Recent trends in infant mortality in China [abstract].

The China 1988 Two per Thousand Survey data on infant mortality are evaluated for validity as far back as the 1950s. Existing data available from official statistics of registration and census materials have been found to be inadequate for the coverage between the 1950s and 1980. The conclusion is that the 2 per 1000 data are valid for the study of the pace distribution and causes of the decline of infant mortality in China between the 1950s and 1988. The sample of more than 500000 women included 2 ever-married women aged 15-57 years for every 1000 population and their pregnancy histories by province. The data are considered suitable for multivariate analysis and time-series analysis. Validity is examine for omitted events and misplaced events and age selectivity. Data analysis shows that infant mortality declined at a different pace across provinces. The result of the Great Leap Forward and the associated famine was great excess infant mortality in general which occurred in only some provinces. The age pattern of infant mortality shows an upward concave pattern which also varies across provinces. The sex differences in mortality vary across provinces and the potential meaning is discussed. Analysis indicates that the declines in infant mortality occurred before the early 1970s and consequently before substantial economic development. These economic conditions and infant mortality decline have occurred in China and in Cuba.