Interpretation and identification of causal mediation
Open accessThe DOI matches its Crossref record; 88 of the 88 references it cites were found in a registry or an index.
Abstract
This article reviews the foundations of causal mediation analysis and offers a general and transparent account of the conditions necessary for the identification of natural direct and indirect effects, thus facilitating a more informed judgment of the plausibility of these conditions in specific applications. I show that the conditions usually cited in the literature are overly restrictive and can be relaxed substantially without compromising identification. In particular, I show that natural effects can be identified by methods that go beyond standard adjustment for confounders, applicable to observational studies in which treatment assignment remains confounded with the mediator or with the outcome. These identification conditions can be validated algorithmically from the diagrammatic description of one's model and are guaranteed to produce unbiased results whenever the description is correct. The identification conditions can be further relaxed in parametric models, possibly including interactions, and permit one to compare the relative importance of several pathways, mediated by interdependent variables.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Citations
Cites
Baron, R. M., & Kenny, D. A. (1986) · Journal of Personality and Social Psychology · doi:10.1037/0022-3514.51.6.1173 · Cited by 91,532
The central role of the propensity score in observational studies for causal effects
ROSENBAUM, P. R., & RUBIN, D. B. (1983) · Biometrika · doi:10.1093/biomet/70.1.41 · Cited by 31,241
Structural Equations with Latent Variables.
Clogg, C. C., & Bollen, K. A. (1991) · Contemporary Sociology · doi:10.2307/2072165 · Cited by 17,442
Causality: Models, Reasoning, and Inference
Pearl, J. (2009) · doi:10.1017/cbo9780511803161 · Cited by 11,653
Self-citationAddressing Moderated Mediation Hypotheses: Theory, Methods, and Prescriptions
Preacher, K. J., Rucker, D. D., & Hayes, A. F. (2007) · Multivariate Behavioral Research · doi:10.1080/00273170701341316 · Cited by 9,973
Beyond Baron and Kenny: Statistical Mediation Analysis in the New Millennium
Hayes, A. F. (2009) · Communication Monographs · doi:10.1080/03637750903310360 · Cited by 9,757 · cited 4×
“, Imai, Keele, & Yamamoto, 2010) or uncorrelated error terms. 3 Imai, Keele, and Yamamoto (2010) and Imai, Keele, Tingley, and Yamamoto (2011) discussed similarities and differences among several versions of the identifying assumptions, and Shpitser and VanderWeele (2011) delineated the context under which a restricte…”
InfluentialEstimating causal effects of treatments in randomized and nonrandomized studies
Rubin, D. B. (1974) · Journal of Educational Psychology · doi:10.1037/h0037350 · Cited by 9,436 · cited 2×
“Counterfactual thinking in statistics (Holland, 1988; Rubin, 1974) and epidemiology (Robins and Greenland, 1992), together with a formal semantics based on non-parametric structural equations (Balke and Pearl, 1995; Halpern, 1998; Pearl, 2001) have given causal mediation analysis a sound theoretical basis and have ext…”
Introduction to Statistical Mediation Analysis
MacKinnon, D. P. (2012) · doi:10.4324/9780203809556 · Cited by 5,618
Fisher, R. A. (1936) · BMJ · doi:10.1136/bmj.1.3923.554-a · Cited by 4,980
CAUSALITY: MODELS, REASONING AND INFERENCE
Hitchcock, C. (2001) · The Philosophical Review · doi:10.1215/00318108-110-4-639 · Cited by 4,741
Self-citation
Cited by
mediation:R Package for Causal Mediation Analysis
Tingley, D., Yamamoto, T., Hirose, K., Keele, L., & Imai, K. (2014) · Journal of Statistical Software · doi:10.18637/jss.v059.i05 · Cited by 4,155
Do Not Cross Me: Optimizing the Use of Cross-Sectional Designs
Spector, P. E. (2019) · Journal of Business and Psychology · doi:10.1007/s10869-018-09613-8 · Cited by 1,366
“The notion that the observed connection between X and Y might be attributable to additional causal variables has been discussed in other contexts such as endogeneity (Antonakis, Bendahan, Jacquart, & Lalive, 2010) and causal mediation (Imai, Keele, & Tingley, 2010; Pearl, 2014).”
BackgroundVaccari, C., & Chadwick, A. (2020) · Social Media + Society · doi:10.1177/2056305120903408 · Cited by 978
Advances in Mediation Analysis: A Survey and Synthesis of New Developments
Preacher, K. J. (2015) · Annual Review of Psychology · doi:10.1146/annurev-psych-010814-015258 · Cited by 793
Hair, J. F., Babin, B. J., & Krey, N. (2017) · Journal of Advertising · doi:10.1080/00913367.2017.1281777 · Cited by 780
“Pearl (2009, 2010, 2012, 2014) provides a comprehensive review of causality and SEM, and points out that the greater accessibility to statistical techniques coincides with users who may lack a fundamental understanding of key principles.”
