Theory and Analysis of Total, Direct, and Indirect Causal Effects

Axel Mayer, Felix Thoemmes, Norman Rose, Rolf Steyer, Stephen West

Research output: Contribution to journalArticle

15 Citations (Scopus)

Abstract

Mediation analysis, or more generally models with direct and indirect effects, are commonly used in the behavioral sciences. As we show in our illustrative example, traditional methods of mediation analysis that omit confounding variables can lead to systematically biased direct and indirect effects, even in the context of a randomized experiment. Therefore, several definitions of causal effects in mediation models have been presented in the literature (Baron & Kenny, 1986; Imai, Keele, & Tingley, 2010; Pearl, 2012). We illustrate the stochastic theory of causal effects as an alternative foundation of causal mediation analysis based on probability theory. In this theory we define total, direct, and indirect effects and show how they can be identified in the context of our illustrative example. A particular strength of the stochastic theory of causal effects are the causality conditions that imply causal unbiasedness of effect estimates. The causality conditions have empirically testable implications and can be used for covariate selection. In the discussion, we highlight some similarities and differences of the stochastic theory of causal effects with other theories of causal effects.

Original languageEnglish (US)
Pages (from-to)425-442
Number of pages18
JournalMultivariate Behavioral Research
Volume49
Issue number5
DOIs
StatePublished - Sep 1 2014

Fingerprint

Causal Effect
Causality
Mediation
Probability Theory
Behavioral Sciences
Confounding Factors (Epidemiology)
Randomized Experiments
Unbiasedness
Confounding
Biased
Covariates
Causal
Imply
Alternatives
Model
Estimate

ASJC Scopus subject areas

  • Experimental and Cognitive Psychology
  • Statistics and Probability
  • Arts and Humanities (miscellaneous)

Cite this

Theory and Analysis of Total, Direct, and Indirect Causal Effects. / Mayer, Axel; Thoemmes, Felix; Rose, Norman; Steyer, Rolf; West, Stephen.

In: Multivariate Behavioral Research, Vol. 49, No. 5, 01.09.2014, p. 425-442.

Research output: Contribution to journalArticle

Mayer, Axel ; Thoemmes, Felix ; Rose, Norman ; Steyer, Rolf ; West, Stephen. / Theory and Analysis of Total, Direct, and Indirect Causal Effects. In: Multivariate Behavioral Research. 2014 ; Vol. 49, No. 5. pp. 425-442.
@article{6ecb41894bbe4914ab9134c3866efccf,
title = "Theory and Analysis of Total, Direct, and Indirect Causal Effects",
abstract = "Mediation analysis, or more generally models with direct and indirect effects, are commonly used in the behavioral sciences. As we show in our illustrative example, traditional methods of mediation analysis that omit confounding variables can lead to systematically biased direct and indirect effects, even in the context of a randomized experiment. Therefore, several definitions of causal effects in mediation models have been presented in the literature (Baron & Kenny, 1986; Imai, Keele, & Tingley, 2010; Pearl, 2012). We illustrate the stochastic theory of causal effects as an alternative foundation of causal mediation analysis based on probability theory. In this theory we define total, direct, and indirect effects and show how they can be identified in the context of our illustrative example. A particular strength of the stochastic theory of causal effects are the causality conditions that imply causal unbiasedness of effect estimates. The causality conditions have empirically testable implications and can be used for covariate selection. In the discussion, we highlight some similarities and differences of the stochastic theory of causal effects with other theories of causal effects.",
author = "Axel Mayer and Felix Thoemmes and Norman Rose and Rolf Steyer and Stephen West",
year = "2014",
month = "9",
day = "1",
doi = "10.1080/00273171.2014.931797",
language = "English (US)",
volume = "49",
pages = "425--442",
journal = "Multivariate Behavioral Research",
issn = "0027-3171",
publisher = "Psychology Press Ltd",
number = "5",

}

TY - JOUR

T1 - Theory and Analysis of Total, Direct, and Indirect Causal Effects

AU - Mayer, Axel

AU - Thoemmes, Felix

AU - Rose, Norman

AU - Steyer, Rolf

AU - West, Stephen

PY - 2014/9/1

Y1 - 2014/9/1

N2 - Mediation analysis, or more generally models with direct and indirect effects, are commonly used in the behavioral sciences. As we show in our illustrative example, traditional methods of mediation analysis that omit confounding variables can lead to systematically biased direct and indirect effects, even in the context of a randomized experiment. Therefore, several definitions of causal effects in mediation models have been presented in the literature (Baron & Kenny, 1986; Imai, Keele, & Tingley, 2010; Pearl, 2012). We illustrate the stochastic theory of causal effects as an alternative foundation of causal mediation analysis based on probability theory. In this theory we define total, direct, and indirect effects and show how they can be identified in the context of our illustrative example. A particular strength of the stochastic theory of causal effects are the causality conditions that imply causal unbiasedness of effect estimates. The causality conditions have empirically testable implications and can be used for covariate selection. In the discussion, we highlight some similarities and differences of the stochastic theory of causal effects with other theories of causal effects.

AB - Mediation analysis, or more generally models with direct and indirect effects, are commonly used in the behavioral sciences. As we show in our illustrative example, traditional methods of mediation analysis that omit confounding variables can lead to systematically biased direct and indirect effects, even in the context of a randomized experiment. Therefore, several definitions of causal effects in mediation models have been presented in the literature (Baron & Kenny, 1986; Imai, Keele, & Tingley, 2010; Pearl, 2012). We illustrate the stochastic theory of causal effects as an alternative foundation of causal mediation analysis based on probability theory. In this theory we define total, direct, and indirect effects and show how they can be identified in the context of our illustrative example. A particular strength of the stochastic theory of causal effects are the causality conditions that imply causal unbiasedness of effect estimates. The causality conditions have empirically testable implications and can be used for covariate selection. In the discussion, we highlight some similarities and differences of the stochastic theory of causal effects with other theories of causal effects.

UR - http://www.scopus.com/inward/record.url?scp=84907468729&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84907468729&partnerID=8YFLogxK

U2 - 10.1080/00273171.2014.931797

DO - 10.1080/00273171.2014.931797

M3 - Article

AN - SCOPUS:84907468729

VL - 49

SP - 425

EP - 442

JO - Multivariate Behavioral Research

JF - Multivariate Behavioral Research

SN - 0027-3171

IS - 5

ER -