Distribution of birth intervals according to the sampling frame

Estimates of birth intervals are complicated both by their dependence upon marital duration and survival and by the diversity of sampling frames which are realistically feasible or desirable.

Abstract

Observed distributions of the lengths of intervals between successive births (or of the open intervals since the most recent birth) are considerably affected by: (i) the sampling frame, which includes the method of ascertaining persons to be investigated and the kind of data obtained from each individual, (ii) the composition of the population sampled, and (iii) the effects of competing risks of such events as death, marriage, marital dissolution and the ending of previous intervals. Some of these factors also affect distributions of other duration variables such as age (at death or some point in time) marital duration, length of residence, etc. Estimates of birth intervals are complicated both by their dependence upon marital duration and survival and by the diversity of sampling frames which are realistically feasible or desirable.