Over the last four decades in Sri Lanka, age at first marriage for males increased marginally, and for females it rose considerably, but there has been no recent increase in female age at marriage. Among the younger cohorts, estimated ages at which some married indeed show a declining mean age at marriage. Coale-McNeil estimates indicate that there is no real decline in female age at marriage in Sri Lanka. Marriages were delayed as a result of economie hardship or increased mortality to the mid-1970s, but once the overall economy improved after 1977, more marriages took place. Female age at marriage in Sri Lanka remained slightly below 24 years, the age which is anticipated by younger cohorts, and this level is likely to persist for some time.

Users of the six major methods of contraception are compared across a broad range of variables using data from a community sample. Differences between the groups were apparent for a range of socioeconomic and reproductive variables, and current users of the various methods also differed in their past use of contraception. Users of barrier methods fared particularly well. Few differences were observed for current health status or for the sociocultural variables examined, although users of natural methods differed from all others in their religious affiliation and commitment.

Discriminant analysis showed that the most predictive variables distinguishing women who had opted for permanent methods of contraception (female sterilization and vasectomy) were the woman's stated reason for using her current method and her past contraceptive patterns; the inclusion of social, health and reproductive indicators did little to improve the prediction. It is argued that heightened expectations for contraceptive efficacy in the face of increasing concerns about long-term health consequences have contributed to the increased use of permanent methods.

Data from 157 men in Nepal who had vasectomy reversal are analysed. Most sought reversal within 5 years of vasectomy. Half of the men sought reversal because of the death of a male child, and about one-fourth because of the loss of a female child. Re-marriage was the primary reason for only 10% of the men. Those having reversal because of the loss of a male child were generally younger, and for almost half of this group, the age of their last child at the time of their vasectomy was under 2 years. The results suggest that the demand for reversal could be considerably reduced by more careful screening of the potential vasectomy acceptors.

In a survey in Ankole (Uganda) in 1985, 1664 women in samples of the three socioeconomic groups (pastoralists, cultivators and mixed farmers) were asked questions on their reproductive history, and 1029 men questions on educational and wedding costs and their expectations of help from relatives and obligations to help relatives. It is shown that the extended family system allows a great deal of mutual support, while the power of the elders allows them to influence younger married relatives to follow the pronatalist traditions of the society.

This paper analyses the relative importance of demographic and socioeconomic factors with respect to their role in reducing infant mortality in Egypt.

Logit analyses of data from a nationally representative sample of Egyptian households, and for urban and rural households separately, indicate that demographic factors have more effect on infant mortality than socioeconomic factors. The results also show the need to improve housing in urban areas and sewerage systems in rural areas in order to reduce infant mortality. One of the most important policy conclusions, however, concerns the importance of providing a vigorous educational campaign to enlighten mothers and prospective mothers in both rural and urban areas on the positive effects of breast-feeding, longer birth intervals, and fewer children on the survival of infants.

Data from the 1985–86 Japanese census are analysed to explore the determinants of the divorce rates in Japan's forty-seven prefectures, using two theoretical models: (a) the social integration model, which is shown to have a greater utility in predicting Japanese divorce levels than (b), the human capital model. Female emigration patterns play a significant role in affecting the divorce rate. Population increase and net household income are also important predictors of the Japanese divorce rate and urbanization has a great influence in modern Japan. Demographic and aggregate variables such as migration, urbanization, and socioeconomic factors are useful when organized under a social integration model.

About 10% of 3887 ever-married women included in the 1984–85 Malaysian Population and Family Survey revealed that they were influenced by the new population policy to desire more children than they had originally wanted. These women were more likely to be rural Malays from the lower socioeconomic class. Ideal family size was more than four children. Children are desired for economic benefits and emotional support. The natality of the Malays has risen since 1980: their total fertility rate has increased while their contraceptive prevalence rate has dropped sharply. Coupled with a decline in the crude death rate, the present fertility preferences and behaviour of the Malays will render the target of the population policy more attainable than is reflected by the survey data.

Different methods of obtaining information in medical and social research present problems of interpretation for the researcher. However, there are few systematic studies on the extent of the lack of concordance yielded by different methods. This paper uses data from a randomized controlled trial of social support in pregnancy to examine this issue in relation to three methods of data collection—medical records, home interviews and a postal questionnaire—on the following topics: obstetric history; smoking, alcohol use, number of antenatal hospital visits, bleeding and depression in pregnancy; length of labour; baby's sex, birthweight and neonatal health problems; and ethnicity. Considerable discrepancies are found comparing the different data sources. These suggest that mothers may be more reliable sources of information than medical records, and that the anonymity of a postal questionnaire may provide higher estimates than home interviews on some sensitive topics, such as smoking in pregnancy.

Evidence on infant weaning processes provided by field research in Addis Ababa, Ethiopia, and among a refugee population of Eritreans in the Sudan is presented. The study in Addis Ababa, where households were economically disadvantaged but the situation was nutritionally stable, allowed the identification of factors constraining weaning, which includes both the introduction of supplementary foods and the termination of breastfeeding. While the timing of each aspect of weaning was the outcome of the interaction between mother and infant factors, ‘infant-centred’ factors were more important in the commencement and ‘mother-centred’ factors in the completion of weaning. The refugee population had in normal times a similar weaning regime, but under the unstable conditions of flight and life in a refugee camp the process had been significantly altered. Investigation of the factors which have led to altered weaning processes elaborates and confirms the model derived from the Addis Ababa study.

The reproductive pattern of a sample of nuclear families from La Alpujarra (Andalusia, Spain) is analysed. The origin of the wife or consanguinity of the couple does not influence fertility. Variability in number of pregnancies is most closely associated with marriage duration followed by the wife's year of birth. Differences in the number of births are explained more by the number of pregnancies than by the number of miscarriages. The number of survivors to the first birthday is dependent on the number of births and to a lower extent on infant mortality. Data from incomplete families show that reproductive performance of the wife below the age of 45 cannot be accepted as a reliable estimate of complete reproduction.

Analysis of the general fertility rate in the counties of China shows significant associations with infant mortality, illiteracy, and the percentage of the population employed in industry, but not with industrial or agricultural output. Urban counties differ from rural in the importance of the illiteracy rate. The associations in the rural counties vary according to the region of China in which they are located.