5. Example of a Basic Test of Mediation

The simplest mediation analysis involves a single independent variable, a dependent variable, and a hypothesized mediator.

The unmediated model is represented by the direct effect of x on y, quantified as c.

However, the effect of X on Y may be mediated by a process, or mediating variable M.

Basic Mediation
Complete mediation is said to occur when X no longer affects Y after M has been controlled for. In this case path c’ is zero.

Partial mediation is the case in which the path from X to Y is reduced in absolute size but is still different from zero when the mediator is controlled for.

In path analysis, an independent variable is called an exogenous variable. Any variable that is predicted by another variable acts as a dependent variable and is called an endogenous variable.

Example: Effects of visual anonymity on attraction to the group

Here's an example of a simple mediation analysis relating to my own research.

In this example I test some ideas about deindividuation theory that derive from a social identity approach to group behaviour.

My basic hypothesis was that visual anonymity among group members increases attraction to the group.

I test this by regressing group attraction onto a measure of visual anonymity, and found that visual anonymity significantly affected group attraction with a Beta of .40.

mediation analysis example of direct effect

However, I also hypothesized that this effect was mediated by the extent to which group members perceived or categorized themselves as part of the group – a self-categorization variable. I next test this mediated effect and measure the change in the direct effect, with the following results….

Mediated effect, an example

As you can see from the path diagram, visual anonymity had an effect on self-categorization, which in turn had an effect on group attraction. At the same time, the direct effect of anonymity on group attraction (which was previously .50) is now reduced to just about zero, after the mediated effect is taken into account. In other words the hypothesized mediated effect accounted for just about all of the effect of anonymity on group attraction.

Note that self-categorization is considered here to be a mediator, not a moderator. That is, my model was not that visual anonymity increases group attraction when group members self-categorize as part of the group.

Instead, my model was that anonymity increases group attraction because it increased self-categorization.

In other words, my model addressed how and why anonymity achieves its effect, not when it achieves its effect.

NEXT: 6. Mediation Analysis: Procedures and Tests


Statistics Training: Introduction to Path Analysis


If you found this article useful, please share so others can find it.