BackgroundCausal Interpretations of Black-Box Models
Zhao, Q., & Hastie, T. (2021) · Journal of Business & Economic Statistics · doi:10.1080/07350015.2019.1624293 · Cited by 462
Kline, R. B. (2015) · Basic and Applied Social Psychology · doi:10.1080/01973533.2015.1049349 · Cited by 460
“…or potential outcomes model (Rubin, 2005), principal stratification with propensity scores (Jo, Stuart, MacKinnon, & Vinokur, 2011), instrumental variable methods (Angrist, Imbens, & Rubin, 1996), conditional process modeling (Hayes, 2013), and causal mediation analysis (Pearl, 2014).”
BackgroundMethodChoi, S. L., Goh, C. F., Adam, M. B. H., & Tan, O. K. (2016) · Human Resources for Health · doi:10.1186/s12960-016-0171-2 · Cited by 353
“The mediating analysis is useful in predicting how the causal effect of transformational leadership on job satisfaction is intervened by employee empowerment and thus has tangible policy implications [19].”
BackgroundCausal Intervention for Weakly-Supervised Semantic Segmentation
Zhang, D., Zhang, H., Tang, J., Hua, X., & Sun, Q. (2020) · arXiv · doi:10.48550/arxiv.2009.12547 · Cited by 353
“CONTA is based on the backdoor adjustment [37] for P (Y ∣do(X)).”
BackgroundMethodCEO social capital, risk-taking and corporate policies
Ferris, S. P., Javakhadze, D., & Rajkovic, T. (2017) · Journal of Corporate Finance · doi:10.1016/j.jcorpfin.2017.09.003 · Cited by 286
Wang, T., Huang, J., Zhang, H., & Sun, Q. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.01077 · Cited by 291
“There has been a growing amount of efforts in marrying complementary strengths of deep learning and causal reasoning [49, 48] and have been explored in several contexts, including image classification [8, 40], reinforcement learning [46, 11, 5] and adversarial learning [28, 26].”
Backgroundmedflex: An R Package for Flexible Mediation Analysis using Natural Effect Models
Steen, J., Loeys, T., Moerkerke, B., & Vansteelandt, S. (2017) · Journal of Statistical Software · doi:10.18637/jss.v076.i11 · Cited by 278
“1 Pearl (2001, 2014) offers a weaker set of conditions, which does not require a common set of baseline covariates to deconfound each of the possibly confounded relations, but allows for separate adjustment sets for the exposure-mediator relation on the one hand and the exposure-outcome and…”
A novel measure of effect size for mediation analysis
Lachowicz, M. J., Preacher, K. J., & Kelley, K. (2018) · Psychological Methods · doi:10.1037/met0000165 · Cited by 271
“The issue of confounding of the causal effect in mediation analysis has received substantial attention in the methodological literature over the last decade (e.g., Imai, Keele, & Tingley, 2010; Pearl, 2014; Vanderweele, 2015).”
BackgroundBelief in COVID‐19 Conspiracy Theories Reduces Social Distancing over Time
Bierwiaczonek, K., Kunst, J. R., & Pich, O. (2020) · Applied Psychology: Health and Well-Being · doi:10.1111/aphw.12223 · Cited by 265
Agyeiwaah, E., Otoo, F. E., Suntikul, W., & Huang, W.-J. (2019) · Journal of Travel & Tourism Marketing · doi:10.1080/10548408.2018.1541775 · Cited by 238
“through one or more intervening variables, resulting in direct and indirect effects on an outcome variable (Little, 2013; Pearl, 2014), examination of mediation is pertinent.”
BackgroundComment: understanding Simpson’s paradox
Pearl, J. (2014) · The American Statistician · doi:10.1080/00031305.2014.876829 · Cited by 206
Self-citationAntibiotics and Antibiotic Resistance in Agroecosystems: State of the Science
Williams-Nguyen, J., Sallach, J. B., Bartelt-Hunt, S., Boxall, A. B., Durso, L. M., McLain, J. E., Singer, R. S., Snow, D. D., & Zilles, J. L. (2016) · Journal of Environmental Quality · doi:10.2134/jeq2015.07.0336 · Cited by 195
Imbens, G. W. (2020) · Journal of Economic Literature · doi:10.1257/jel.20191597 · Cited by 186
Causal Mediation Programs in R, M plus, SAS, SPSS, and Stata
Valente, M. J., Rijnhart, J. J. M., Smyth, H. L., Muniz, F. B., & MacKinnon, D. P. (2020) · Structural Equation Modeling: A Multidisciplinary Journal · doi:10.1080/10705511.2020.1777133 · Cited by 186
Causal Attention for Interpretable and Generalizable Graph Classification
Sui, Y., Wang, X., Wu, J., Lin, M., He, X., & Chua, T.-S. (2022) · Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining · doi:10.1145/3534678.3539366 · Cited by 189
“Even if 𝐶 has no direct link to 𝑌 , the backdoor path will cause 𝐶 to establish a spurious correlation with 𝑌 , e.g., making wrong predictions based on shortcut feature 𝑆 instead of causal feature 𝐶 .”
