Interpreting moderation

[meri2106]

Hi all,
I'm having trouble interpreting and reporting my current findings. Just to give you a brief idea: I'm looking at the change in a certain measurement between Time 1 and Time 2 as a result of a manipulation and I'm interested in whether mood moderates this change. As my first step I conducted a 2 (experimental condition 1 vs. 2) x 2 (Time 1 vs. Time 2) mixed model and found the interaction to be significant (p=.022). So in one of my experimental groups, the change from Time 1 to Time 2 is significant (also evidenced by a paired t-test). Then I entered mood as a covariate to the model. This time, the interaction is only marginally significant (p=.065). Does this mean that mood moderates my effect? What kind of follow-up analysis should I do to further prove this? How do I report this finding?
I would appreciate any and all help!

 

[css]

Hi meri,
I am not a truly expert, but I hope this is useful.
I feel that the results you obtain after co-varying by mood indicate a possible mediatory (not moderating) effect of this variable. However, to effectively demonstrate this fact, I would suggest using the linear regression method described by Baron & Kenny (Baron RM, Kenny D a. The Moderator-Mediator Variable Distinction in Social The Moderator-Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations. J Pers. 1986;51: 1173–1182. doi:10.1037/0022-3514.51.6.1173).
All the best,

 

[hlsmith]

Was the manipulation randomized?

Another issue you are butting against is that you are just looking at p-values. Perhaps there is just correlation between the terms and effects are not changing just standard errors. You should be the content expert and have the best idea how mood is related to your variable of interest.