Table 4.
Model testing mediation of the association between obesity and perceived weight discrimination by psychological wellbeing (see Figure 2)
Coeff. | SE | p* | Bootstrap 95% CI | Effect ratio | |
---|---|---|---|---|---|
Total effect (path c) | 0.112 | 0.006 | <.001 | - | - |
Direct effect (path c’) | 0.106 | 0.006 | <.001 | - | - |
Indirect effect (via mediators) | 0.0063 | 0.001 | <.001 | [0.004; 0.009] | 0.056 |
Indirect effect (via quality of life) | 0.0059 | 0.001 | <.001 | [0.004; 0.009] | 0.053 |
Indirect effect (via life satisfaction) | −0.0009 | 0.0005 | .072 | [−0.002; −0.0001] | −0.008 |
Indirect effect (via depressive symptoms) | 0.0013 | 0.0007 | .067 | [0.0001; 0.003] | 0.012 |
Model uses z-scores for all psychological wellbeing variables.
Model is adjusted for age, sex, and wealth.
Coeff. = coefficient; SE = standard error; CI = confidence interval.
p values shown for indirect effects are derived from the Sobel test for consistency with total and direct effects, however bootstrap 95% confidence intervals provide a more robust indication of significant mediation (see Method for more details).