[[A hazard analysis of birth timing in Japan since the 1960s]]

"This paper carried out a multivariate hazard analysis of the first, second and third birth timings by applying [a] log-logistic model. The results are as follows: (1) When wives are employed, the first, second and third birth hazards are all smaller than others....(2) Coresiding with couple's parents is likely to raise the first, second and third birth hazards irrespective of wife's employment status; (3) Wives with tertiary level education are accompanied by a smaller first birth hazard compared with others, while wife's educational attainment does not show any effect on the second and third birth hazards; (4) White-collar husbands are associated with a smaller first birth hazard when their wife is employed; (5) Even after controlling for several background variables, the first, second and third birth timings were accelerated in the cohorts of the late 1960s. This result suggests that the birth timing change in the late 1960s was caused by some heterogeneity which is not considered in this study." An English summary is available from the author.