Skip to main content

Table 4 Results of correlated component regression analyses for unhealthy dietary intake outcome measures

From: Describing socioeconomic gradients in children’s diets – does the socioeconomic indicator used matter?

   BOYS (n = 275)      GIRLS (n = 353)  
Variable CV predictor counta β Model goodness of fit indicesb Variable CV predictor counta β Model goodness of fit indicesb
Non-core food intake (1 predictor)   Non-core food intake (1 predictor)   
 Mother’s education* 62 −0.111 R2 = 0.012 Mother’s education* 100 −0.119 R2 = 0.012
 Employment 35   R2(CV) = 0.011  Marital status 22   R2(CV) = 0.006
 Marital status 21   SD (CV) = 0.008  Employment 8   SD (CV) = 0.003
 Mother’s age 20     Income 7   
 Income 12     Mother’s occupation 6   
 Child age 10     SEIFA 3   
 SEIFA 10     Child age 2   
 Mother’s occupation 10     Mother’s age 2   
Sweetened drink intake (6 predictors)   Sweetened drink intake (7 predictors)  
 Child age* 90 0.087 R2 = 0.037  Income* 100 −0.085 R2 = 0.069
 Mother’s occupation* 79 0.070 R2(CV) = 0.008  Mother’s education* 100 −0.063 R2(CV) = 0.049
 SEIFA* 78 −0.070 SD (CV) = 0.005  Employment* 100 −0.067 SD (CV) = 0.005
 Mother’s age* 71 −0.051    Mother’s age* 100 −0.066  
 Income* 70 −0.057    SEIFA* 99 −0.064  
 Employment* 66 −0.048    Mother’s occupation* 71 0.049  
 Mother’s education 34     Marital status* 70 −0.045  
 Marital status 32     Child age^ --   
Unhealthy behavioursC (4 predictors)   Unhealthy behavioursC (3 predictors)  
 Child age* 100 0.128 R2 = 0.063  Mother’s education* 100 −0.184 R2 = 0.063
 Mother’s education* 100 −0.121 R2(CV) = 0.037  Mother’s occupation* 100 −0.149 R2(CV) = 0.039
 Mother’s occupation* 90 0.088 SD (CV) = 0.005  Employment* 100 −0.255 SD (CV) = 0.003
 Income* 87 −0.084    Mother’s age 20   
 Employment 3     Marital status 20   
 SEIFA^ --     Child age 19   
 Mother’s age^ --     Income 18   
 Marital status^ --     SEIFA 3   
  1. *Predictor retained in final model.
  2. ^Predictor not retained in any model.
  3. ß = standardised regression coefficient.
  4. aCross-validation predictor count - Represents number of regressions in which predictor appeared. Predictor count of 100 indicates that predictor was present in all 100 regressions. Indicates importance of predictor together with standardised regression coefficient (β).
  5. bModel goodness of fit indices: R2(CV) = cross-validated R2; SD (CV) = Standard deviation for cross-validated R2.
  6. cUnhealthy behaviours: Eat dinner in front of TV, Eat snacks in front of TV, Eat fast food.