This paper uses multilevel modelling and Demographic and Health Survey data from five African countries to investigate the relative contributions of compositional and contextual effects of socioeconomic status and place of residence in perpetuating differences in the prevalence of malnutrition among children in Africa. It finds that community clustering of childhood malnutrition is accounted for by contextual effects over and above likely compositional effects, that urban–rural differentials are mainly explained by the socioeconomic status of communities and households, that childhood malnutrition occurs more frequently among children from poorer households and/or poorer communities and that living in deprived communities has an independent effect in some instances. This study also reveals that socioeconomic inequalities in childhood malnutrition are more pronounced in urban centres than in rural areas.

Two sample populations, one refugee and one resident, were studied. The frequencies of consanguineous marriages came out to be 49·8% and 55·4%, respectively, for the refugees and the residents. Caste endogamy was dominant both in the residents and the refugees. The mean coefficient of inbreeding was calculated to be 0·0303 for the refugee population and 0·0332 for the resident population samples. First cousin marriage was the dominant type of marriage in both samples; father’s brother’s daughter (FBD) marriage was more frequent among the refugees while mother’s brother’s daughter (MBD) marriage was more frequent among the residents. Education has no decreasing effect on the incidence of consanguineous marriages. A significant difference in the pattern of marriages in the refugees is observed after the Saur Revolution of 1979.

Female- and male-headed households were compared using data from a Demographic and Health Survey conducted in Tanzania in 1996. Chi-squared tests showed that sex of head of household was highly significantly associated with: residence, household size and composition, radio ownership, having enough food to eat, and age and marital status of head of household. An analysis by the logit regression model showed that female-headed households were more likely than male-headed households to be in rural areas, be small, have fewer men, not have radios and not have enough food to eat. The majority of female heads of households were unmarried and older than male heads of households. The implication is that female-headed households are poorer than male-headed households.

In its recently adopted Global Strategy on Infant and Young Child Feeding, the World Health Assembly called for urgent action in addressing the barriers to optimal feeding practices. This paper examines mothers’ concerns about milk insufficiency as a major contributor to suboptimal infant feeding decisions, using survey data from periurban areas of two Bolivian cities. Mothers in the lowland modernizing city of Santa Cruz were more likely than mothers in the highland traditional city of Cochabamba to express concern about insufficient milk, and also less likely to feed their infants according to international recommendations. Furthermore, perceived milk insufficiency was particularly common among mothers of infants younger than 6 months of age – an age at which infants are particularly vulnerable to the adverse effects of breast-feeding cessation and complementary feeding initiation. The paper concludes with policy, programme and research recommendations to address the critical problem of perceived insufficient breast milk.

The objective of this study was to identify the key determinants of condom use with regular and casual partners among youth in Madagascar. Data stem from a reproductive health survey conducted in October–December 2000 among a representative sample of 2440 youth aged 15–24 living in Toamasina province. Following theoretical models of behaviour change, logistic regression was used to assess the effect of AIDS awareness, personal risk perception, condom access, perceived condom effectiveness, self-efficacy and social support on condom use. Among sexually experienced youth, only about four in ten males and two in ten females have ever used condoms. Fewer than 15% of youth used a condom in last intercourse with their regular partner. Whether youth will try condoms appears to depend largely on the perceived effectiveness of condoms for family planning, access to a nearby condom source, parental support for condom use, and patterns of risky sexual behaviour. Young males’ likelihood of using a condom with a regular partner increases significantly if they perceive condoms to be effective for family planning (OR=11·4; p=0·019). For females, it increases with level of self-efficacy (OR=2·1; p=0·042) and having discussed HIV prevention with someone in the last year (OR=2·8; p=0·022). Among males, condom use with casual partners is significantly higher among those who perceive themselves to be at high risk of sexually transmitted infections (OR=2·3; p=0·014), who believe condoms are effective for family planning (OR=2·8; p=0·048), who have good access to condoms (OR=2·9; p=0·002) and who perceive their parents support condom use (OR=1·7; p=0·048). In conclusion, very few youth in Toamasina are using condoms, highlighting the need to continue and expand adolescent reproductive health interventions. In this low HIV prevalence setting, it is important for these programmes to emphasize that condoms are effective for both pregnancy prevention and STI/HIV prevention.

