Skip to main content

Table 4 Differences between analyses with datasets containing complete data (CD) and imputed data (ID)

From: Skipping breakfast, overconsumption of soft drinks and screen media: longitudinal analysis of the combined influence on weight development in primary schoolchildren

   Skipping breakfast Soft drinks > 1 time per week Screen media > 1h/day
   OR 95% CI OR 95% CI OR 95% CI
Abdominal obesity CD 2.06 (1.23; 3.47)** 1.46 (0.92; 2.32) 2.00 (1.23; 3.23)**
ID 1.87 (1.19; 2.96)** 1.37 (0.92; 2.04) 1.81 (1.19; 2.75)**
Overweight CD 1.71 (1.04; 2.80)* 1.29 (0.84; 1.96) 2.01 (1.33; 3.03)***
ID 1.60 (1.02; 2.50)* 1.38 (0.94; 2.01) 1.76 (1.18; 2.61)**
Obesity CD 0.90 (0.39; 2.07) 1.57 (0.82; 3.03) 1.87 (0.96; 3.67)
ID 1.02 (0.50; 2.07) 1.56 (0.87; 2.80) 1.65 (0.90; 3.01)
  B (SE) p-value B (SE) p-value B (SE) p-value
Changes in WHtRa,b,c CD 0.50 (0.19) 0.007** -0.01 (0.15) 0.966 0.29 (0.16) 0.074
ID 0.51 (0.17) 0.003** -0.07 (0.13) 0.600 0.17 (0.15) 0.240
Changes in weight [kg]b,d CD 0.39 (0.12) 0.001*** -0.08 (0.09) 0.385 0.19 (0.10) 0.054
ID 0.51 (0.17) 0.003** -0.07 (0.13) 0.600 0.17 (0.15) 0.240
Changes in BMI percentilesb,e CD 2.01 (0.90) 0.027* -0.75 (0.70) 0.282 0.70 (0.78) 0.373
ID 2.58 (0.83) 0.002** -1.12 (0.64) 0.083 0.21 (0.72) 0.780
  1. OR odds ratio, CI confidence interval, B (SE) B regression coefficient (standard error), a multiplied by 102 for better interpretability, badjusted for school, migration background, family education level, household income, age, gender, participation in the intervention, and cbaseline WHtR, dbaseline weight, ebaseline BMI percentiles
  2. *** p< 0.001, ** p< 0.01, * p< 0.05