PIP: During the summer of 1983, a provincewide representative sample of all ever-married women and single women with children, under the age of 60, was interviewed to examine variations in family size in Northern Ireland. The valuation list of domestic dwellings was used as the sample frame and a clustered sample was drawn from this using the 526 electoral wards as individual clusters. Wards were stratified according to socioeconomic structure, religion, location within Northern Ireland, and whether they were predominatly rural or urban. Sampling within selected clusters was on a systematic basis and proportional to population size. For women currently married, 1 in 5 of the husbands was interviewed using a shorter questionnaire focusing on attitudes to fertility. Of the 3914 houseolds where contact was made and which contained an eligible female, there were 2997 successful interviews with females, giving an effective response rate of 77%. An additional 392 interviews were obtained from husbands. The most persistent fertility differential in Northern Ireland going back at least to the beginning of the 20th century has been that between Protestants and Roman Catholics. According to these preliminary findings, this remains the case. The average number of children born alive to ever-married Roman Catholic females was 3.24 in 1983, compared with 2.29 children for the corresponding group of non-Catholic women. Yet, the survey relates to an instant during a peirod of considerable flux in Roman Catholic fertliity, and comparison of these 1983 data with those collected at the 1971 census of population shows that Roman Catholic family size declined by 11% over the intervening 12 years, although this is marginally reduced after standardization for changes in the age structure of ever-married women. Non-Catholic family size, by contrast, remained virtually static during the same time period. The pace of change has been more pronounced in Belfast and its suburbs. Data are not yet available from the survey to chart these processes by marriage cohort, but some pointers can be obtained from the tabulation of family size by denomination and age of mother. With the exception of the under-20 age group, the average size of Roman Catholic families was consistently and substantially larger than that of the corresponding non-Catholic age group. The absolute difference in terms of mean numbers of children widened steadily with increasing age. These preliminary data also show that there is still a strong geographical dimension to the religious differential in fertility. Average family size was larger in rural than urban areas but moreso for Roman Catholics (16% larger) than for non-Catholics (9% larger). Although differences in family size between Roman Catholics and non-Catholics still prevail, there has been a considerable degree of overall convergence since 1971.