Mediation versus Moderation – What’s the difference?

Mediation vs. Moderation

It’s easy to confuse the two. They sound similar, and while they both look at how a third variable fits into a relationship of interest, they are not the same. Let’s break it all down. In this post, we will highlight some key characteristics of mediation and moderation. We will also talk about some of the key differences between these analyses.


A mediation analysis is an extension of multiple regression. We start to think about mediation when we want to explain why or how X affects Y. This tells us more information about how or why an independent variable affects a dependent variable. The relationship between X and Y is the total effect.

The bivariate regression or Pearson correlation between X and Y is the total effect.
The total effect is the relationship between X and Y when a mediator is not present.

In mediation, we add an independent variable called the mediator. Mediators mediate the relationship between X and Y. This occurs by X affecting M leading to M affecting Y, which is called the indirect effect. The direct effect is the relationship between X and Y in the presence of a mediator. Mediation occurs when (1) there is a statistically significant indirect effect (2) the direct effect is smaller than the total effect.

In a mediation analysis, we want to obtain the zero-order or bivariate correlations between X and M, and between M and Y. Next, a multiple regression is used to get the direct and indirect effects where X and M are independent variables and Y as the dependent variable.


Moderation analyses look at interactions. In other words, we’re interested in whether the effect of X on Y varies depending on another variable (i.e., the moderator).

Moderator variables modify the relationship between X and Y. They affect the strength and direction of the relationship between X and Y. That means that X‘s effect on Y can change depending on the moderator.

In a moderation analysis, we want to calculate the interaction term (X*M).
Using hierarchical multiple regression analysis, we enter the two independent variables (X and W) in Step 1, the interaction term in Step 2, and Y as the dependent variable.

An interaction or product term represents the moderator effect. We can calculate the interaction term by multiplying the independent variable by the moderator (X*W).

Key Differences

Mediators are possible explanations for a relationship between X and Y. Moderators affect the magnitude of the effect of X on Y. Another difference is in the relationship that mediators and moderators have with the independent variable. In theory, mediators result from the independent variable (i.e., X M). On the other hand, there is no directional relationship assumed between X and a moderator (i.e., X M).

Overall, we use mediation analyses to explain relationships. We use moderation analyses to understand what variables affect the strength and direction of a relationship.


Cheyenne Chooi, PhD Student (Neuropsychology)

Cheyenne is a graduate of the University of Western Australia where she received a Bachelor's degree in Science, and a Bachelor's degree in Science with Honours in psychology. She is currently a third year PhD Student researching factors related to Alzheimer's disease. Her current research focuses on the concept of cognitive reserve.

Leave a Reply

Your email address will not be published. Required fields are marked *