Table 3.
Predictors of size underestimation (feeling ‘too light’) among normal-weight adolescents (multivariate analysisc)
| Whole sample (n = 429) | Boys (n = 212) | Girls (n =217) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| %a (n) | OR [95% CI] | p | %a (n) | OR [95% CI] | p | %a (n) | OR [95% CI] | p | |
| Age (years)b | - | 0.892 [0.730–1.089] | .261 | - | 0.813 [0.621–1.065] | .133 | - | 1.068 [0.778–1.468] | .683 |
| Gender | |||||||||
| Female | 19.4 (42) | 1.00 | - | - | - | - | - | - | |
| Male | 17.9 (17) | 0.810 [0.481–1.363] | .427 | - | - | - | - | - | - |
| Ethnicity | |||||||||
| European | 12.7 (21) | 1.00 | 15.4 (12) | 1.00 | 10.3 (9) | 1.00 | |||
| Melanesian | 22.5 (57) | 1.682 [0.906–3.124] | .100 | 18.9 (24) | 1.390 [0.568–3.398] | .470 | 26.2 (33) | 1.989 [0.823–4.805] | .127 |
| Polynesian | 18.2 (2) | 1.590 [0.287–8.814] | .596 | 28.6 (2) | 1.748 [0.258–11.836] | .567 | - | nd | - |
| SES | |||||||||
| Higher | 11.0 (16) | 1.00 | 13.2 (10) | 1.00 | 8.7 (6) | 1.00 | |||
| Intermediate | 21.5 (23) | 2.122 [1.020–4.416] | .044 | 25.0 (15) | 2.342 [0.888–6.174] | .085 | 17.0 (8) | 2.017 [0.610–6.674) | .250 |
| Lower | 23.2 (41) | 2.137 [1.063–4.296] | .033 | 17.1 (13) | 1.674 [0.588–4.760] | .334 | 27.7 (28) | 2.723 [0.976–7.596] | .056 |
| Residence | |||||||||
| Urban | 12.0 (1) | 1.00 | 18.5 (12) | 1.00 | 5.0 (3) | 1.00 | |||
| Rural | 21.4 (65) | 1.360 [0.638–2.901] | .426 | 17.7 (26) | 0.603 [0.224–1.623] | .317 | 24.8 (39) | 5.264 [1.287–21.532] | .021 |
| Weight statusd | |||||||||
| Upper-normal weight | 13.4 (38) | 1.00 | 10.4 (14) | 1.00 | 16.0 (24) | 1.00 | |||
| Lower-normal weight | 29.0 (42) | 3.442 [2.025–5.851] | <.001 | 30.8 (24) | 4.043 [1.893–8.634] | <.001 | 26.9 (18) | 3.306 [1.489–7.338] | .003 |
OR odds ratio, CI confidence interval, SES socioeconomic status
aIndicates the percentage of overweight participants in each group perceiving themselves to be too light
bEntered into the model as a continuous variable
cVariables in the models are: age (years), gender, ethnicity, SES, residence and weight status category
nd: not determined due to small size of the subgroup
d‘Lower normal weight’ and ‘upper normal weight’ category were subcategories of the normal weight category