Table 3.
Multilevel logistic regression models for individual and household/community factors associated with child marriage in Nigeria
| Variable Individual-level variables | Model 0 | Model I | Model II | Model III |
|---|---|---|---|---|
| aOR [95% CI] | aOR [95% CI] | aOR [95% CI] | ||
| Age, y | ||||
| 20 | 1 | 1 | ||
| 21 | 0.90 [0.70–1.15] | 0.92 [0.71–1.18] | ||
| 22 | 0.80* [0.64–0.99] | 0.81* [0.65–1.00] | ||
| 23 | 0.67** [0.53–0.84] | 0.71** [0.57–0.90] | ||
| 24 | 0.52*** [0.41–0.66] | 0.56*** [0.44–0.71] | ||
| Educational level | ||||
| No education | 1 | 1 | ||
| Primary education | 0.86 [0.66–1.11] | 0.94 [0.72–1.22] | ||
| Secondary and above | 0.37*** [0.29–0.46] | 0.48*** [0.38–0.61] | ||
| Partner's level of education | ||||
| No education | 1 | 1 | ||
| Primary | 0.68** [0.51–0.89] | 0.72* [0.55–0.96] | ||
| Secondary and above | 0.48*** [0.38–0.60] | 0.57*** [0.45–0.73] | ||
| Working status | ||||
| Not working | 1 | 1 | ||
| Working | 1.38*** [1.17–1.62] | 1.31** [1.11–1.55] | ||
| Mass media exposure | ||||
| No | 1 | 1 | ||
| Yes | 0.81* [0.68–0.97] | 0.95 [0.79–1.15] | ||
| Religious affiliation | ||||
| Christianity | 1 | 1 | ||
| Islam | 1.22 [0.98–1.53] | 1.20 [0.94–1.55] | ||
| Traditionalist and other | 2.16 [0.93–5.04] | 2.02 [0.84–4.84] | ||
| Ethnicity | ||||
| Hausa | 1 | 1 | ||
| Yoruba | 0.20*** [0.14–0.27] | 0.39*** [0.23–0.65] | ||
| Igbo | 0.25*** [0.18–0.37] | 0.41** [0.23–0.74] | ||
| Other | 0.48*** [0.39–0.59] | 0.52*** [0.40–0.67] | ||
| Household/community level | ||||
| Wealth index | ||||
| Poorest | 1 | 1 | ||
| Poorer | 0.91 [0.71–1.15] | 1.15 [0.90–1.47] | ||
| Middle | 0.60*** [0.47–0.78] | 0.92 [0.70–1.21] | ||
| Richer | 0.39*** [0.29–0.52] | 0.65** [0.48–0.89] | ||
| Richest | 0.17*** [0.11–0.26] | 0.37*** [0.24–0.57] | ||
| Region | ||||
| North Central | 1 | 1 | ||
| North East | 2.48*** [1.93–3.18] | 1.51** [1.16–1.95] | ||
| North West | 3.32*** [2.62–4.21] | 1.43* [1.06–1.93] | ||
| South East | 0.49*** [0.34–0.71] | 0.86 [0.48–1.55] | ||
| South South | 0.97 [0.71–1.33] | 1.32 [0.95–1.83] | ||
| South West | 0.57** [0.41–0.78] | 0.74 [0.47–1.18] | ||
| Place of residence | ||||
| Urban | 1 | 1 | ||
| Rural | 0.94 [0.76–1.17] | 0.95 [0.77–1.18] | ||
| Sex of household head | ||||
| Male | 1 | 1 | ||
| Female | 0.70* [0.53–0.93] | 0.80 [0.60–1.06] | ||
| Community socioeconomic status | ||||
| Low | 1 | 1 | ||
| High | 1.31 [0.99–1.73] | 1.14 [0.86–1.50] | ||
| Community literacy level | ||||
| Low | 1 | 1 | ||
| Medium | 0.62*** [0.47–0.80] | 0.75* [0.58–0.96] | ||
| High | 0.46*** [0.37–0.57] | 0.73** [0.59–0.90] | ||
| Random effects results | ||||
| PSU variance (95% CI) | 1.96 [1.56–2.46] | 0.17 [0.07–0.41] | 0.29 [0.16–0.52] | 0.14 [0.05–0.41] |
| ICC | 0.37 | 0.05 | 0.08 | 0.04 |
| LR test | χ2=333.57, p<0.001 | χ2=5.94, p<0.01 | χ2=17.33, p<0.001 | χ2=4.29, p<0.05 |
| Wald χ2 | Reference | 749.73*** | 620.38*** | 749.21*** |
| Model fitness | ||||
| Log–likelihood | −2609.00 | −2193.74 | −2254.55 | −2152.75 |
| AIC | 5221.99 | 4421.48 | 4541.11 | 4367.51 |
| Number of clusters | 1173 | 1173 | 1173 | 1173 |
Weighted NDHS, 2018
Abbreviations: AIC, Akaike's information criterion; ICC, intra-class correlation; LR, likelihood ratio; PSU, primary sampling unit.
Exponentiated coefficients; 95% CIs in brackets.
*p<0.05; **p<0.01; ***p<0.001.
Model 0 is the null model, a baseline model without any explanatory variable included.
Model I is adjusted for individual-level variables (age, educational level, partner's educational level, working status, mass media exposure, religious affiliation, ethnicity).
Model II is adjusted for household/community level variables (wealth index, region of residence, place of residence, sex of household head, community socioeconomic status and community literacy level).
Model III is the final model adjusted for both individual and household/community level variables.