Table 3. Estimates of coefficients of nutritional variables for Models in Group I.

Model	Variables other than controls	WAZ8	HAZ8	EGRA	Math	PPVT
C	R-squared ( $R_C^2$)	0.314	0.282	0.175	0.345	0.281
0	Norm. birth weight (y0)	0.153*** (0.031)	0.134*** (0.027)	-0.005 (0.032)	0.009 (0.036)	-0.017 (0.025)
	${\mathrm{R}}^2-R_C^2$	0.014	0.011	0.001	0.004	0.002
1W	Weight-for-age at 1 year	0.561*** (0.022)	0.376*** (0.020)	0.119*** (0.022)	0.097*** (0.022)	0.024 (0.022)
	R2	0.489	0.402	0.188	0.357	0.283
1H	Height-for-age at 1 year	0.436*** (0.027)	0.478*** (0.021)	0.106*** (0.021)	0.100*** (0.017)	0.041** (0.018)
	R2	0.436	0.500	0.187	0.359	0.284
2W	Unpredicted weight gain (y1)	0.507*** (0.021)	0.336*** (0.018)	0.113*** (0.021)	0.090*** (0.019)	0.025 (0.020)
	R2	0.482	0.394	0.189	0.357	0.283
2H	Unpredicted height growth (y2)	0.418*** (0.026)	0.461*** (0.022)	0.106*** (0.022)	0.099*** (0.017)	0.043** (0.017)
	R2	0.430	0.493	0.188	0.359	0.285

Notes: The variables on early childhood nutrition and on test scores in EGRA, Math and PPVT are normalized by the sample standard deviations; Number of observations: 1758; the standard errors (in parentheses, underneath the coefficients) are robust to heteroscedasticity;

^***p<0.01,

^**p<0.05,

^*p<0.1.