Dear researchers, If the direct impact of an independent variable (X) is statistically insignificant but when it is used as a moderating variable(X * M) then the interacting term is statistically significant.
What did these results show?
whether it can be called partial moderation?
How we can interpret such results?
Regards,
Salam
Dear,
The direct relation as you are mentioning is the relationship with no effect of moderator (i.e. when moderator is 0). This shows that in the absence of moderator the relationship is insignificant.
The significant moderating effect is showing that the beta of independent variable in the presence of moderator is significantly different than the beta of intendent variable in absence of moderator (Im assuming that the moderator is binary. For continues variable interpretation can be different). However, the p value of moderating effect do not confirm the significant affect of independent variable in the presence of moderating variable. For such effect you can use margin effects in STATA.
For margin effects see the commands https://www.stata.com/manuals13/rmargins.pdf
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation