Table 2. Predictors of energy drink consumption (multivariate analysis^$).

	Whole sample (n = 678)			Boys (n = 314)			Girls (n = 364)
	%* (n)	OR [95% CI]	P	%* (n)	OR [95% CI]	P	%* (n)	OR [95% CI]	P
Age (years)†		1.070 [0.946–1.211]	.281		1.291 [1.083–1.539]	.004		0.866 [0.718–1.044]	.866
Sex
Female	37.1 (135)	1.00
Male	48.4 (152)	1.707 [1.221–2.386]	.002
Ethnicity
European	33.8 (77)	1.00		39.4 (41)	1.00		29.0 (36)	1.00
Melanesian	47.2 (195)	1.066 [0.718–1.582]	.751	52.4 (100)	1.051 [0.585–1.885]	.868	42.8 (95)	1.124 [0.650–1.945]	.676
Polynesian	40.5 (15)	1.501 [0.687–3.277]	.308	57.9 (11)	2.749 [0.912–8.282]	.072	22.2 (4)	0.749 [0.210–2.668]	.656
SES
High	29.5 (62)	1.00		39.6 (40)	1.00		20.2 (22)	1.00
Intermediate	42.6 (72)	1.306 [0.820–2.078]	.261	48.3 (42)	1.159 [0.598–2.244]	.663	36.6 (30)	1.667 [0.835–3.330]	.148
Low	51.2 (153)	1.807 [1.171–2.788]	.008	55.6 (70)	1.318 [0.688–2.526]	.406	48.0 (83)	2.324 [1.260–4.290]	.007
Residence
Urban	29.8 (53)	1.00		37.4 (34)	1.00		21.8 (19)	1.00
Rural	46.8 (234)	1.498 [0.939–2.390]	.090	52.9 (118)	1.865 [0.973–3.575]	.060	41.9 (116)	1.121 [0.544–2.308]	.757
Energy drink perception
Bad for health	32.9 (170)	1.00		37.8 (88)	1.00		28.9 (82)	1.00
No effect on health	70.3 (52)	4.247 [3.003–8.127]	< .001	74.4 (29)	3.920 [1.772–8.673]	< .001	65.7 (23)	4.052 [1.873–8.766]	< .001
Good for health	74.7 (65)	5.119 [2.452–7.355]	< .001	83.3 (35)	7.582 [3.151–18.243]	< .001	66.7 (30)	3.884 [1.936–7.794	< .001
Weight status
Overweight	44.7 (106)	1.00		50.0 (52)	1.00		40.6 (54)	1.00
Not overweight	41.0 (181)	0.893 [0.625–1.277]	.537	47.6 (100)	0.969 [0.567–1.656]	.909	35.1 (81)	0.804 [0.492–1.314]	.384

OR: odds ratio; CI: confidence interval; SES: socioeconomic status

*Indicates the percentage of adolescents in each group consuming energy drinks.

†Entered into the model as a continuous variable.

$Variables in the models are: age (years), sex, ethnicity, SES, residence, energy drink perception and weight status category.