TY  - JOUR
A1  - Lehrer, Evelyn
A1  - Son, Yeon-jeong
T1  - Women’s age at first marriage and marital instability in the United States: Differences by race and ethnicity
Y1  - 2017/08/02
JF  - Demographic Research
JO  - Demographic Research
SN  - 1435-9871
SP  - 229
EP  - 250
DO  - 10.4054/DemRes.2017.37.9
VL  - 37
IS  - 9
UR  - https://www.demographic-research.org/volumes/vol37/9/
L1  - https://www.demographic-research.org/volumes/vol37/9/37-9.pdf
L2  - https://www.demographic-research.org/volumes/vol37/9/37-9.pdf
L3  - https://www.demographic-research.org/volumes/vol37/9/files/readme.37-9.txt
L3  - https://www.demographic-research.org/volumes/vol37/9/files/demographic-research.37-9.zip
N2  - Background: Numerous studies have shown that the age at which women enter first marriage is a major factor in marital instability, but possible differences by race/ethnicity have not been examined to date. 

Objective: Our aim was to study whether the association between women’s age at first marriage and marital instability varies by race/ethnicity. 

Methods: We used data from the 2006–2010 National Survey of Family Growth to examine differences by race/ethnicity in the shape of the curve relating women’s age at entry into first marriage to marital instability. 

Results: We found that for non-Hispanic white women, the probability of dissolution falls with age up to ages 30–32 and thereafter the curve flattens out. For black women, marital instability decreases with age only up to ages 24–26. For Hispanic women, marital instability falls from age ≤20 to 21–23 and then the curve flattens out; beyond ages 30–32 the curve turns upward. 

Conclusions: We suggest explanations for these patterns, based in part on differentials in the associations of age at marriage with education and nonmarital fertility. For white women, but not for their black and Hispanic counterparts, delayed entry into marriage is associated with a small increase in nonmarital fertility and a pronounced increase in education. 

Contribution: The common practice in studies with US data of conducting pooled analyses – with simple controls for black, Hispanic, and other – can lead to misleading conclusions. Our findings underscore the desirability of conducting separate analyses by race/ethnicity wherever possible. 

ER  -