a result of the analysis, I found that the Ewbank model captures the mortality pattern of the Japanese, and that the four parameters of the model can be effectively estimated by using Simplex method and Marquardt method. As to the applicability of the Ewbank model, it might be possible to estimate the life table at any point of time from the past to the future by changing the values of the parameters of this model. I ascertained this by altering parameter a. As a result, the feasibility of this model is found to be quite high."
Correspondence: T. Otsuka, Nihon University, Tokyo 102, Japan. Location: Princeton University Library (Gest).

64:40745 Raab, G. M.; Yang, S.; Allardice, G.; Goldberg, D. J.; McMenamin, J. Modelling human immunodeficiency virus infection and acquired immune deficiency syndrome cases in Scotland: data sources, prior information and Bayesian estimation. Journal of the Royal Statistical Society, Series A: Statistics in Society, Vol. 161, No. 3, 1998. 367-84 pp. London, England. In Eng.
"The paper describes the methodology developed to carry out predictions of the acquired immune deficiency syndrome (AIDS) epidemic in Scotland. Information on the human immunodeficiency virus (HIV) epidemic comes from formal case reports of AIDS cases and HIV positive tests, reports from surveillance schemes and from special studies. These sources of information, up to the end of 1994, are reviewed.... A simple model of the HIV epidemic in Scotland is proposed and the information is summarized in terms of this model. Bayesian methodology, using Markov chain Monte Carlo methods, is described and used to predict future cases of AIDS in Scotland and people who will be living with AIDS in the years 1995-1999."
Correspondence: G. M. Raab, Napier University, Applied Statistics Group, Department of Mathematics, 10 Colinton Road, Edinburgh EH10 5DT, Scotland. E-mail: g.raab@napier.ac.uk. Location: Princeton University Library (PF).

64:40746 Ryder, Norman B. Migration and population replacement. Canadian Studies in Population, Vol. 24, No. 1, 1997. 1-26 pp. Edmonton, Canada. In Eng. with sum. in Fre.
The author describes "three stable population models.... The first two are closed to migration, with fertility fixed at the current level and at replacement. The third has fixed amounts of migration, with fertility at the current level. An alternative using fixed migration rates is considered. The models are exemplified with current Canadian data, for the first century and ultimately, paying special attention to age distribution. The role of migrant age is examined. Replacement by migration is an acceptable alternative to replacement by fertility, and much preferable to a closed model with current fertility. Limitations on the usefulness of the exercise for policy are assessed."
Correspondence: N. B. Ryder, Princeton University, Office of Population Research, 21 Prospect Avenue, Princeton, NJ 08544-2091. Location: Princeton University Library (SPR).

64:40747 Smith, Jonathan D. H. Demography and the canonical ensemble. Mathematical Biosciences, Vol. 153, No. 2, Nov 1998. 151-61 pp. New York, New York. In Eng.
"The Gibbs canonical ensemble of statistical mechanics is used to describe the probability distribution of the age classes of mothers of new-borns in an age-structured population. The Malthusian parameter emerges as a Lagrange multiplier corresponding to a generation time constraint, while a new perturbation parameter appears as the Lagrange multiplier corresponding to a maternity constraint. Classical Lotka stability reduces to the unperturbed case of the more general canonical ensemble model. The model is used in a case study of the female (peninsular) Malaysian population of 1970. The Malthusian parameter and perturbation are calculated easily by linear regression. Use of the model identifies an anomaly in the population due to the effects of World War II."
Correspondence: J. D. H. Smith, Iowa State University, Department of Mathematics, 400 Carver Hall, Ames, IA 50011-2064. E-mail: jdsmith@pollux.math.iastate.edu. Location: Princeton University Library (SM).

64:40748 Tan, Wai-Yuan; Xiang, Zhihua. A state space model for the HIV epidemic in homosexual populations and some applications. Mathematical Biosciences, Vol. 152, No. 1, Aug 15, 1998. 29-61 pp. New York, New York. In Eng.
"In this paper we have developed a state space model for the HIV epidemic in homosexual populations which have been divided into subpopulations according to sexual activity levels. In this model, the stochastic dynamic system model is the stochastic model of the HIV epidemic in terms of the chain multinomial model whereas the observation model is a statistical model based on the observed AIDS incidences. This model is applied to the San Francisco homosexual population for estimating the numbers of susceptible people, infective people and AIDS cases and for estimating the probabilities of HIV transmission from infective people to susceptible people given sexual contacts. The results show that the estimated numbers of AIDS incidences trace closely the observed numbers indicating the usefulness of the model."
Correspondence: W.-Y. Tan, University of Memphis, Department of Mathematical Sciences, Memphis, TN 38152. E-mail: tanwy@msuvx1.memst.edu. Location: Princeton University Library (SM).

64:40749 Wiley, Andrea S.; Pike, Ivy L. An alternative method for assessing early mortality in contemporary populations. American Journal of Physical Anthropology, Vol. 107, No. 3, Nov 1998. 315-30 pp. New York, New York. In Eng.
Some problems with the demographic approach to the measurement of infant and child mortality are first reviewed, particularly those associated with lack of data in developing countries. Instead, the authors "offer an alternative to age-based early mortality statistics that makes use of human biologists' interest in, and skill at, assessing human growth and development. Our proposal is to use developmental stages of juveniles instead of relying exclusively on age as the basis for mortality statistics. Death or survival according to a developmental stage (such as crawling or weaning) may provide more accurate data that are also more closely related to the cause of death. Developmental stages have the added advantage of putting infants and children back at the center of the discussion of early mortality by focusing on their activities in relation to their environment. A case study from the Turkana population of Kenya illustrates the use of developmental stages in describing early mortality."
Correspondence: A. S. Wiley, James Madison University, Department of Anthropology and Sociology, MSC-7501, Harrisonburg, VA 22807. E-mail: wileyas@jmu.edu. Location: Princeton University Library (PR).

64:40750 Xia, Hong; Carlin, Bradley P. Spatio-temporal models with errors in covariates: mapping Ohio lung cancer mortality. Statistics in Medicine, Vol. 17, No. 18, Sep 30, 1998. 2,025-43 pp. Chichester, England. In Eng.
"In this paper we blend methods for spatial-temporal mapping with those for handling errors in covariates in a single hierarchical model framework. Estimated posterior distributions for the resulting highly-parameterized models are obtained via Markov chain Monte Carlo (MCMC) methods, which also play a key role in our approach to model evaluation and selection. We apply our approach to a data set of county-specific lung cancer rates in the state of Ohio during the period 1968-1988. Our model uses age-adjusted death rates, and incorporates recent information regarding smoking prevalence, population density, and the socio-eoconomic status of the counties. This information is critical to understanding the role played by a certain depleted uranium fuel processing facility on the elevated lung cancer rates in the counties that neighbour it."
Correspondence: B. P. Carlin, University of Minnesota, School of Public Health, Division of Biostatistics, Box 303 Mayo Memorial Building, Minneapolis, MN 55455-0392. E-mail: brad@muskie.biostat.umn.edu. Location: Princeton University Library (SPR).

Copyright © 1998-1999, Office of Population Research, Princeton University.