Hello,

Generally, when you got that, it means there are multicollinearity issues (highly correlated variables).

you must test it with the VIF technique (for example) and remove the concerned variables from the models, or you can use alternative models.

It is possible to obtain a significant overall model in PROCESS SPSS but have insignificant predictor effects. This can occur when the significance of the predictor effects is masked by the presence of other variables in the model. In other words, the predictor may have a weak or non-significant direct effect on the outcome variable, but it may have a significant indirect effect through one or more mediators.

In such cases, it is important to examine the indirect effects of the predictor variable through the proposed mediators. If the indirect effects are significant, it suggests that the predictor variable is affecting the outcome variable through the mediators. However, if the indirect effects are not significant, it suggests that the predictor variable is not affecting the outcome variable through the proposed mediators.

It is also important to consider other factors that may be contributing to the overall model's significance, such as the inclusion of other predictors or interaction effects. These factors may be driving the overall model's significance, even if the predictor variable's direct effect is not significant.

In summary, it is possible to have a significant overall model in PROCESS SPSS but insignificant predictor effects. This can occur when the predictor variable has a weak or non-significant direct effect but has a significant indirect effect through proposed mediators. It is important to examine the indirect effects and consider other factors contributing to the overall model's significance.

Please say more about your explanatory variables, Type and Altru_av. Type sounds like it might be categorical. Is it? If so, how many types are there? Your output shows that it was treated as if it is continuous. That's fine if it is dichotomous and was coded 0/1. But otherwise, it could be problematic.

Assuming it is fine to treat Type as continuous:

• The coefficient for Type shows the effect on Y of increasing Type by one unit when Altru_av = 0.
• The coefficient for Altru_av shows the effect on Y of increasing Altru_av by one unit when Type = 0.

Depending on the nature of the variables, it may not be sensible to examine the simple effects of Type or Altru_av when the other is equal to 0. You may need to examine the simple effects at other values of the other variable. To that end, I believe the PROCESS macro has an option to show the simple effect of one variable at the mean, the mean - 1SD and the mean + 1SD, does it not? Perhaps that will show you more meaningful tests of simple effects. But as I said at the start, it would help others to help you if you said more about the explanatory variables. HTH.

Bruce Weaver: variable type is categorical, yes, with 2 levels: 0=natural condition, 1=artificial condition (dummy coded). Even if I changed the variable to nominal, the results from the regression are the same. Simple slope analysis confirmed insignificant effects for variable Altru_av (the moderator in my model).

Ma'Mon Abu Hammad: I am using model 1 in Process which is regarding moderation (Altru_av is the moderator). Do you suggest that I use another model in order to check whether Altru_av is, in fact, a mediator, instead of a moderator?

I would like to thank all of you for the useful feedback!