BackgroundMethodInfluentialKock, N., Mayfield, M., Mayfield, J., Sexton, S., & De La Garza, L. M. (2019) · Journal of Leadership & Organizational Studies · doi:10.1177/1548051818806290 · Cited by 183
“The model presents a causal statement in line with Pearl (2012, 2014), Bollen and Pearl’s (2013), and Goldthorpe’s (2001) suggestions for establishing causal linkages between constructs.”
BackgroundValente, M. J., & MacKinnon, D. P. (2017) · Structural Equation Modeling: A Multidisciplinary Journal · doi:10.1080/10705511.2016.1274657 · Cited by 173
“Pearl, 2014, for alternate assumptions regarding identification of causal mediated effects).”
BackgroundWhen Celebrities Count: Power Distance Beliefs and Celebrity Endorsements
Winterich, K. P., Gangwar, M., & Grewal, R. (2018) · Journal of Marketing · doi:10.1509/jm.16.0169 · Cited by 168
“Because we measure the mediators after exposing participants to the experimental stimuli, the mediators should be endogenous; this endogeneity results in bias for the relationship between the mediator and the outcome variables (Pearl 2014).”
BackgroundDistilling Causal Effect of Data in Class-Incremental Learning
Hu, X., Tang, K., Miao, C., Hua, X.-S., & Zhang, H. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00395 · Cited by 167
“Based on the graph, we can formulate the causal effect between variables by causal intervention, which is denoted as the do(·) operation [30, 31].”
MethodSianoja, M., Syrek, C. J., de Bloom, J., Korpela, K., & Kinnunen, U. (2018) · Journal of Occupational Health Psychology · doi:10.1037/ocp0000083 · Cited by 166
Can Emotions Capture the Elusive Gain-Loss Framing Effect? A Meta-Analysis
Nabi, R. L., Walter, N., Oshidary, N., Endacott, C. G., Love-Nichols, J., Lew, Z. J., & Aune, A. (2020) · Communication Research · doi:10.1177/0093650219861256 · Cited by 165
MacKinnon, D. P., Valente, M. J., & Gonzalez, O. (2020) · Prevention Science · doi:10.1007/s11121-019-01076-4 · Cited by 165
“In general, prevention has been slow to adopt new causal methods with some exceptions (Imai et al. 2010; Jo 2008; Liu et al. 2013; Pearl 2014; Stuart et al. 2015; Valeri and VanderWeele 2013), at least in part because the links between the potential outcomes framework and traditional analysis have…”
BackgroundChang, C.-J., Barr, D. B., Ryan, P. B., Panuwet, P., Smarr, M. M., Liu, K., Kannan, K., Yakimavets, V., Tan, Y., Ly, V., et al. (2022) · Environment International · doi:10.1016/j.envint.2021.106964 · Cited by 166
Economic Shocks and Crime: Evidence from the Brazilian Trade Liberalization
Dix-Carneiro, R., Soares, R. R., & Ulyssea, G. (2018) · American Economic Journal Applied Economics · doi:10.1257/app.20170080 · Cited by 198
Stavola, B. L. D., Daniel, R., Ploubidis, G. B., & Micali, N. (2014) · American Journal of Epidemiology · doi:10.1093/aje/kwu239 · Cited by 177
Exposure to concentrated ambient PM2.5 alters the composition of gut microbiota in a murine model
Wang, W., Zhou, J., Chen, M., Huang, X., Xie, X., Li, W., Cao, Q., Kan, H., Xu, Y., & Ying, Z. (2018) · Particle and Fibre Toxicology · doi:10.1186/s12989-018-0252-6 · Cited by 158
“[25] The mediation was then calculated based on two linear mixed effect (LME) models as demonstrated below [26]:”
Blood pressure lowering and prevention of dementia: an individual patient data meta-analysis
Peters, R., Xu, Y., Fitzgerald, O., Aung, H. L., Beckett, N., Bulpitt, C., Chalmers, J., Forette, F., Gong, J., Harris, K., et al. (2022) · European Heart Journal · doi:10.1093/eurheartj/ehac584 · Cited by 142
The “true” indirect effect won't (always) stand up: When and why reverse mediation testing fails
Lemmer, G., & Gollwitzer, M. (2017) · Journal of Experimental Social Psychology · doi:10.1016/j.jesp.2016.05.002 · Cited by 145
Self−leadership in the context of part−time teleworking.
