Your search found 2 Results
Voorburg, Netherlands, Netherlands Interuniversity Demographic Institute, 1985 Sep. ix, 56 p. (Working Paper of the N.I.D.I. No. 63)The objective of this report is to introduce the available techniques of life history analysis to study the data collected by the national migration surveys and to demonstrate the relevance of such techniques to provide more insight into to the problems addressed by the ESCAP migration and urbanization project. The 2nd section of the report introduces the basic concepts, with special reference to migration, and contains a simple example. Section 3 deals with 3 further issues which may arise while modeling migration histories: the alternative definitions of the state-space; the definition of the time dependence; and heterogeneity considerations and ways of dealing with heterogeneity for discrete state stochastic models of migration. The 4th section focuses on some major problems which may arise while estimating stochastic models of migration histories with ESCAP migration his. 2 issues are emphasized in this section: problems with the measurement of the timing of the events and issues related to using the information on the covariates of migration. Continous time stochastic models provide a powerful means of modeling event sequences. Migration histories consist of information on the times and the characteristics of migration experienced by individuals. More conventional ways of modeling such data are the dummy variables regression, the logit regression, or aggregation of the data are to form contingency tables and application of the log-linear models. Continous time event history models easily be generalized to incorporate complex designs of the state space, which express the moves between residences, and to provide detailed and cross nationally comparable information on the patterns of time dependence. Additionally, they are based on the estimation techniques which do not require unrealistic assumptions. These models aim at identifying a dynamic process that underlies the observed data. Estimated parameters of these models provide a description of the time dependence and also provide quantitative information about the effects of exogenous variables on the phenomenon of interest. The dependent variable of the continous time event history models is usually the instantaneous transition rate which is not directly observable. The estimated coefficients of the exogenous variables may be interpreted the same as the coefficients of a regression model, except that they usually have a multiplicative relation with the dependent variable. Once models of fundamental parameters of the underlying process are designed and estimated, many implications of such a process may be derived.
In: Population prospects in developing countries: structure and dynamics, edited by Atsushi Otomo, Haruo Sagaza, and Yasuko Hayase. Tokyo, Japan, Institute of Developing Economies, 1985. 39-57, 326-7. (I.D.E. Statistical Data Series No. 46)A comparative study on mortality trends of developing countries was conducted by making use of UN projections of mortality measures. These mortality measures projected by the UN were used to observe future prospects of general mortality trends in selected countries. Under several research projects of the Institute of Developing Economies (IDE), some attempts were made to analyze recent trends of cause-specific mortality covering several selected countries. Estimates of future changes in cause-specific mortality may be considered useful to supply basic information needed for social and economic development planning of a country. Trends of mortality changes in the 1950s and 1960s were characterized in many countries by a rapid decline. Such a declining trend of mortality was brought about initially by a successful control of infectious and parasitic diseases accompanied by improvements in living conditions of the people in general. Thus the mortality of less developed countries that had been affected to a greater extent directly by infectious and parasitic diseases could be improved more drastically at such a stage. After the 1970s the pace of decline in mortality slowed down gradually to a considerable extent all over the world but was more prominent among more developed countries. In most countries mentioned in this discussion, regardless of whether they are more or less developed, the crude death rate is expected to reach the lowest level within a few decades. In many instances of developing countries, the crude death rate is assumed to reach such a minimum level in and around 2000. After reaching the lowest level, the crude death rate will turn to increase in varying degrees. Such a rise in crude death rate does not imply deterioration of mortality conditions. The crude death rate is often affected by the sex-age composition of the population. In contrast to the crude death rate, in most countries selected here, the expectation of life at birth is expected to expand steadily towards the future during the whole duration of this projection. An analytical observation was made on the cause-specific mortality for 10 selected countries covering the period from 1970 to the latest year for which basic data were available on the 8th (1965) revision of International Statistical Classification. Future prospects of cause structure of deaths will be very much influenced by proccesses effected by policy making and planning, and projections of cause-specific mortality should be made with an aim toward providing useful information for policy making and planning for national development.