How to test the significance of a mediation effect?, To perform a sobel test in R is covered in this lesson.
Using R to run a Sobel test
We can use the bda package to run a sobel test in R.
How to test the significance of a mediation effect
The following is the fundamental syntax for performing a sobel test.
where iv stands for the independent variable, dv for dependent variable, and mv is for the mediator variable.
Using a list of 50 normal random variables as the mediator variable, independent variable, and dependent variable, the following code performs a Sobel test.
Sobel Aroian Goodman z.value -0.5312235 -0.4533736 -0.6708299 p.value 0.5952639 0.6502797 0.5023289
In this instance, the values in the Sobel column are what we are most concerned with. The p-value that corresponds to the z-value of -0.5312235 is 0.5952639.
We would be unable to reject the null hypothesis that there is no mediation effect because this p-value is higher than the alpha level of 0.05.
There is therefore no statistically significant mediation effect.
Note: In your own test, you can use a different alpha level. The usual options for alpha are 0.01, 0.05, and 0.10.