Dear all,
I used PROCESS model 7 (Hayes, 2018) in order to test a moderated mediation. The attached is the output of my data.
X= dichotomous (1 = yes, 2= no)
Y = continuous (likert scale)
M = continuous (likert scale)
W = multicategorical (1 = low, 2 = medium, 3 = high)
Firstly, in my output, why isn't my dichotomous variable X, incentive, coded using indicator coding as shown in other studies?
Secondly, my multicategorical variable W has 2 values, namely w1 and w2, but I do not know which one to report. Similarly, for the interaction of variable X and W, there are Int_1 and Int_2. For instance, when considering the effect of W on M and index of moderated mediation, I am confused about which to present.
Thank you in advance for your help.
I recommend you to use the free program JAMOVI because it provides more information about p-values of significance and standardized effects for your model of mediation. Your scores report no moderate mediation effect, because the confidence intervals cross 0. Your X*W effect is not significant, w1 and w2 moderate spillover effects at low and medium levels are also not significant. Your model is not methodologically valid because your spillover effect "a" = non-significant, so all other interactive effects are meaningless, and should be applied to the significant spillover effect "b".
Dear Cristian Ramos-Vera ,
Thank you so much for your detailed and timely reply.
I do not understand how you interpreted that "w1 and w2 moderate spillover effects at low and medium levels are also not significant", can you explain to me with my data, please? And what does that indicate?
Moreover, when you say " the model should be applied to the significant spillover effect b", do you mean that I should conduct a simple mediation analysis using Model 4 of Hayes (2018) as the attached photo?
I am doing this data analysis for my master's dissertation and relatively new to this field. Appreciate any help!
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude