Page 26 - IJPS-11-3

International Journal of
Population Studies Nuptiality patterns among Ethiopian women

Table 5. Decomposition analyses of cumulative incidence of first marriage at age 25
Components of Decomposition Coefficient p-value 95% CI
Overall
1980 – 1984 0.909 0.000 (0.896, 0.921)
1965 – 1969 0.960 0.000 (0.951, 0.969)
Difference −0.051 0.000 (−0.066, −0.035)
Explained −0.014 0.004 (−0.024, −0.005)
Unexplained −0.037 0.000 (−0.053, −0.020
Characteristics Endowment effect Coefficient effect
Coefficient p-value 95% CI Coefficient p-value 95% CI
Rural −0.002 0.296 (−0.006, 0.002) −0.088 0.008 (−0.152, −0.023)
Education −0.007 0.003 (−0.012, −0.002) 0.007 0.485 (−0.012, 0.025)
Regions −0.004 0.003 (−0.006, −0.001) 0.035 0.018 (0.006, 0.065)
Occupation −0.004 0.011 (−0.007, −0.001) 0.001 0.438 (−0.002, 0.005)
Religion 0.000 0.382 (−0.001, 0.000) 0.006 0.053 (0.000, 0.012)
Wealth 0.002 0.083 (0.000, 0.005) −0.002 0.349 (−0.006, 0.002)
Community Education 0.001 0.776 (−0.009, 0.012) 0.006 0.021 (0.001, 0.011)
Community Affluence −0.001 0.514 (−0.003, 0.001) −0.005 0.101 (−0.011, 0.001)
Constant 0.003 0.937 (−0.064, 0.069)
Abbreviation: CI: Confidence interval.

occupations between the two cohorts resulted in a Nearly one-fourth of this reduction (−0.027/−0.102 =
far greater decrease in early marriage (12.6%) than in 26.0%) was attributed to differences in characteristics
education. The increase in the relative share of urban between the two birth cohorts of women. The remainder of
regions also contributed to the reduction in early marriage the reduction (74%) resulted from the effects of variables
to a smaller extent (2.9%) (Table 3). or coefficient effects. A detailed decomposition of the
Regarding the contributions of the coefficient effects, reduction in the cumulative incidence of first marriage
occupation, wealth, and community education played at age 20 showed that women’s education, community
significant roles in the marriage dynamics in Ethiopia. These affluence, and occupation significantly contributed to the
factors contributed to the widening gap in the incidence of reduction. The positive coefficient for residence indicates
marriage between the two birth cohorts rather than reducing that the disparity in cumulative incidence of marriage
it. The positive coefficient value of community education between urban and rural areas was significantly higher
in the detailed decomposition analysis (0.017) reflects its in the former birth cohort than in the more recent cohort
significant role in reducing the incidence of marriage in the when other factors were controlled.
former cohort but a loss of significance in the latter cohort. The inter-cohort gap in women’s educational
Similarly, the positive coefficient for women’s occupations achievement contributed to a 12.7% reduction in the
(0.008) indicates a decline in the significance of women’s overall gap. Differences in women’s occupations between
employment, particularly in the agricultural sector, in the two cohorts resulted in a greater reduction in the
promoting early marriage. Contrarily, for household wealth, incidence of first marriage (10.7%) than differences in
the positive contribution (0.015) reflects an increase in its community affluence (3.9%). In addition, the increase
influence on the incidence of early marriage. This suggests in the relative share of urban regions (urbanization)
that wealthier households in the recent cohort tend to marry contributed significantly to the reduction in cumulative
off their daughters earlier than those in earlier cohorts. The incidence of first marriage, although the magnitude of this
higher significant coefficient value of the constant suggests contribution was small (3.9%) (Table 4).
that certain factors that could reduce the gap were not The differential effects of variables were also evaluated,
included in the model at age 15 (Table 3). and the results indicate that religion and community

The decomposition analysis of the cumulative incidence affluence played significant roles. Regarding the
of first marriage at age 20 revealed a 10.2% reduction. contributions of the coefficient effect, religion significantly

Volume 11 Issue 3 (2025) 20 https://doi.org/10.36922/ijps.3531