Physical activity in early childhood: a five-year longitudinal analysis of patterns and correlates

Table 4 Linear mixed effect multivariate models I and II

Linear mixed intercept only models. Child total PA was included as the dependent variable, nested to each individual code.
Parental activity was imputed as a fixed factor as was; parental education (high education as reference), motor skill (high as reference), preschool care (part time as reference), season (summer as reference vs. winter), sex (boy as reference), BMISDS, risk group ..(Low risk as reference) and Nordic family (Nordic as reference).
Model I is a full model including all variables of interest and additional potential confounding factors.
Model II is a forward stepwise selection model with the threshold set to p-value <0.3 from bivariate models (Additional file 8) this model also takes model I in consideration by including the significant (p-values <0.05) variables from this model. * p-value significant at 0.05 level
^a PA=physical activity
^b Parental education was considered high if at least one parent had an academic education
^c Movement ABC test for motor skills High considered >15^th percentile
^d Full-time preschool care considered ≥30h per week
^e Current season for PA measurement, spring (March-May), summer (June-August), fall (September-November) and winter (December-February)
^f Risk group classified as obesity risk based on parental BMI Low risk parental BMI <25kg/m high risk parental BMI >25kg/m
^g Nordic family was considered if both parents originate from a Nordic country
^h Coef.= unstandardized regression coefficient
ⁱ CI= confidence interval

ISSN: 1479-5868