Hi again :),
I am doing hierarchical regression in SPSS.
In one of my models, I have a non significant Beta for moderator. Shall I delete the moderator from variables and do the regression without moderator again and only report that data?
Thanks a lot for your time,
Sara
Hi Sara, what do you mean with "no effect of the moderator"? Do you mean the first order effect of the moderator itself or the product term. If it is the latter, than yes, you can keep the less complex (linear) model.
Best,
Holger
I mean that when Beta for moderator is nonsignificant, shall I remove it from my variables in regression and report results based on variables that are only significant? Since by removing variables which are not significant, R changes as well. Holger Steinmetz
It is better to keep nonsignificant path(s ) in your model, and it can be your research limitation and suggestion for future studies. However, I think nonsignificant path can be removed from the model to obtain good model fit (if the indices of model fit are not in acceptable level ) .
Best,
Alireza
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression