In my study , I hypothesized X to have a negative effect on Y and that the moderator W would moderate the negative relationship between X and Y. However, my initial hypothesis of X having a negative effect on Y is not supported. So should I go for moderation effect?
Even if I did the moderation testing the interaction effect between X and W had a significant negative impact on variable Y. So how should I interpret the results?
It is absolutely possible to find a significant moderation (interaction) effect when the overall ("main") effect of X on Y is zero. I would go ahead and plot the interaction effect (i.e., plot the regression lines for the relationship between X and Y for different values of W) to find out what the interaction means. Perhaps the relationship between X and Y is positive for some values of W and negative for other values of W thus leading to a zero (or close to zero) X --> Y relationship overall.
- Badges
- Science topic
- Similar topics
- Psychology