## Table 3 |
|||||||||||

Associations of objectively measured MVPA with self reported Variety and Frequency |
|||||||||||

Variety and Frequency in separate models |
Variety and Frequency in the same model |
||||||||||

Outcome |
Exposure |
Exp. β |
95% CI |
P-value |
% variance explained by exposure |
Exp. β |
95% CI |
P-value |
% variance explained by exposure |
||

MVPA | Variety | 1.04 | 1.02 | 1.06 | <0.001 | 1.18 | 1.01 | 0.98 | 1.04 | 0.402 | 1.36 |

Frequency | 1.04 | 1.02 | 1.06 | <0.001 | 1.32 | 1.03 | 1.00 | 1.06 | 0.051 |

Variety and Frequency were transformed to z-scores so effect sizes are directly comparable.

Associations were tested using multilevel linear regression.

Models were adjusted for age, sex, parental education level and age standardised body mass index (BMI).

MVPA; moderate-to-vigorous physical activity was log transformed for analysis. The β coefficients and 95% confidence intervals (95% CI) were then exponentiated for interpretation. Exponentiated β (Exp. β) and 95% CI represents a ratio of geometric means and can be interpreted in the same way as odds ratios. i.e. Exp. β and 95% CI <1 equates to a negative association whereas Exp. β and 95% CI >1 equates to a positive association.

P-value for significance of the association of Variety or Frequency with physical activity.

Brooke * et al.*

Brooke * et al.* *BMC Public Health* 2013 **13**:761 doi:10.1186/1471-2458-13-761