I conducted a parallel mediation model with two candidate mediators using Hayes’s PROCESS. The indirect pathways for both mediators are statistically significant. However, I need to compare the effect sizes of these mediation effects to see if one of them is significantly larger than the other one (please note that I don’t want to simply report and compare them, but in addition to test if one is significantly larger than the other). Do you know how that could be done? Please let me know if more information might be needed to answer the question. I use R and SPSS. Thanks in advance for your help!
Dear Mr. Mazidi,
First of all, you must check if indirect effects are different, and if they are, I believe that you can only interpret that one effect is larger than the other one if they have the same sign, i.e., + or - (Hayes, 2013).
You can carry out the contrasts in PROCESS clicking on Options and selecting "Pairwaise contrasts of indirect effects/Compare indirect effects".
At the end of the output, you will find the pairwise comparisons between the specific indirect effects, called (C1), (C2)... The legend of these contrasts are at the end of the output -Specific indirect effect constrast definition(s)-. If CI (LLCI-ULCI) of a comparison (e.g., C1) does not contain zero, its indirect effects are different. Now, you must check if indirect effects have the same sign and, if they have, see wich one is larger.
Kind regards.
Take the difference between each chi-square value and the difference between each df, and if the difference exceeds the critical value for that df you can say it is a significantly better fitting model. For example: suppose the mediation model was x2(8) = 47.15, while moderation model was x2(5) = 40.15
Hello,
I have the same question as the poster (i.e., testing whether one specific indirect effect is statistically stronger than a different specific indirect effect). Does anybody happen to know how to figure this out using AMOS?
Thank you so much for your time!
Arash
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling