Table 2. Association between emotional subscale (depressive symptomology) and diet quality factor tertiles for boys at baseline and follow-up.

	Baseline emotional subscalea					Follow-up emotional subscale
	Variable	B	SE	p	95% CI	Variable	B	SE	p	95% CI
Unhealthy diet quality factor
	0 –low					0 –low
	1 –medium	0.1	1.1	0.96	-2.4, 2.5	1 –medium	-1.4	1.3	0.30	-4.1, 1.3
	2—high	0.7	0.8	0.38	-1.0, 2.5	2—high	-0.9	1.4	0.56	-3.9, 2.2
Healthy diet quality factor
	0 –low					0 –low
	1 –medium	3.4	0.9	< 0.00	1.4, 5.4	1 –medium	2.3	1.0	0.04	0.2, 4.4
	2—high	7.9	1.0	< 0.00	5.8, 9.9	2—high	5.0	0.9	< 0.00	3.0, 6.9
Covariates
	Age (years)	-0.1	0.3	0.88	-0.7, 0.6	Age (years)	-0.3	0.5	0.51	-1.3, 0.6
	Ethnicity	0.4	0.8	0.58	-1.2, 2.1	Ethnicity	3.1	1.4	0.04	0.2, 6.0
	Condition	1.0	1.1	0.39	-1.4, 3.4	Condition	-2.2	1.0	0.05	-4.3, 0.0
	BMI-z score	0.7	0.2	< 0.00	0.3, 1.2	BMI-z score	0.3	0.3	0.28	-0.3, 0.9
	Walk to school	0.0	0.1	0.99	-0.3, 0.3	Walk to school	0.3	0.1	0.03	0.0, 0.5
	Active at recess	-0.1	0.5	0.84	-1.2, 1.0	Active at recess	-0.5	1.3	0.68	-3.2, 2.1
	Active at lunch	0.2	0.4	0.59	-0.7, 1.1	Active at lunch	2.01	1.0	0.08	-0.3, 4.2
	Active after school	0.4	0.2	0.04	0.0, 0.8	Active after school	0.6	0.4	0.19	-0.3, 1.4

Data from the 2005–2008 Pacific Obesity Prevention in Communities Project.

B: Unstandardized coefficient; SE: Standard error; CI: Confidence interval.

Results significant where p<0.05.

^aMultiple linear regression model adjusted for age (years); ethnicity; study condition; BMI-z score; and PA (active transport to and/or from school; activity levels at recess; activity levels at lunchtime; frequency of after school activities).