When I do a moderation analysis, I put the IV (e.g., condition) and moderator in Step 1, and their interaction in Step 2, then I get the Hierarchical Regression model.
When I report the results, for example, the main effect of the condition, should I report the β in Step 1 or in Step 2 ?? (e.g., Step 1 has a β of IV, Step 2 also has a β of IV, their values are different for the same variable). I find them different, and even on different significant level sometimes.
How can we explain the difference? Which one should I adopt? Any rules? Thanks.
Step 2 of the interaction is the moderating effect that is most important.
One common format for reporting this kinds of analysis is to have a separate column for each stage in the hierarchical regression. The first regression in the set would include only the main effect of your moderator, and the second would add the interaction terms to the first set of variables.
The value of your b can change because it is now defined in combination with the interaction term. for example, if your moderator is scored 0 and 1, the b will now be the case where the main effect and interaction effect are both equal to 0.
To test for moderation, the key things you want to examine are the significance of the interaction effect and the change in R-sq.
Also, this kind of analysis is usually reported with metric b's, rather than standardized betas.
David L Morgan Thanks much for your great answers! Before reporting the interaction, I want to first report the main effect. So for the main effect, I should use the b in the first regression (not in the second regression), right?
If using Hayes' Process macro, it seems that both the main and interaction effects are reported directly in the second regression (i.e., this macro does not show the main effect b in the first regression). I guess that's why I usually find different value when using different methods (e.g., the Macro or the normal SPSS)?
In the format I recommend, the first column would show the main effects only, along with the corresponding explained variance.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics