Mediation analysis allowing for exposure-mediator interactions and causal interpretation: theoretical assumptions and implementation with SAS and SPSS macros

By L. Valeri, Tyler J. VanderWeele. Psychological methods, 2013

1,990 citations recorded by the source corpus.

Mediation analysis is a useful and widely employed approach to studies in the field of psychology and in the social and biomedical sciences. The contributions of this paper are several-fold. First we seek to bring the developments in mediation analysis for non linear models within the counterfactual framework to the psychology audience in an accessible format and compare the sorts of inferences about mediation that are possible in the presence of exposure-mediator interaction when using a counterfactual versus the standard statistical approach. Second, the work by VanderWeele and Vansteelandt (2009, 2010) is extended here to allow for dichotomous mediators and count outcomes. Third, we provide SAS and SPSS macros to implement all of these mediation analysis techniques automatically and we compare the types of inferences about mediation that are allowed by a variety of software macros.

Subfields: Clinical Psychology, Experimental and Cognitive Psychology, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/863a3b679cfc587fab082f741c26f8dc7af52104

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Clinical Psychology

In Experimental and Cognitive Psychology

In Statistics and Probability

Citations in the corpus

Meaningful Mediation Analysis: Plausible Causal Inference and Informative Communication (2017, background): “For instance, the homogeneity condition specifies that the average treatment effect holds similarly for all participants, which may be more an exception than the rule (Imai et al. 2010; Valeri and VanderWeele 2013).”
Mendelian randomisation for mediation analysis: current methods and challenges for implementation (2019, methodology): “6 Counterfactual reasoning has been used to develop methods that can address some of the previously described strong assumptions [10-14].”
The Mediation Myth (2015, methodology): “Statistical significance is too often taken as a ‘‘gold standard’’ for establishing mediation, but this decision 210 KLINE criterion is extreme flawed.”
Two-condition within-participant statistical mediation analysis: A path-analytic framework. (2017, background): “…new commands as “macros,” and there are numerous published articles describing macros researchers have created for use in mediation analysis (e.g., Fairchild, MacKinnon, Taborga, & Taylor, 2009; Hayes, 2013; MacKinnon et al., 2007; Preacher & Hayes, 2004, 2008; Valeri & VanderWeele, 2013).”
Interpretation and identification of causal mediation. (2014, positive): “(16) and (17) (Coffman and Zhong, 2012; Imai et al., 2010c; Muthén, 2011; Pearl, 2010a, 2012a; Valeri and VanderWeele, 2013; VanderWeele and Vansteelandt, 2009).”
Mediation Analysis with Multiple Mediators (2014, methodology): “Similar expressions also hold if the binary outcome is replaced by a count outcome with a Poisson or negative binomial model and rate ratios are used (Valeri and VanderWeele, 2013).”
Regression-based statistical mediation and moderation analysis in clinical research: Observations, recommendations, and implementation. (2017, background): “…recently, the “causal inference,” “counterfactual,” or “potential outcomes” tradition in mediation analysis that has been gaining some traction in the literature (though less in practice) allows X and M to interact (Imai, Keele, Tingley, 2010; Muthén & Asparouhov, 2015; Valeri & VanderWeele, 2013).”
Parent-Child Information Frictions and Human Capital Investment: Evidence from a Field Experiment (2020, neutral): “[51] Valeri, Linda and Tyler VanderWeele (2013).”
Mediation Analysis: A Practitioner's Guide. (2016, background): “It assumes that exposure-outcome confounding (Assumption A1) and exposure-mediator confounding (Assumption A3) are controlled but that there might be unmeasured mediator-outcome confounders denote the maximum risk ratio relating U and on a u s e the outcome Y among the exposed subjects across…”