The study set out to document child care practices in as many different countries and cultures as possible with the aim of providing baseline child care data and stimulating new hypotheses to explain persisting differences in sudden infant death syndrome (SIDS) rates between countries. The protocol, piloted in four countries in 1992, was distributed to 80 potential centres in 1995. Data from 19 centres were received. This paper describes the demographic characteristics of the data from the different centres. Comparison showed significant differences for a number of variables including mean age of completion of the study, response rate, mean gestation, mean birth weight, method of delivery and incidence of admission to neonatal intensive care units. High caesarean section rates identified in the Chinese samples (44 and 40%) were unexpected and have important public health implications. This finding warrants further study but may be related to China's one child policy. We consider that international comparison of child care practice is possible using standardised data collection methods that also allow some individual variation according to local circumstances. However, in view of the heterogeneity of the samples, it will be important to avoid over-interpreting differences identified and to view any differences within the qualitative context of each individual sample. Provided there is acknowledgement of limitations, such ecological studies have potential to produce useful information especially for hypothesis generation.