Figure 1.

Adapted from Montoya and Hayes (2017) and depicts the path-analytic model of a two-condition (passage of time) within-participant mediation model. The independent variable (“X”) does not exist in the dataset and its effect is carried by the difference scores in the mediator and outcome. As recommended by Curran and Bauer (2007), the triangle in the path model signifies that X is quantified by the regression constants in the equations of the difference in the mediator and the difference in the outcome between the two conditions (i.e., pre- and post-intervention).