We can use Causal Mediation to analyze how treatments work. There is a treatment that has an effect on our outcome of interest, and there is an intermediate variable that maybe 'mediating this effect'. By 'mediating this effect', I mean that the treatment first affects the intermediate variable, and then the intermediate variable affects the outcome. Not only is there a temporal ordering of the effects, but there is a chain of causal effects. We often call the intermediate variable a 'mediator'.
Using tools from Causal Inference, we can reveal the assumptions we are making when we make statements like "the treatment worked because it impacted the mediator, and the mediator in turn impacted the outcome".
My research focuses on functional causal mediation, designs of experiments for mediation, and measurement error and propensity score estimation.
Using tools from Causal Inference, we can reveal the assumptions we are making when we make statements like "the treatment worked because it impacted the mediator, and the mediator in turn impacted the outcome".
My research focuses on functional causal mediation, designs of experiments for mediation, and measurement error and propensity score estimation.