Table 4

Logistic regression predicting whether parental PC use predicts any PC use by child
All children Sons Daughters
Weekday N OR [95% CI] N OR [95% CI] N OR [95% CI] P for heterogeneitya
Fathers: Model 1b 485 2.2 [1.5 to 3.3] 265 1.7 [1.0 to 2.8] 220 3.0 [1.6 to 5.6] 0.157
Fathers: Model 2c 485 2.2 [1.5 to 3.3] 265 1.7 [1.0 to 2.7] 220 3.1 [1.7 to 5.7] 0.130
Mothers: Model 1 781 1.7 [1.3 to 2.4] 400 1.7 [1.1 to 2.6] 381 1.8 [1.1 to 2.9] 0.852
Mothers: Model 2 781 1.7 [1.2 to 2.4] 400 1.7 [1.1 to 2.6] 381 1.8 [1.1 to 2.8] 0.853
Weekend
Fathers: Model 1 483 1.8 [1.1 to 2.9] 264 1.1 [0.59 to 1.9] 219 3.4 [1.5 to 7.4] 0.029
Fathers: Model 2 483 1.8 [1.1 to 2.9] 264 1.0 [0.57 to 1.9] 219 3.5 [1.6 to 7.7] 0.027
Mothers: Model 1 782 0.78 [0.43 to 1.4] 400 0.57 [0.23 to 1.4] 382 1.0 [0.48 to 2.2] 0.317
Mothers: Model 2 782 0.77 [0.44 to 1.4] 400 0.56 [0.22 to 1.4] 382 0.98 [0.47 to 2.0] 0.339

aTesting that associations are different in daughters and by sons; tested by adding an interaction term (parent exposure variable*child gender) into the regression model.

bModel 1: Unadjusted association.

cModel 2: Adjusted for child’s BMI-z score parent’s age, parent’s BMI, and household IMD.

Jago et al.

Jago et al. International Journal of Behavioral Nutrition and Physical Activity 2014 11:54   doi:10.1186/1479-5868-11-54

Open Data