Müller, T., & Niessen, C. (2019) · Journal of Organizational Behavior · doi:10.1002/job.2371 · Cited by 144
O'Neill, D. G., James, H., Brodbelt, D. C., Church, D. B., & Pegram, C. (2021) · BMC Veterinary Research · doi:10.1186/s12917-021-02775-3 · Cited by 170
IDENTIFYING MECHANISMS BEHIND POLICY INTERVENTIONS VIA CAUSAL MEDIATION ANALYSIS
Keele, L. J., Tingley, D. H., & Yamamoto, T. (2015) · Journal of Policy Analysis and Management · doi:10.1002/pam.21853 · Cited by 166
The Cambridge Handbook of Personality Psychology
(2020) · Cambridge University Press eBooks · doi:10.1017/9781108264822 · Cited by 161
Flexible Mediation Analysis With Multiple Mediators
Steen, J., Loeys, T., Moerkerke, B., & Vansteelandt, S. (2017) · American Journal of Epidemiology · doi:10.1093/aje/kwx051 · Cited by 123
Ferguson, K. K., Chen, Y., VanderWeele, T. J., McElrath, T. F., Meeker, J. D., & Mukherjee, B. (2016) · Environmental Health Perspectives · doi:10.1289/ehp282 · Cited by 138
“The NIE refers to the change in the odds of preterm birth (Y) when urinary phthalate concentration is held at a specific level (e.g., a) and 8-isoprostane is set at what it would have naturally been at for the defined change in urinary phthalate concentration (e.g., a*) (Pearl 2014).”
Kim, Y.-S. G., & Graham, S. (2021) · Journal of Educational Psychology · doi:10.1037/edu0000564 · Cited by 132
High-dimensional multivariate mediation with application to neuroimaging data
Chén, O. Y., Crainiceanu, C. M., Ogburn, E. L., Caffo, B. S., Wager, T. D., & Lindquist, M. A. (2017) · Biostatistics · doi:10.1093/biostatistics/kxx027 · Cited by 147
“See Robins and Richardson (2010) and Pearl (2014) for detailed discussion of these assumptions, and for a critical evaluation of these assumptions in the high-dimensional setting see Huang and Pan (2015).”
BackgroundYang, Z., Khatibi, E., Nagesh, N., Abbasian, M., Azimi, I., Jain, R., & Rahmani, A. M. (2024) · Smart Health · doi:10.1016/j.smhl.2024.100465 · Cited by 120
“Briefly, mediator analysis in causality involves investigating the intermediate processes through which an independent variable influences a dependent variable [53].”
BackgroundConfounding in statistical mediation analysis: What it is and how to address it.
Valente, M. J., Pelham, W. E., Smyth, H., & MacKinnon, D. P. (2017) · Journal of Counseling Psychology · doi:10.1037/cou0000242 · Cited by 111
“If there are confounders of the mediator – outcome relation that are affected by X, traditional methods such as ANCOVA, would not provide an unbiased estimate of the mediated or direct effect without further assumptions (Imai & Yamamoto, 2013; Moerkerke, et al., 2015; Pearl, 2014; Tchetgen & VanderWeele, 2014).”
BackgroundMethodJin, A., Wang, S., Li, J., Wang, M., Lin, J., Li, H., Meng, X., Wang, Y., & Pan, Y. (2023) · Stroke · doi:10.1161/strokeaha.122.039542 · Cited by 129
On the (In)Validity of Tests of Simple Mediation: Threats and Solutions
Pek, J., & Hoyle, R. H. (2016) · Social and Personality Psychology Compass · doi:10.1111/spc3.12237 · Cited by 108
“Typically, levels of X undergo random assignment such that causality in X→M and X→Y could be inferred because randomization reduces the possibility of confounding in these pathways (c.f., Pearl, 2014).”
BackgroundCano-Gamez, E., Burnham, K. L., Goh, C., Allcock, A., Malick, Z. H., Overend, L., Kwok, A. J., Smith, D. A. B., Sengers, H. P., Antcliffe, D. B., et al. (2022) · Science Translational Medicine · doi:10.1126/scitranslmed.abq4433 · Cited by 117
“Mediation analysis is a statistical tool that tests the compatibility of a hypothesis with existing data by simulating how the response variablewould change if other variables were altered one at a time (18, 19).”
MethodA cautionary note on the power of the test for the indirect effect in mediation analysis
Loeys, T., Moerkerke, B., & Vansteelandt, S. (2015) · Frontiers in Psychology · doi:10.3389/fpsyg.2014.01549 · Cited by 134
Daugherty, A. M., Haacke, E. M., & Raz, N. (2015) · Journal of Neuroscience · doi:10.1523/jneurosci.4717-14.2015 · Cited by 111
“Yet, for testing the temporal order hypothesis, a longitudinal study with multiple occasions is needed, and to what extent observational studies can assess causality remains questionable (Winship and Morgan, 1999; Rubin, 2005; Pearl, 2014a,b).”
