Table 4. Estimates of coefficients of nutritional variables for Models in Group II.

Model	Variables other than controls	WAZ8	HAZ8	EGRA	Math	PPVT
3W	Birth weight (y0)	0.141*** (0.025)	0.126*** (0.022)	-0.007 (0.033)	0.006 (0.035)	-0.018 (0.025)
	Unpredicted weight gain (y1)	0.505*** (0.020)	0.334*** (0.018)	0.113*** (0.022)	0.090*** (0.019)	0.025 (0.020)
	R2	0.489	0.403	0.189	0.357	0.284
3H	Birth weight (y0)	0.140*** (0.027)	0.120*** (0.022)	-0.008 (0.032)	0.005 (0.035)	-0.019 (0.025)
	Unpredicted height growth (y2)	0.415*** (0.026)	0.459*** (0.022)	0.107*** (0.022)	0.099*** (0.017)	0.043** (0.017)
	R2	0.436	0.500	0.188	0.359	0.285
4HW	Birth weight (y0)	0.139*** (0.025)	0.119*** (0.022)	-0.008 (0.033)	0.005 (0.035)	-0.019 (0.025)
	Unpredicted height growth (y2)	0.415*** (0.022)	0.459*** (0.022)	0.107*** (0.022)	0.099*** (0.017)	0.043** (0.017)
	Conditional unpredicted WAZ1 (y3)	0.303*** (0.023)	0.026 (0.023)	0.055*** (0.020)	0.030* (0.017)	-0.007 (0.018)
	R2	0.494	0.501	0.191	0.360	0.285
4WH	Birth weight (y0)	0.139*** (0.025)	0.119*** (0.022)	-0.008 (0.033)	0.005 (0.035)	-0.019 (0.025)
	Unpredicted weight gain (y1)	0.506*** (0.020)	0.335*** (0.015)	0.113*** (0.021)	0.090*** (0.019)	0.025 (0.020)
	Conditional unpredicted HAZ1 (y4)	0.093*** (0.025)	0.316*** (0.027)	0.040* (0.020)	0.051*** (0.015)	0.036** (0.015)
	R2	0.494	0.501	0.191	0.360	0.285

Notes: The variables on early childhood nutrition and ontest 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.