Hello, I could tell you with my experience that people with lower CI also have the ability to decide and make decisions by themselves. Therefore, mediation in these cases is efficient and significant.

Greetings,

Pilar Munuera

Pilar, I'm sorry, I think you misunderstood the question: it was about statistics, not about the psychological constructs

when the lower limit and upper limit of CI has zero e.g in your case LLCI is -0611, PLCI is. 0004, it refers to insignificant effect. You can refer this paper of mine too.

Article "Empirical Investigation of Moderating and Mediating Variabl...

Thank you, Jeevan. I mean, how to interpret it meaningfully with regards of other results: there is simple partial mediation of these constructs, there is simple moderation effect and there are conditional indirect effects - all significant with p-values and CIs. Does it mean that moderation so to say "breaks" mediation? Or is it something else?

if the conditional indirect is significant, it shows that moderation effect is being mediated

well, then how to understand all this altogether? here is the whole analysis:

Outcome: anx

Model Summary

R R-sq MSE F df1 df2 p

,3787 ,1434 1,3309 10,9390 3,0000 196,0000 ,0000

Model

                coeff        se        t            p          LLCI ULCI

constant 1,4323 1,4426 ,9929 ,3220 -1,4127 4,2773

pp ,1403 ,3823 ,3669 ,7141 -,6137 ,8942

fear 1,0208 ,3789 2,6942 ,0077 ,2736 1,7680

int_1 -,1998 ,0978 -2,0418 ,0425 -,3927 -,0068

Product terms key:

int_1 pp X fear

******************** DIRECT AND INDIRECT EFFECTS *************************

Conditional indirect effect(s) of X on Y at values of the moderator(s):

Mediator

             fear Effect Boot SE BootLLCI BootULCI

pp 2,7686 -,0518 ,0250 -,1123 -,0144

pp 4,0507 -,0840 ,0302 -,1588 -,0354

pp 5,3329 -,1162 ,0444 -,2316 -,0476

Values for quantitative moderators are the mean and plus/minus one SD from mean.

Values for dichotomous moderators are the two values of the moderator.

******************** INDEX OF MODERATED MEDIATION ************************

Mediator

       Index SE(Boot) BootLLCI BootULCI

pp -,0251 ,0153 -,0611 ,0004

Hi Ksenia, As I see the Output and as previously mentioned, there won't be Moderated mediation because othe Confidence intervals (BootLLCI & Boot ULCI) of the Index of Moderated Mediation CROSS the zero, although this result is almost marginal and should be interpreted cautiously. In the original Hayes Paper it is further explained the functioning of the index.

Article An Index and Test of Linear Moderated Mediation

thank you, Juan!

Hi Ksenia, could you find an interpretation of the situation that the conditional indirect effect is being significant at all the levels o the moderator (-1, mean and +1)? Could you please send me the source of reference? 

Hi, Payal, i can't send you the reference, but as far as i understood, it means that the mediation is present, but moderation is not - there is no difference between the slopes, the effect is present at all levels

Presence of moderation cannot be analyzed from Index of moderated mediation. For this, you have to see the following three.

1- X=>Y (if significant, the move further). Two steps.

2- Moderated => Y (If significant, then move further)

3- Mod * X => Y (If significant, then go below).

Decision: If

If 1,2 and 3 are significant, the Partial moderation. 

If 1 and 2 are significant but not 3, No moderation. 

If 1 and 3 are significant but not 2, Full moderation. 

Hope this will help.

Sir Khurram , can you please illustrate the above with example.

If you wish to add a better precision to your analysis besides the level of significance and interval you can perform through SEM analysis using AMOS that will allow you to check the fit of your proposed general model beyond the regression coefficients.

https://www.youtube.com/watch?v=9Lk3ObsmZho&t=1006s

You can consult this from

https://statisticalhorizons.com/wp-content/uploads/MMA-Sample-Materials.pdf