Table 3.

Results for multilevel growth curve models for infant WFL z-score trajectories.

		Model 1 Unconditional Means	Model 2 Unconditional Growth	Model 3 Conditional Growth	Model 4 Conditional Growth
	n	n = 666	n = 666	n = 532	n = 532
Fixed Effects
Intercept
Initial Status (Time1)		0.52 (0.04) **	0.38 (0.05) **	0.52 (0.06) **	0.54 (0.06) **
Slope (change in WFL z-scores)
Time1			0	0	0
Time2			0.17 (0.03) **	0.20 (0.07) **	0.16 (0.04) **
Time3			0.25 (0.03) **	0.30 (0.08) **	0.19 (0.04) **
Baby Snacks × Time1
Never	364			0
Sometimes	91			0.11 (0.07)
Often	36			−0.08 (0.10)
Baby Snacks × Time2
Never	76			0
Sometimes	174			−0.12 (0.11)
Often	201			0.07 (0.12)
Baby Snacks × Time3
Never	57			0
Sometimes	132			−0.14 (0.11)
Often	276			0.02 (0.13)
Sweets × Time1
Never	456				.0
Sometimes	24				0.11 (0.12)
Often	11				−0.46 (0.17) *
Sweets × Time2
Never	320				.0
Sometimes	66				−0.09 (0.14)
Often	65				0.48 (0.18) *
Sweets × Time3
Never	245				.0
Sometimes	109				0.04 (0.14)
Often	111				0.53 (0.18) *
Random Effects
Level 1		0.22 (0.01) **	0.21 (0.01) **	0.20 (0.01) **	0.20 (0.05) **
Intercept		0.84 (0.06) **	0.85 (0.06) **	0.79 (0.05) **	0.79 (0.05) **
Model Fit
BIC		3023.7	2978.4	2926.3	2919.0

Table 3 includes parameter estimates with standard errors in parentheses. Models 3 (baby snacks) and 4 (sweets) adjusted for birth weight-for-gestational age z-scores and total weeks any breastfeeding. Baby snacks and sweets were modeled as never [reference], sometimes, and often. Time was modeled as infant age (time₁ = 4–6 months [reference]; time₂ = 7−9 months; time₃ = 10−12 months). Covariance Structure = VC; Estimation Method = REML; Between-within degrees of freedom. Model 1 ICC = 0.79; * p < 0.01, ** p < 0.001.