How can I interpret this?

General help and assistance with jamovi. Bug reports can be made at our issues page: https://github.com/jamovi/jamovi/issues . (If you're unsure feel free to discuss it here)

by Ceecy » Wed Jun 16, 2021 9:21 am

Dear Jamovi forum,

I'm having some issues trying to interpret these results.

The independent variable is brand image, the dependent variable is brand loyalty, and the moderators are hedonic brand attribute and functional brand attribute.

As I conducted regression analyses, the first regression result showed that brand image does not have a significant effect on brand loyalty. However, after adding the interaction term of functional brand attribute and hedonic brand attribute, the effect of brand image on brand loyalty becomes significant.

I'm having a hard time interpreting this, I've tried to search online, but their theories are very difficult for me to comprehend. Can someone help me explain this issue? Thank you!
Attachments
Picture3.png
Moderating effect of hedonic brand attribute
Picture3.png (60.97 KiB) Viewed 5225 times
Picture2.png
Moderating effect of functional brand attribute
Picture2.png (65.42 KiB) Viewed 5225 times
Picture1.png
Effect of brand image on brand loyalty
Picture1.png (47.29 KiB) Viewed 5225 times
Ceecy
 
Posts: 7
Joined: Sun May 16, 2021 2:23 pm

by Ceecy » Wed Jun 16, 2021 9:51 am

Also, how can I interpret the change of coefficient directions of brand image ? The coefficient was positive, as 0.0856, but turns out to be negative in both cases as -1.820 and -1.171
Ceecy
 
Posts: 7
Joined: Sun May 16, 2021 2:23 pm

by bcollins » Thu Jun 17, 2021 8:34 pm

When you have significant interaction effects, which you do, you need to probe the interaction effects. You can do this in the Estimated Marginal Means options where you can include your IV and your moderator as an interaction term. It will generate a plot such that you can see differences in the relationship of your of your IV on your DV at different levels of the moderator. Traditionally, when you have a significant interaction effect, you don't worry so much about the main effects since you have evidence to suggest the relationship between your IV changes as a function of the moderator. So I wouldn't stress too much about the change in sign associated with brand image when an interaction term is included in the model. It will be easier to interpret when you plot it.

You should see something like this:
example_interaction.png
example_interaction.png (62.67 KiB) Viewed 5170 times


In this case, changes in the relationship between NAQ_bel and EVLN_voice change drastically based on levels of NAQ_pun.
bcollins
 
Posts: 5
Joined: Wed Sep 06, 2017 6:09 pm


Return to Help