Skip to main content
. 2022 Oct 31;38(5):1243–1275. doi: 10.1007/s10680-022-09645-0

Table 3.

Results of discrete-time clog-log models of the transition to teen marriage by granular cohorts

HR of teen union
(1) (2) (3)
Conflict measure * Born in
1974–1977 0.866 0.992 0.969
[0.50,1.49] [0.97,1.00] [0.92,1.02]
1978–1980 0.327** 0.958* 0.846*
[0.16,0.69] [0.92,0.97] [0.74,0.96]
1981–1984 0.668 0.986 0.958
[0.37,1.20] [0.96,1.01] [0.87,1.06]
Overall conflict-affectedness (ref: Not-affected)
Conflict-affected 1.246
[0.73,2.12]
Conflict frequency (events) 1.037*
[1.00,1.07]
Conflict intensity (fatalities per 1,000) 1.107
[0.94,1.30]
District dummies Yes Yes Yes
Year of birth dummies Yes Yes Yes
Controls Yes Yes Yes
σu2 1.181 1.152 1.134
N person-years 44,885 44,885 44,885

Source 2006 AZ-DHS. Notes: Sample consists of women born 1957–1984 (aged 22–49 in 2006), excluding women born 1972–1973. Subjects enter analysis at age 12. Columns represent hazard ratios. 95% confidence intervals are in parentheses. Robust standard errors clustered at the PSU level. The “War-cohort” includes women born 1974–1984. The binary indicator “overall conflict-affectedness” is equal to “1” for IDP/refugee women, non-migrant women residing in Upper-Karabakh and non-displaced women with at least one male member of their family of origin (or mother) who identified as IDP/refugee, and “0” otherwise. All regressions control for duration since start of exposure to the risk of teen marriage (< 5 years, 5–6 years and > 6 years) and rural/urban residence and include a constant not shown. Models are specified with individual-level frailty terms (σu2) and are weighted using provided sampling weights. *p < 0.05, **p < 0.01, ***p < 0.001