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

Table 2.

Results of discrete-time cloglog models of the transition to teen marriage

HR of teen union
(1) (2) (3) (4)
War cohort (1974–1984)* conflict measure 0.659* 0.635* 0.983* 0.942*
[0.45,0.97] [0.43,0.94] [0.97,0.99] [0.88,0.98]
Overall conflict-affectedness (ref: Not-affected)
Conflict-affected 1.001 1.268
[0.72,1.39] [0.74,2.17]
Conflict events 1.037*
[1.00,1.07]
Conflict fatalities per 1,000 1.093
[0.96,1.24]
District dummies No Yes Yes Yes
Year of birth dummies No Yes Yes Yes
Controls No Yes Yes Yes
σu2 1.559 1.127 1.141 1.121
N person-years 44,885 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