Mediation midwives a statistical relationship and interaction, in order words the relationship or effect of the independent variable with the dependent is observed on the mediator rather than directly on the dependent variable. However if the independent variable can still affect or influence the dependent variable in any way then, it is a partial mediation but if it cannot, it is a full mediation. This differs from Moderation anyway.

Best Mohammad Irshad

Hello Mohammad,

As Chukwuemeka E. Etodike indicates, proposing that variable W mediates the relationship between scores X and Y suggests that there should be little or no direct influence of X on Y (which can be tested), and that there should be a non-zero linkage between X and W, and also between W and Y. That is, the indirect influence of X on Y (via W) should be sufficiently large, relative to the direct influence (if any) to support the claim of mediation.

However, I can't tell if you're really asking about how to interpret signs and/or values of your estimated model parameters. If that's your question, please elaborate the query to indicate: (a) what variables you have; (b) the proposed relationships; and (c) the estimated model coefficients for each of the paths. I bet you'll get a more focused response.

Good luck with your work.

The following link may help:

(20) (PDF) Making Sense of Mediating Analysis: A Marketing Perspective (researchgate.net)

Values between the variables are indicators of a positive or negative relationship/ cause-effect relationship. Basically, the independent variable is the cause. Its value is independent of other variables in your study. The dependent variable is the effect. Its value depends on changes in the independent variable. Positive values mean, one variable is causing another variable to be high (check the direction of the arrow to know which variable is causing which variable), and negative variable in general means, one variable is causing another variable to be low. EXAMPLE 1: Stress (1st variable) is causing job resignation(2nd variable) to be more if there is a +ve value. EXAMPLE 2: Stress has no effect on job resignation (if the value between these two is 0.00/neutral). EXAMPLE 3: Stress is causing job resignations to be low (if the value is -ve). You can get an accurate interpretation from SEM AMOS itself by Clicking the left-handed icon ON A PATH ANALYSIS DIAGRAM ( VIEW TEXT ) then clicking on ESTIMATES, then on your right-hand side there are many +ve or -ve values in tables, check for values under CR or PLable, here if you click on any value AMOS will give you text interpretation. Good Luck Mohammad Irshad