Table 1.

Best-Fitting Linear and Logistic Regression Models Predicting BMI-Based Outcomes at Age 15

Outcome: BMI Percentile	Beta	SE	t-value	P
BMI percentile trajectory classes, 2 to 13 (reference: low stable class)
High rising	0.43	0.02	22.55	***
Median stable	0.22	0.02	11.75	***
Low-to-high	0.25	0.02	10.38	***
Child is female	0.03	0.01	1.83
Child’s birth weight for age	< −0.01	< 0.01	−1.41
Child’s percent weight increase, 0–15 mos.	−0.16	0.09	−1.77
Estimated maternal BMI, 15 mos.	0.01	< 0.01	2.88	**
Maternal average sensitivity, 6–24 mos.	−0.02	< 0.01	−2.75	**
Child has avoidant attachment style	−0.01	0.02	−0.78
Child has anxious-resistant attachment style	−0.04	0.02	−2.03	*
Income-to-needs ratio, 0–24 mos.	−0.01	< 0.01	−2.32	*
Child’s birth weight for age X Child’s percent weight increase, 0–15 mos.	< 0.01	< 0.01	2.05	*
Outcome: Overweight	OR	95% CI		P
BMI percentile trajectory classes, 2 to 13 (reference: low stable class)
High rising	59.72	27.77	128.43	***
Median stable	4.90	2.20	10.91	***
Low-to-high	8.94	3.75	21.30	***
Child is female	0.61	0.41	0.91	*
Estimated maternal BMI, 24 mos.	1.24	1.09	1.41	***
Maternal years of education	0.92	0.83	1.02
Income-to-needs ratio, 0–24 mos.	0.90	0.82	1.00	*
Outcome: Obesity	OR	95% CI		p
BMI percentile trajectory classes, 2 to 13 (reference: high rising class)1
Median stable	0.08	0.04	0.17	***
Low-to-high	0.22	0.10	0.45	***
Estimated maternal BMI, 15 mos.	1.48	1.28	1.71	***
Maternal years of education	0.88	0.79	0.97	*

Note. SE = standard error; BMI = body mass index; mos. = months. Hg = weight measured in hectograms; OR = odds ratio; CI = confidence interval. Variables with at least marginal significance were retained per stepwise AIC reduction method.

¹No children with obesity were classified in the low stable class and so it was dropped from the analysis and the high rising group was used as the reference.

^*p < .05

^**p <.01

^***p <.001.