
Elizabeth J. answered 07/07/20
MA Cognitive and Social Processes with 10+ years using SPSS
A moderator variable uniquely changes the relationship between the independent variable and the dependent variable. There can be a relationship (or not) between the independent and dependent variable but when a moderator variable is entered into the equation, the outcome (dependent variable) acts differently than when the IV stands by itself. Rather than a mediator variable that is linear from IV -- Mediator -- DV, the moderator shifts to "outside" the linear equation by changing how the IV impacts the DV.
For example, if I am interested in how a certain grade level (1st, 2nd, or 3rd) responds to a certain exam item, but I think that a third variable such as their IQ scores could change the outcome of the grade level on the exam item, then I might use IQ as a moderator variable. This would be a good prediction to make if the grade level by itself is known to not impact the outcome variable. However, when IQ is entered into the analysis as a moderator, the IV has a new (and actual) impact on the DV.