You need to show your models in order for one to know what you are talking about. Thanks, D. Booth

Where there are contradictory results from PROCESS macro and regression analysis, check the computations again. There could be a mistake somewhere.

Hi David, here is the figure. When I use PROCESS macro Model 4 a and b come out positive and statistically significant. c is positive but not significant. When I do the regression analysis between Personally perceived injunctive norms and Intentions to donate the relationship is significant. So I have two questions:

1. Is it possible that the relationship is significant but become insignificant when you add a mediator (in this case Attitudes towards the post).

2. Should I go ahead and do moderation analysis with PROCESS macro Model 1 where I only add X(personally perceived injunctive norms), Y (Intentions to donate) and M (Attitudes towards the post). Thanks for your help!

There are rigid boundries between classifying a variable as a mediator or moderator. If a variable is a cause of a predictor, then this is classified as a mediator. On the other hand, if a variable has a tendency of a situational factor, or contextual factor but not the direct cause of predictor, i.e. classified as a moderator. In your model, you have to revisit the basic quantitative assumptions first such as reliability and validity, look for the model fitness, and also test the path direction of your hypotheses, VIF, multicollinearity, data screening etc. Once you are done with these assumption, this is the stage to go for the regression analysis.

You can also run a mediation analysis using regression. In mediation, you need 3 regressions.

DVˆ = a + b(IV)

DVˆ = c + d(IV)+e(MV)

MVˆ = f + g(IV).

https://www.youtube.com/watch?v=8CjJosnbEU0&t=978s

https://www.youtube.com/watch?v=znhs2dqnruk

https://www.youtube.com/watch?v=9Lk3ObsmZho&t=1006s

I recommend performing the analysis by means of structural equations to check the proposed global theoretical model beyond path coefficients but with global SEM indicators.

Thank you, everyone! I appreciate your help.