BackgroundText and Causal Inference: A Review of Using Text to Remove Confounding from Causal Estimates
Keith, K. A., Jensen, D. D., & O’Connor, B. (2020) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2020.acl-main.474 · Cited by 170
“Mediator variables require a different set of adjustments than confounders to find the “natural direct effect” between treatment and outcome (VanderWeele, 2015; Pearl, 2014).”
BackgroundImbens, G. W. (2019) · National Bureau of Economic Research · doi:10.3386/w26104 · Cited by 118
Lee, S., Zimmerman, M. E., Narkhede, A., Nasrabady, S. E., Tosto, G., Meier, I. B., Benzinger, T. L. S., Marcus, D. S., Fagan, A. M., Fox, N. C., et al. (2018) · PLoS ONE · doi:10.1371/journal.pone.0195838 · Cited by 105
Zeng, P., Shao, Z., & Zhou, X. (2021) · Computational and Structural Biotechnology Journal · doi:10.1016/j.csbj.2021.05.042 · Cited by 112
“In particular, besides the implicit assumption of temporal ordering between the exposure, mediator and outcome, we assume that all confounders in the three equations are correctly controlled for and that no confounders affecting both the mediator and the outcome are affected by the exposure [72,78,101,103,104].”
BackgroundGraetz, N., Boen, C., & Esposito, M. H. (2022) · Journal of Health and Social Behavior · doi:10.1177/00221465211066108 · Cited by 113
“…systems characterizing the time and place through which the Add Health cohort aged?” to be based on a “well-described exposure” (i.e., racialization in a racist system) rather than necessarily a “well-defined intervention” (Glymour 2006; Holland 2008; Morgan and Winship 2014; Pearl 2014).”
BackgroundLee, H., McAuley, J. H., Hübscher, M., Kamper, S. J., Traeger, A. C., & Moseley, G. L. (2016) · Pain · doi:10.1097/j.pain.0000000000000472 · Cited by 101
“analysis state that, for causal interpretation of mediation effects, careful consideration for potential confounders is required to satisfy the “no unmeasured confounding” assumption.(39,58) We further explored this inconsistent finding by testing our theoretical assumptions for the selection of 2 proposed confounders…”
BackgroundTemperament and brain networks of attention
Posner, M. I., & Rothbart, M. K. (2018) · Philosophical Transactions of the Royal Society B Biological Sciences · doi:10.1098/rstb.2017.0254 · Cited by 99
The Value of Descriptive Analytics: Evidence from Online Retailers
Berman, R., & Israeli, A. (2020) · Marketing Science · doi:10.1287/mksc.2022.1352 · Cited by 130
Sex-specific effects of microglial activation on Alzheimer’s disease proteinopathy in older adults
Casaletto, K. B., Nichols, E., Aslanyan, V., Simone, S. M., Rabin, J. S., Joie, R. L., Brickman, A. M., Dams-O’Connor, K., Palta, P., Kumar, R. G., et al. (2022) · Brain · doi:10.1093/brain/awac257 · Cited by 99
Insights into the Cross-world Independence Assumption of Causal Mediation Analysis
Andrews, R. M., & Didelez, V. (2020) · Epidemiology · doi:10.1097/ede.0000000000001313 · Cited by 119
Daugherty, A. M., & Raz, N. (2015) · NeuroImage · doi:10.1016/j.neuroimage.2015.12.045 · Cited by 83
“…that iron-related oxidative stress causes degeneration (e.g., Zhang et al., 2009) and our models of iron accumulation preceding shrinkage replicated the observed data well, but our conclusions about a causal mechanism are limited (see Winship and Morgan, 1999; Rubin, 2005; Pearl, 2014a, 2014b).”
Direction of effects in mediation analysis.
Wiedermann, W., & von Eye, A. (2015) · Psychological Methods · doi:10.1037/met0000027 · Cited by 87
Tofighi, D., & Kelley, K. (2019) · Multivariate Behavioral Research · doi:10.1080/00273171.2019.1618545 · Cited by 85
Hirt, M., Craig, G. C., Schäfer, S. A. K., Savre, J., & Heinze, R. (2020) · Quarterly Journal of the Royal Meteorological Society · doi:10.1002/qj.3788 · Cited by 84
“It is also possible to identify different indirect causal effects for nonlinear or non-parametric models (Pearl, 2014).”
BackgroundMethodInfluentialStrategic entrepreneurial behaviors: Construct and scale development
Anderson, B. S., Eshima, Y., & Hornsby, J. S. (2018) · Strategic Entrepreneurship Journal · doi:10.1002/sej.1306 · Cited by 87
“alternative explanations influencing the path from learning to SEBs and from SEBs to growth (Pearl, 2014).”
BackgroundInfluentialDirected Acyclic Graphs for Oral Disease Research
Akinkugbe, A. A., Sharma, S., Ohrbach, R. K., Slade, G. D., & Poole, C. (2016) · Journal of Dental Research · doi:10.1177/0022034516639920 · Cited by 82
“For more, refer to Robins and Greenland (1992), Pearl (2012, 2014), and VanderWeele (2009, 2015).”
BackgroundPrenatal fine particulate matter exposure, placental DNA methylation changes, and fetal growth
Zhao, Y., Wang, P., Zhou, Y., Xia, B., Zhu, Q., Ge, W., Li, J., Shi, H., Xiao, X., & Zhang, Y. (2020) · Environment International · doi:10.1016/j.envint.2020.106313 · Cited by 78
Wang, C., Cai, J., Chen, R., Shi, J., Yang, C., Li, H., Lin, Z., Meng, X., Liu, C., Niu, Y., et al. (2017) · Environmental Pollution · doi:10.1016/j.envpol.2017.02.068 · Cited by 72
A Tutorial in Bayesian Potential Outcomes Mediation Analysis
Miočević, M., González, O., Valente, M. J., & MacKinnon, D. P. (2017) · Structural Equation Modeling A Multidisciplinary Journal · doi:10.1080/10705511.2017.1342541 · Cited by 79
“…in the potential outcomes framework will have a causal interpretation that does not depend on the statistical models used for the mediator or the outcome (i.e., nonparametric identi fi cation of effects; Imai et al., 2010; Pearl, 2001, 2014; VanderWeele, 2015; VanderWeele & Vansteelandt, 2009).”
BackgroundMethodMi, Y., Burnham, K. L., Charles, P. D., Heilig, R., Vendrell, I., Whalley, J. P., Torrance, H. D. T., Antcliffe, D. B., May, S. M., Neville, M. J., et al. (2024) · Science Translational Medicine · doi:10.1126/scitranslmed.adh0185 · Cited by 91
Ethnic and social class discrimination in education: Experimental evidence from Germany
Wenz, S. E., & Hoenig, K. (2019) · Research in Social Stratification and Mobility · doi:10.1016/j.rssm.2019.100461 · Cited by 90
Wang, C., Chen, R., Shi, M., Cai, J., Shi, J., Yang, C., Li, H., Lin, Z., Meng, X., Liu, C., et al. (2017) · American Journal of Epidemiology · doi:10.1093/aje/kwx277 · Cited by 66
Moreno‐Navarrete, J. M., Blasco, G., Puig, J., Biarnés, C., Rivero, M., Gich, J., Fernández‐Aranda, F., Garre‐Olmo, J., Ramió‐Torrentà, L., Alberich‐Bayarri, Á., et al. (2017) · International Journal of Obesity · doi:10.1038/ijo.2017.162 · Cited by 65
Song, X. (2016) · Demography · doi:10.1007/s13524-016-0515-5 · Cited by 66
“Borrowing terms from causal mediation analysis (Baron and Kenny 1986; Pearl 2014), I distinguish between two effects that explain the role of family structure in three-generation social mobility: mediation and moderation, the latter of which is the focus of the present study.”
MethodRø, A. D. B., Simpson, M. R., Rø, T. B., Storrø, O., Johnsen, R., Videm, V., & Øien, T. (2017) · Clinical & Experimental Allergy · doi:10.1111/cea.12930 · Cited by 71
“Recent methods of mediation analysis use a counterfactual framework and allow for the natural variation of the level of the mediator between the subjects.(34) The natural direct effect (NDE) is then defined as the effect of the exposure on the outcome if the mediator is fixed to the level it would be in the absence of…”
MethodLord’s Paradox Revisited – (Oh Lord! Kumbaya!)
Pearl, J. (2016) · Journal of Causal Inference · doi:10.1515/jci-2016-0021 · Cited by 73
“These decision-theoretic considerations have indeed been cited as the core of causal mediation analysis (Pearl, 2001, 2014), where the value of estimating the indirect effect is tied to our ability to suppress it (or suppress the direct effect).”
BackgroundMethodSelf-citationStress-sensitive inference of task controllability
Ligneul, R., Mainen, Z. F., Ly, V., & Cools, R. (2022) · Nature Human Behaviour · doi:10.1038/s41562-022-01306-w · Cited by 60
“In the vocabulary of causal mediation analysis 27 , positive TE values entail the existence of a nat-ural direct effect linking actions to future states of the environment (Supplementary Note 2).”
TRYGVE HAAVELMO AND THE EMERGENCE OF CAUSAL CALCULUS
Pearl, J. (2014) · Econometric Theory · doi:10.1017/s0266466614000231 · Cited by 79
“Finally, mediation and moderation effects in nonlinear parametric systems (Pearl, 2014) and attribution problems in monotonic systems (Pearl, 2009a, Ch.”
