Thank you for this video! helped me a lot but the problem I am facing is that in some cases, my interaction is significant, the model is significant but one of the other coefficients (lets say zbooks or zattend) is not significant. How should I interpret that?
That's fine - the individual main effects can be non-significant and the interaction be significant - that happens just means that the main effect was not important in context of the interaction.
Hi Erin, thanks for your very useful videos. You have done a video on moderated moderation called "3 Way Interactions in SPSS with PROCESS". would you be able to do a similar one on the moderated moderation model using JASP/Excel?
I'll add it to the request list - you'd have to do a lot of breaking down for each interaction! (it would very similar to this video but with multiple new high and low columns based on what was significant).
In mediation, you are mostly interested in the paths, so you would use the t-tests associated with those paths. You can include the overall F tests if you want, but it would be redundant to the t-tests for the a and c paths.
You explain 100% of the variance in the survival of humanity. You're Great!
Thank you for the kind words!
Thank you for this video! helped me a lot but the problem I am facing is that in some cases, my interaction is significant, the model is significant but one of the other coefficients (lets say zbooks or zattend) is not significant. How should I interpret that?
That's fine - the individual main effects can be non-significant and the interaction be significant - that happens just means that the main effect was not important in context of the interaction.
Hi Erin, thanks for your very useful videos. You have done a video on moderated moderation called "3 Way Interactions in SPSS with PROCESS". would you be able to do a similar one on the moderated moderation model using JASP/Excel?
I'll add it to the request list - you'd have to do a lot of breaking down for each interaction! (it would very similar to this video but with multiple new high and low columns based on what was significant).
If the interaction is not significant, does the rest of it matter?
Correct - in a real project, I would not follow up the interaction with the simple slopes if the interaction coefficient was not significant.
Thank you! I have based my whole dissertation on moderation regression and this video has been great!
Hello, would you please clarify why we should you t test instead of f tests in this case?
In mediation, you are mostly interested in the paths, so you would use the t-tests associated with those paths. You can include the overall F tests if you want, but it would be redundant to the t-tests for the a and c paths.