Marriage or cohabitation: A competing risks analysis of first-partnership formation among the 1958 British birth cohort

(2000) - Ann Berrington, Ian Diamond

Journal: Journal of the Royal Statistical Society: Series A (Statistics in Society)
Link:: https://onlinelibrary.wiley.com/doi/10.1111/1467-985X.00162
DOI:: 10.1111/1467-985X.00162
Links::
Tags:: #paper #NCDS #Family #Transition
Cite Key:: [@Berrington2000]

Abstract

A discrete time competing risks hazards model is used to analyse entry into ®rst partnership among men and women born in Britain in 1958. Using a life course approach we identify family background and current life experiences which affect the timing and type of ®rst-partnership formation. Education is a key factor in¯uencing the age of entry into ®rst partnership and whether or not the respondent will experience pregnancy before forming the partnership. Religiosity, experience of parental separation and the geographical region of residence are more important in affecting the decision to cohabit rather than to marry directly. The analyses highlight the importance of transitions in other domains such as leaving the parental home in encouraging cohabitation.

Notes

“A discrete time competing risks hazards model is used to analyse entry into ®rst partnership among men and women born in Britain in 1958.” (Berrington and Diamond, 2000, p. 127)

“we identify familybackground and current life experiences which affect the timing and type of ®rst-partnership formation” (Berrington and Diamond, 2000, p. 127)

“Since the early 1970s ®rst-marriage rates have fallen dramatically.” (Berrington and Diamond, 2000, p. 127)

“Premarital cohabitation is nowthe norm in Britain” (Berrington and Diamond, 2000, p. 127)

“The diversi®cation of the life course which has taken place over the last few decades is re ̄ected in the Arelatively low percentage #39% of men and 43% of women) who followed a `traditional' course of direct marriage” (Berrington and Diamond, 2000, p. 128)

“interest is primarily related to household formation, then what is of concern is the time when the respondent ®rst enters a co-residential partnership” (Berrington and Diamond, 2000, p. 129)

“most academics, commentators and policy makers tend to emphasize di€erences between cohabitation and marriage, citing for example the lower reported levels of quality of relationship” (Berrington and Diamond, 2000, p. 129)

“and the greater instability among cohabiting partnerships” (Berrington and Diamond, 2000, p. 129)

“we may consider direct entry into marriage and entry into cohabitation to be two distinct processes or competing risks.” (Berrington and Diamond, 2000, p. 129)

“In this case the occurrence of one event removes the individual from the possibility of experiencing the other Yamaguchi's #1991) `competing risk ideal type II').” (Berrington and Diamond, 2000, p. 129)

“the NCDS refers to a single birth cohort it is impossible to di€erentiate between age and period e€ects Hobcraft et al., 1982).” (Berrington and Diamond, 2000, p. 130)

“Evidence from cross-sectional data suggests that cohabitation is more common in the south and south-east of Britain than in Scotland and the north Haskey and Kiernan, 1989).” (Berrington and Diamond, 2000, p. 132)

“We use a discrete time multinominal logistic hazards model to estimate the log-odds of remaining single, marrying or cohabiting within each month” (Berrington and Diamond, 2000, p. 134)

“Individuals contribute 1-month intervals of exposure from exact age 16 years until the time when they enter a partnership or reach exact age 33 years when they are censored Allison, 1982).” (Berrington and Diamond, 2000, p. 134)

“By including dummy variables representing each 12-month interval we estimate a nonparametric hazard rate which can change with age.” (Berrington and Diamond, 2000, p. 134)

“Advantages of the discrete time approach include the ability to handle tied data, the ease with which time-varying covariates can be included in the model and the ability to test explicitly for proportionality in the e€ects of covariates by modelling interactions between age and the other explanatory variables Allison, 1982).” (Berrington and Diamond, 2000, p. 134)

“Since some of the cohort member's characteristics such as occupational social class and religiosity are only recorded at age 23 years, we analyse entry into ®rst partnership within two separate age ranges: 16±22 and 23±32 years.” (Berrington and Diamond, 2000, p. 135)

“Although there remains in Britain a cultural norm for students to remain unmarried, cohabitation appears to be more compatible with full-time education” (Berrington and Diamond, 2000, p. 148)

“the move away from the parental home, and the shared living arrangements which often accompany participation in higher education, may encourage entry into cohabitation.” (Berrington and Diamond, 2000, p. 148)

“The lower marriage rates among men and women living at age 16 years in the south-east of Britain indicate that the greater proportion of never-married individuals in the south-east noted by Berrington and Murphy #1994) cannot be wholly explained by selective migration of single adults to the region.” (Berrington and Diamond, 2000, p. 149)

“In summary, we have found that men and women born in 1958 who cohabited in their ®rst partnership are a self-selected group who tend to be less religious, more likely to have lived outside a family and more likely to have experienced the separation of their parents.” (Berrington and Diamond, 2000, p. 150)