Truncation effect in closed and open birth interval data.

Abstract

A numerical investigation using a flexible simulation model to establish interval analysis as an index for changing natality patterns. Such an index should reflect parity distribution, the age at which women start reproduction, and the spacing of their births. The simulated statistical results illustrate the truncation effect that reflects a negative correlation between parity and the length of closed and open intervals in a birth or marriage cohort. Truncation is related to the duration of marriage at survey, but this duration interacts with other assumptions. Holding duration constant does not ensure that the data on intervals will reflect postulated changes in the distributions. For complete birth orders, this analysis does reflect patterns of child spacing. However, it ignores changes in the parity distribution, whether produced by deliberate limitation of family size or by the onset of secondary sterility. This difficulty is not overcome by life table analysis except under highly restrictive assumptions. It is doubtful whether the current emphasis on securing such data is justified. Further investigation is needed to provide a better basis for the definition and analysis of interval data if they are to be used.