Self-citationLakerveld, J., Palmeira, A. L., van Duinkerken, E., Whitelock, V., Peyrot, M. F., & Nouwen, A. (2020) · Diabetic Medicine · doi:10.1111/dme.14228 · Cited by 77
Instrumental Variables and Causal Mechanisms: Unpacking The Effect of Trade on Workers and Voters
Dippel, C., Gold, R., Heblich, S., & Pinto, R. (2017) · doi:10.3386/w23209 · Cited by 59
Causal Mediation Analysis: Warning! Assumptions Ahead
Keele, L. J. (2015) · American Journal of Evaluation · doi:10.1177/1098214015594689 · Cited by 81
“Readers should see Pearl (2014) for a more complete discussion of the connections between the two frameworks.”
BackgroundInfluentialReligiousness, Spirituality, and Substance Use: A Genetically Sensitive Examination and Critique
Nedelec, J. L., Richardson, G. B., & Silver, I. A. (2017) · Journal of Drug Issues · doi:10.1177/0022042617693382 · Cited by 60
“It is important to note that although random assignment to treatments like 12-step facilitation provides assurance that omitted variables do not confound the local average effects between AA participation and substance abuse outcomes, similar assurance does not extend to the effects between mediators of these effects…”
BackgroundInfluentialBody Mass Index, Adverse Pregnancy Outcomes, and Cardiovascular Disease Risk
Khan, S. S., Petito, L. C., Huang, X., Harrington, K. A., McNeil, R. B., Bello, N. A., Merz, C. N. B., Miller, E. C., Ravi, R., Scifres, C., et al. (2023) · Circulation Research · doi:10.1161/circresaha.123.322762 · Cited by 64
“41,42 This approach clarifies the assumptions needed for a causal interpretation, namely that there is no unmeasured confounding of the exposure-outcome, mediator-outcome, or exposure-mediator relationships, and that none of the mediator-outcome confounders are affected by the exposure.”
BackgroundWhy Do Tougher Caseworkers Increase Employment? The Role of Program Assignment as a Causal Mechanism
Huber, M., Lechner, M., & Mellace, G. (2016) · The Review of Economics and Statistics · doi:10.1162/rest_a_00632 · Cited by 68
“Recently, Robins (2003), Petersen, Sinisi, and Laan (2006), VanderWeele (2009), Imai, Keele, and Yamamoto (2010), Albert and Nelson (2011), and Imai and Yamamoto (2013), among others, consider rather general mediation models under sequential conditional independence of the treatment and the mediator.”
BackgroundA Framework for Testing Causality in Personality Research
Costantini, G., & Perugini, M. (2018) · European Journal of Personality · doi:10.1002/per.2150 · Cited by 60
“…possible fourth step to disentangle these possibilities could involve inspecting whether the causal effects of experimental conditions on trait-relevant outcomes are mediated by changes in personality traits or whether the causal conditions affect outcomes independent of traits (Pearl, 2014).”
BackgroundLouie, P., & Wheaton, B. (2019) · Journal of Health and Social Behavior · doi:10.1177/0022146519845069 · Cited by 66
Ogawa, K., Matsushima, S., Urayama, K. Y., Kikuchi, N., Nakamura, N., Tanigaki, S., Sago, H., Satoh, S., Saito, S., & Morisaki, N. (2019) · Scientific Reports · doi:10.1038/s41598-019-38999-5 · Cited by 69
Johnson, T. M., Sayles, H. R., Baker, J. F., George, M., Roul, P., Zheng, C., Sauer, B. C., Liao, K. P., Anderson, D. R., Mikuls, T. R., et al. (2021) · Annals of the Rheumatic Diseases · doi:10.1136/annrheumdis-2021-220125 · Cited by 51
The Blessings of Unlabeled Background in Untrimmed Videos
Liu, Y., Chen, J., Chen, Z., Deng, B., Huang, J., & Zhang, H. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.00611 · Cited by 58
“Ideally, if we can observe all the possible confounders, we can estimate the average effect of the video evenly associated with each confounder [27, 45], mimicking the physical intervention such as “recording” the videos of LONGJUMP everywhere.”
BackgroundContextual Debiasing for Visual Recognition with Causal Mechanisms
Liu, R., Liu, H., Li, G., Hou, H., Yu, T., & Yang, T. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.01242 · Cited by 70
“Causal inference [28,29] is aimed at pursuing the causal effect, which has been seen in medical, political and psychological research for years.”
BackgroundMethodMetzler‐Baddeley, C., Mole, J. P., Sims, R., Fasano, F., Evans, J., Jones, D. K., Aggleton, J. P., & Baddeley, R. (2019) · Scientific Reports · doi:10.1038/s41598-018-37658-5 · Cited by 54
Causal Disentangled Graph Neural Network for Fault Diagnosis of Complex Industrial Process
Liu, R., Zhang, Q., Lin, D., Zhang, W., & Ding, S. X. (2024) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2024.3452246 · Cited by 48
“Considering each sensor as a node, this implicit network can be conceptualized as an explicit graph structure.”
