Skip to main content

Table 2 Results regarding the mediating role of dietary cost in the association between individual and household educational level and the DHD15-index and DASH score

From: To what extent do dietary costs explain socio-economic differences in dietary behavior?

Independent variable Mediator Dependent variable Total effect
(c-path)
Direct effect
(c’-path)
Indirect effect Proportion mediated
β 95%CI β 95%CI β Bootstrap 95%CI
  Dietary cost (€/d) DHD15- index (score) Individual educational level
Middle vs. low 3.96* 2.93; 5.00 3.78* 2.75; 4.82 0.18* 0.10; 0.28 4.5
High vs. low 9.07* 8.37; 9.77 8.86* 8.16; 9.57 0.20* 0.11; 0.30 2.2
High vs. middle 5.10* 4.01; 6.19 5.07* 4.00; 6.16 0.02 −0.03; 0.09 N/A
  Household educational level
Middle vs. low 2.96* 1.91; 4.01 2.84* 1.79; 3.89 0.12* 0.05; 0.21 4.1
High vs. low 8.25* 7.59; 8.90 8.04* 7.37; 8.70 0.21* 0.11; 0.32 2.5
High vs. middle 5.20* 4.15; 6.24 5.28* 4.24; 6.33 0.09* 0.03; 0.16 1.7
  DASH diet (score) Individual educational level
Middle vs. low 1.00* 0.71; 1.29 0.93* 0.64; 1.22 0.07* 0.04; 0.10 7.0
High vs. low 2.25* 2.06; 2.45 2.17* 1.98; 2.37 0.08* 0.05; 0.11 3.6
High vs. middle 1.25* 0.95; 1.56 1.24* 0.94; 1.54 0.01 −0.01; 0.03 N/A
  Household educational level
Middle vs. low 0.75* 0.46; 1.05 0.71* 0.41; 1.00 0.05* 0.03; 0.08 6.6
High vs. low 2.02* 1.84; 2.21 1.94* 1.76; 2.12 0.08* 0.05; 0.11 4.0
High vs. middle 1.27* 0.98; 1.56 1.23* 0.94; 1.52 0.03* 0.01; 0.06 2.4
  1. Abbreviations: B beta regression coefficient, CI confidence interval, N/A Not Applicable
  2. aSample size for analyses with individual educational level is 9275
  3. bSample size for analyses with household educational level is 9282
  4. *P < 0•05
  5. All analyses were adjusted for age, sex, study center and energy intake
  6. Proportion mediated was not calculated for non-significant indirect effects