Variables, Mediating Types

When to Expect a Mediating Variable

Before testing for mediating variables, it is imperative to understand where a mediating variable might exist. Whereas there is no definite sign to look for, there are some clues that can help in determining whether a mediating variable might be present. For instance, if previous literature or theory suggests that there could be a relationship between all three of the study variables, it might be likely for the mediation to occur again in the current sample with the correlation of the same three variables. Another method for determining if a mediating variable might exist is if all three variables correlate after running a correlation matrix.

Some variables might present themselves as moderating variables in certain scenarios. When trying to distinguish between mediating and moderating variables, it is important to remember the necessity of the third variable. For example, if it seems that the relationship between the independent variable and the dependent variable would exist without the presence of the third variable, then the variable is probably a moderating variable. However, if it seems as though the third variable’s presence is necessary for the other relationship between the two variables to exist, then the third variable is likely a mediator. A mediating variable is the conduit that creates the [Page 1852]connection between two other variables. Without the presence of this mediating variable, the channel linking X to Y would be nonexistent. For example, individuals with eating disorders may also be inclined to share their struggles with close friends. However, if social support is the mediator between engaging in disordered eating and disclosing those behaviors to friends, then the absence of social support from those friends would prevent the link between disordered eating and disclosure to ever occur.

...