Surnames provide a useful method to study the structure of human populations for which biological data are not available. The isonymic method has had multiple applications, but difficulties emerge when dealing with groups where extramarital reproduction is common and the sample size is small, and even more so when only paternal surnames are taken into account. Therefore, it could be of interest to retain female surnames, including those of unmarried mothers. This study was carried out using all birth records from an Argentinian population in the colonial period, which was characterized by the presence of different ethno-social groups (Spanish, Indian and ‘Mestizo’ or mixed Spanish–Indian) and various reproductive patterns regarding legitimacy. Coefficient of relationship by isonymy (Ri) kinship matrices between geographical populations were obtained, and the results derived from sets of surnames (paternal, maternal of legitimate and illegitimate children, and all surnames in the registers) compared. The results show similar surname distribution regardless of the set of surnames and group considered. Kinship Ri matrices using paternal surnames, maternal surnames of legitimate children, maternal surnames of illegitimate children, and the set of whole surnames showed the same relationships among populations, indicating a similar pattern for Spanish, Indian and Mixed ethno-social groups. Mantel test correlation between all pairs of matrices was significant in all different ethno-social groups. The results suggest that in populations with high illegitimacy, such as that studied here, it is possible to include maternal surnames, even corresponding to single mothers, in order to consider total reproduction and therefore maximize sample size.

Over the course of the 20th century the sex differential in life expectancy at birth in the industrialized countries has widened considerably in favour of women. Starting in the early 1970s, the beginning of a reversal in the long-term pattern of this differential has been noted in some high-income countries. This study documents a sustained pattern of narrowing of this measure into the later part of the 1990s for six of the populations that comprise the G7 countries: Canada, France, Germany, Italy, England and Wales (as representative of the United Kingdom) and USA. For Japan, a persistence of widening sex differences in survival is noted. The sex differences in life expectancy are decomposed over roughly three decades (early 1970s to late 1990s) from the point of view of four major cause-of-death categories: circulatory diseases, cancers, accidents/violence/suicide, and ‘other’ (residual) causes. In the six countries where the sex gap has narrowed, this has resulted primarily from reduced sex differences in circulatory disease mortality, and secondarily from reduced differences in male and female death rates due to accidents, violence and suicide combined. In some of the countries sex differentials in cancer mortality have been converging lately, and this has also contributed to a narrowing of the difference in life expectancy. In Japan, males have been less successful in reducing their survival disadvantage in relation to Japanese women with regard to circulatory disease and cancer; and in the case of accidents/violence/suicide, male death rates increased during the 1990s. These trends explain the divergent pattern of the sex difference in life expectation in Japan as compared with the other G7 nations.

Indoor air pollution resulting from the combustion of solid fuels has been identified as a major health threat in the developing world. This study examines how the choice of cooking fuel, place of cooking and behavioural risk factors affect respiratory health infections in Accra, Ghana. About 65·3% of respondents use charcoal and 4·2% use unprocessed wood. A total of 241 (25·4%) respondents who cook had had respiratory health symptoms in the two weeks preceding the study. Household socioeconomic status and educational attainment of respondents were found to have a significant impact on respiratory health through their particular influence on the choice of cooking fuel. Households that use wood and charcoal have a high incidence of respiratory health symptoms. The poor are more affected by respiratory health problems due to their heavy dependence on solid fuels as compared with their wealthy counterparts. Households that cook in multiple purpose rooms are more affected by respiratory health problems than those that cook outdoors. There is a positive correlation between the presence of children in the kitchen during cooking and the incidence of respiratory health symptoms among children (r=0·31, p<0·0001). Poverty and lack of education and awareness are the major factors affecting the choice of cooking fuel, place of cooking and respiratory health in Accra.

Do the sexual inclinations of parents influence those of their children? Of 77 adult children of homosexual parents who volunteered for three different investigations, at least 23 (30%) were currently homosexual: twelve (55%) of 22 daughters and three (21%) of fourteen sons of lesbians; five (29%) of seventeen daughters and three (17%) of eighteen sons of gays; none of six sons with both a gay and a lesbian parent. At least 25 (32%) were currently heterosexual. Of the ten with transsexual parents, one of nine daughters was currently lesbian, one was currently heterosexual, and one was transsexual. The son’s sexual preference was not reported. These findings suggest that parents’ sexual inclinations influence their children’s.

Howard’s (2005) claim that male dominance in chess is ‘consistent with the evolutionary psychology view that males predominate at high achievement levels at least partly because of ability differences’ (p. 378) is based on the premise that top level chess skill depends on a high level of IQ and visuospatial abilities. This premise is not supported by empirical evidence. In 1927 Djakow et al. first showed that world-class chess players do not have exceptional intellectual abilities. This finding has subsequently been confirmed many times. Different participation rates, or differences in the amount of practice, motivation and interest for chess in male and female chess players, may provide a better explanation for gender differences in chess performance.

Bilalić and McLeod’s arguments fall short on several grounds. There are excellent logical reasons to expect strong ability/chess expertise links and specific research evidence to date is sparse, with mixed findings. Data are presented from Georgia, which has a high female participation rate in chess, which suggest that differing gender motivation levels and participation rates impact relatively little on chess performance differences at the extreme.