Skip to main content
. Author manuscript; available in PMC: 2010 Aug 25.
Published in final edited form as: Psychol Methods. 2008 Dec;13(4):314–336. doi: 10.1037/a0014207

Figure 2.

Figure 2

Assumptions to identify causal effects in the structural equation modeling approach. In this hypothetical setting, 6 individuals are randomly assigned either to the treatment or to the control condition. Y = outcome; M = mediator; X = treatment assignment; a = the effect of treatment assignment on M; b = the effect of M on Y conditional on the effect of X on Y; c = the direct effect of X on Y conditional on the effect of M on Y; ab = a × b, which is the indirect effect of X on Y. In this illustrative example, c = 1.0, b = 1.0, and a takes one of the three values (0, 1, 2). Panel A shows the assumption of ignorability (i.e., 6 individuals with different M values are comparable on all measured and unmeasured covariates). In Panel B, constant effect implies that treatment assignment effect is the same across different M values. In Panel C, linearity implies that the relation between Y and M is linear. Panel D shows how the total effect is partitioned into the direct (c) and indirect (ab) effects.