Strøm, I. F., Aakvaag, H. F., Birkeland, M. S., Felix, E. D., & Thoresen, S. (2018) · European Journal of Psychotraumatology · doi:10.1080/20008198.2017.1418570 · Cited by 66
“The total effect of an independent variable on an outcome can be decomposed into two components: the pure natural direct effect (PNDE) and the total natural indirect effect (TNIE) (Muthén & Asparouhov, 2015; Pearl, 2014).”
BackgroundMustieles, V., & Arrebola, J. P. (2020) · Journal of Epidemiology & Community Health · doi:10.1136/jech-2019-213181 · Cited by 49
Powell, S. R., Berry, K. A., Fall, A., Roberts, G. J., Fuchs, L. S., & Barnes, M. A. (2020) · Journal of Educational Psychology · doi:10.1037/edu0000513 · Cited by 46
“However, given the model’s strong theoretical support and following the logic of Pearl (2009, 2014) and others (Imai et al., 2010), it is reasonable to conceptualize the indirect effect as a “unitary” parameter to which students are randomized, allowing for causal inferences across the arc of…”
BackgroundOn Muddled Reasoning and Mediation Modeling
Grice, J. W., Cohn, A. M., Ramsey, R. R., & Chaney, J. M. (2015) · Basic and Applied Social Psychology · doi:10.1080/01973533.2015.1049350 · Cited by 48
“More recent methods based on counterfactual reasoning similarly revolve around parameter estimation and efforts to partition variable relationships into direct and indirect statistical effects (e.g., Imai, Keele, & Tingley, 2010; Pearl, 2014).”
MethodNguyen, T. Q., Schmid, I. T., Ogburn, E. L., & Stuart, E. A. (2022) · Journal of Causal Inference · doi:10.1515/jci-2021-0049 · Cited by 70
Graphical Tools for Linear Structural Equation Modeling
Chen, B., & Pearl, J. (2014) · Defense Technical Information Center (DTIC) · doi:10.21236/ada609131 · Cited by 57
Self-citationLiu, R., Xie, Y., Lin, D., Zhang, W., & Ding, S. X. (2024) · Reliability Engineering & System Safety · doi:10.1016/j.ress.2024.110468 · Cited by 52
Rudolph, K. E., Goin, D. E., Paksarian, D., Crowder, R., Merikangas, K. R., & Stuart, E. A. (2018) · American Journal of Epidemiology · doi:10.1093/aje/kwy248 · Cited by 59
Tereno, S., Madigan, S. L., Lyons‐Ruth, K., Plamondon, A., Atkinson, L. R., Guédeney, N., Greacen, T., Dugravier, R., Saïas, T., & Guédeney, A. (2017) · Development and Psychopathology · doi:10.1017/s0954579417000232 · Cited by 56
“nitions, the total effect of the intervention on the outcome can be decomposed into two components (Muthén & Asparouhov, 2015; Pearl, 2014).”
BackgroundInfluentialHigh-dimensional mediation analysis in survival models
Luo, C., Fa, B., Yan, Y., Wang, Y., Zhou, Y., Zhang, Y., & Yu, Z. (2020) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1007768 · Cited by 50
“Specially, the graphical model of causal mediation models can be illustrated by using a directed acyclic graph (DAG) [4].”
Background
Metadata
- Authors
- Pearl, J.
- Affiliations
- University of California, Los Angeles (US)
- Corresponding author
- Pearl, Judea
- Venue
- Psychological Methods
- Published
- 2014
- Volume
- 19
- Issue
- 4
- Pages
- 459-481
- ISSN
- 1082-989X (ISSN-L)
- 1939-1463
- OpenAlex
- W1967742189
- Cited by
- 554
- References
- 84
- Citations by year
- 2026: 45 · 2025: 52 · 2024: 49 · 2023: 49 · 2022: 56 · 2021: 57 · 2020: 87 · 2019: 38 · 2018: 32 · 2017: 33 · 2016: 25 · 2015: 21 · 2014: 10
- FWCI
- 31.34
- Type
- journal article
- Language
- en
- Open access
- green
- autopapers.ssrn.com
- Locations
- Psychological Methods (publishedVersion)
- PubMed (publishedVersion)
- SSRN Electronic Journal (submittedVersion)
- Funders
- National Institutes of Health (1R01 LM009961-01)
- National Science Foundation (IIS-0914211 and IIS-1018922)
- Office of Naval Research (N000-14-09-1-0665 and N00014-10-1-0933)
- Topics
- Physical Sciences › Mathematics › Statistics and Probability › Advanced Causal Inference Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Bayesian Modeling and Causal Inference
- Social Sciences › Arts and Humanities › History and Philosophy of Science › Philosophy of Biology and Scientific Explanation
- Keywords
- causal mediation analysis, natural direct and indirect effects, identifiability conditions, observational studies, confounding, parametric models, interaction, probabilistic graphical models
- MeSH terms
- Humans, Statistics as Topic, Models, Statistical, Causality
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.