Statistically, I saw that moderator has no or inconsistent correlation with IV and DV in thesis.
Actually in my situation, I measured baseline IV, DV and moderator once(each level2).
And I measured IV and DV 3 times(each level1). It is nested multilevel design.
I am conducting to test cross-level interaction effect with HLM.
IV was Group mean centered and moderator was Grand mean centered.
I found significant cross-level interaction effect.
But the problem is, I found that baseline IV and DV has high correlation with moderator.
Therefore, I can predict that moderator has high correlation with IV and DV.
In this situation, can I tell that there is really significant cross-level interaction effect?
If you are saying that your data doesn't support your initial hypothesis that sometimes happens. But notice baseline IVB having high correlation with the moderator is moderation ie, interaction . I suggest you review this by looking at this link: https://www.bing.com/search?FORM=U528DF&PC=U528&q=moderation+in+regression+form and especially consider this one:
- Moderation Analysis -- Advanced Statistics using R https://advstats.psychstat.org/book/moderation/index.phpA moderation effect indicates the regression slopes are different for different groups. Therefore, if we plot the regression line for each group, they should interact at certain point. Such a plot is called an interaction plot. To get the plot, we first calculate the intercept and slope for each level of the moderator. For this example, we have
- This should help. Best wishes, D. Booth
I applied the moderating effect in my research. You can read it and you will have a good experience.
https://www.emerald.com/insight/content/doi/10.1108/JHTI-09-2018-0058/full/html
- Badges
- Science topic