Hi,
I am struggling to interpret the size of effect for results from mediation analysis. I understand that we calculate coefficients for mediation analyses by multiplying standardised coefficients for the two paths involved. For example, if we are looking to see how attention mediates the relationship between socio-emotional competence and academic skills, then we will multiple standardised beta coefficients for Path A (From socio-emotional competence to attention) with Path B (From attention to academic skills. If β (Path A) = 0.4 and β (Path B) = .4. Standardised beta for mediation will be 0.16.
I wonder how can I interpret the effect size for this mediation coefficient. If I apply these standards: (Small = .1 - .3; Medium = .3 - .5; Large = > .5); then effect size for mediation above will be small. However, for mediation coefficients to be 0.49 ~ .5, both paths would need to be 0.7, which is not likely in social sciences.
Looking forward to some guidance.
The focus of effect-size measures in mediation analysis concentrates on comparing the magnitudes of different effects in the model—the indirect effect, the direct effect, and the total effect—in order to assess the relative contribution of each (Sobel, 1982).
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Waves