When using argument contrasts to compare indirect effects, all possible combinations of indirect effects are currently compared. This is not very meaningful in serial multiple mediator models. For instance, with two serial mediators, some indirect effects are the product of two coefficients whereas another is the product of three coefficients, and therefore those can be expected to be different in magnitude. Furthermore, the number of comparisons quickly explodes since three serial mediators already imply seven indirect effects.
If would be more meaningful to compare only the indirect effects that are a product of the same number of coefficients.
When using argument
contrasts
to compare indirect effects, all possible combinations of indirect effects are currently compared. This is not very meaningful in serial multiple mediator models. For instance, with two serial mediators, some indirect effects are the product of two coefficients whereas another is the product of three coefficients, and therefore those can be expected to be different in magnitude. Furthermore, the number of comparisons quickly explodes since three serial mediators already imply seven indirect effects.If would be more meaningful to compare only the indirect effects that are a product of the same number of coefficients.