Abstract
The advantages of large-scale multi-purpose surveys compared with official divorce records for examining marital breakdown are assessed, and the extent of under-recording of the concept of breakdown in the latter source is estimated. Demographic and socio-economic differentials in breakdown are examined and the former are found to be generally more powerful. A proportional-hazards life-table model is used to establish the impact of childlessness on divorce in a more satisfactory way than hitherto. Among fertile couples, the length of the first birth interval is found to be particularly important as a risk factor influencing breakdown. Controlling for demographic factors, such as age at marriage and fertility status, is shown to modify the observed crude differences between social classes. Housing tenure and personal factors associated with the couple's individual circumstances are more important than social class in explaining marital breakdown, with age at marriage retaining a strongly persisting and remarkably constant effect whatever other variables are included in the analysis.

This publication has 0 references indexed in Scilit: