Table 2. Variance estimates in child anthropometry and hemoglobin level using four-level random intercepts models, and % explained by a comprehensive set of covariates, National Family Health Survey (NFHS-4).

	Primary Analysis (n = 139,116)			Secondary Analysis (n = 25,605)
	Variance estimates (95% CI)		% Explained (95% CI)	Variance estimates (95% CI)		% Explained (95% CI)
	Model 1	Model 2	Model 1 vs Model 2	Model 1	Model 3	Model 1 vs Model 3
Height-for-age z-scores
Between-population	0.38 (0.37, 0.38)	0.25 (0.24, 0.26)	33.3 (32.6, 34.0)%	0.78 (0.76, 0.80)	0.63 (0.62, 0.64)	20.1 (18.7, 21.6)%
Within-population	2.21 (2.20, 2.22)	2.09 (2.07, 2.11)	5.4 (4.8, 6.1)%	1.86 (1.84, 1.88)	1.76 (1.74, 1.78)	5.2 (4.3, 6.1)%
Weight-for-age z-scores
Between-population	0.27 (0.27, 0.27)	0.16 (0.15, 0.17)	39.4 (38.2, 39.6)%	0.49 (0.48, 0.51)	0.35 (0.34, 0.36)	28.9 (27.0, 30.7)%
Within-population	1.17 (1.16, 1.18)	1.08 (1.07, 1.09)	7.5 (7.3, 7.7)%	0.97 (0.96, 0.98)	0.90 (0.89, 0.91)	7.9 (6.7, 9.1)%
Weight-for-height z-scores
Between-population	0.29 (0.29, 0.29)	0.25 (0.25, 0.25)	12.8 (12.1, 13.5)%	0.59 (0.58, 0.60)	0.53 (0.52, 0.54)	9.9 (9.2, 10.7)%
Within-population	1.55 (1.54, 1.56)	1.52 (1.51, 1.53)	2.0 (1.4, 2.6)%	1.27 (1.26, 1.28)	1.24 (1.23, 1.25)	2.4 (1.9, 2.9)%
Standardized Hemoglobin
Between-population	0.21 (0.20, 0.22)	0.20 (0.19, 0.21)	2.1 (1.5, 2.8)%	0.46 (0.45, 0.47)	0.45 (0.44, 0.46)	1.6 (1.1, 2.0)%
Within-population	0.83 (0.83, 0.83)	0.83 (0.83, 0.83)	0.2 (0.0, 0.4)%	0.61 (0.60, 0.62)	0.61 (0.60, 0.62)	0.5 (0.2, 0.7)%

CI, confidence interval.

Model 1: age and sex.

Model 2: Model 1 + birth order, residence (urban/rural), household wealth, mother’s height, mother’s BMI, mother’s age at marriage, child’s dietary diversity, breastfeeding, drinking water availability, sanitation, safe stool disposal, infectious disease, household air quality, iodized salt, vitamin A supplementation, full vaccination, family planning, skilled birth attendant, antenatal care visits, child diarrhea in past 2 weeks, mother sought care for child with cough/fever.

Model 3: Model 2 + father’s height, father’s BMI, father’s education, father’s age.

% Explained refers to the change in variance estimates from Model 1 to Model 2 (or Model 3). For example, % explained in between-population variation from Model 1 and Model 2 is calculated $\left(\frac{[{\sigma }_{{\text{u}}_0}^2+{\sigma }_{{\text{v}}_0}^2+{\sigma }_{{\text{f}}_0}^2{]}_{\text{model }1}-[{\sigma }_{{\text{u}}_0}^2+{\sigma }_{{\text{v}}_0}^2+{\sigma }_{{\text{f}}_0}^2{]}_{\text{model }2}}{[{\sigma }_{{\text{u}}_0}^2+{\sigma }_{{\text{v}}_0}^2+{\sigma }_{{\text{f}}_0}^2{]}_{\text{model }1}}\right)\times 100$ and % explained in within-population variation from Model 1 and Model 2 is calculated as $(\frac{[{\sigma }_{{\text{e}}_0}^2{]}_{\text{model }1}-[{\sigma }_{{\text{e}}_0}^2{]}_{\text{model }2}}{[{\sigma }_{{\text{e}}_0}^2{]}_{\text{model }1}})\times 100$. 95% confidence intervals for % explained calculated via bootstrapping with 1,000 bootstrap re-samples.