Multilevel mediation analysis: The effects of omitted variables in the 1-1-1 model

Davood Tofighi, Stephen West, David Mackinnon

Research output: Contribution to journalArticle

22 Citations (Scopus)

Abstract

Multilevel mediation analysis examines the indirect effect of an independent variable on an outcome achieved by targeting and changing an intervening variable in clustered data. We study analytically and through simulation the effects of an omitted variable at level 2 on a 1-1-1 mediation model for a randomized experiment conducted within clusters in which the treatment, mediator, and outcome are all measured at level 1. When the residuals in the equations for the mediator and the outcome variables are fully orthogonal, the two methods of calculating the indirect effect (ab, c - c′) are equivalent at the between- and within-cluster levels. Omitting a variable at level 2 changes the interpretation of the indirect effect and will induce correlations between the random intercepts or random slopes. The equality of within-cluster ab and c - c′ no longer holds. Correlation between random slopes implies that the within-cluster indirect effect is conditional, interpretable at the grand mean level of the omitted variable.

Original languageEnglish (US)
Pages (from-to)290-307
Number of pages18
JournalBritish Journal of Mathematical and Statistical Psychology
Volume66
Issue number2
DOIs
StatePublished - May 2013

Fingerprint

Multilevel Analysis
Mediation
Mediator
Slope
Model
Randomized Experiments
Clustered Data
Intercept
Equality
Imply
Simulation

ASJC Scopus subject areas

  • Psychology(all)
  • Statistics and Probability
  • Arts and Humanities (miscellaneous)

Cite this

@article{933ede311be945938376e9602293e9a4,
title = "Multilevel mediation analysis: The effects of omitted variables in the 1-1-1 model",
abstract = "Multilevel mediation analysis examines the indirect effect of an independent variable on an outcome achieved by targeting and changing an intervening variable in clustered data. We study analytically and through simulation the effects of an omitted variable at level 2 on a 1-1-1 mediation model for a randomized experiment conducted within clusters in which the treatment, mediator, and outcome are all measured at level 1. When the residuals in the equations for the mediator and the outcome variables are fully orthogonal, the two methods of calculating the indirect effect (ab, c - c′) are equivalent at the between- and within-cluster levels. Omitting a variable at level 2 changes the interpretation of the indirect effect and will induce correlations between the random intercepts or random slopes. The equality of within-cluster ab and c - c′ no longer holds. Correlation between random slopes implies that the within-cluster indirect effect is conditional, interpretable at the grand mean level of the omitted variable.",
author = "Davood Tofighi and Stephen West and David Mackinnon",
year = "2013",
month = "5",
doi = "10.1111/j.2044-8317.2012.02051.x",
language = "English (US)",
volume = "66",
pages = "290--307",
journal = "British Journal of Mathematical and Statistical Psychology",
issn = "0007-1102",
publisher = "Wiley-Blackwell",
number = "2",

}

TY - JOUR

T1 - Multilevel mediation analysis

T2 - The effects of omitted variables in the 1-1-1 model

AU - Tofighi, Davood

AU - West, Stephen

AU - Mackinnon, David

PY - 2013/5

Y1 - 2013/5

N2 - Multilevel mediation analysis examines the indirect effect of an independent variable on an outcome achieved by targeting and changing an intervening variable in clustered data. We study analytically and through simulation the effects of an omitted variable at level 2 on a 1-1-1 mediation model for a randomized experiment conducted within clusters in which the treatment, mediator, and outcome are all measured at level 1. When the residuals in the equations for the mediator and the outcome variables are fully orthogonal, the two methods of calculating the indirect effect (ab, c - c′) are equivalent at the between- and within-cluster levels. Omitting a variable at level 2 changes the interpretation of the indirect effect and will induce correlations between the random intercepts or random slopes. The equality of within-cluster ab and c - c′ no longer holds. Correlation between random slopes implies that the within-cluster indirect effect is conditional, interpretable at the grand mean level of the omitted variable.

AB - Multilevel mediation analysis examines the indirect effect of an independent variable on an outcome achieved by targeting and changing an intervening variable in clustered data. We study analytically and through simulation the effects of an omitted variable at level 2 on a 1-1-1 mediation model for a randomized experiment conducted within clusters in which the treatment, mediator, and outcome are all measured at level 1. When the residuals in the equations for the mediator and the outcome variables are fully orthogonal, the two methods of calculating the indirect effect (ab, c - c′) are equivalent at the between- and within-cluster levels. Omitting a variable at level 2 changes the interpretation of the indirect effect and will induce correlations between the random intercepts or random slopes. The equality of within-cluster ab and c - c′ no longer holds. Correlation between random slopes implies that the within-cluster indirect effect is conditional, interpretable at the grand mean level of the omitted variable.

UR - http://www.scopus.com/inward/record.url?scp=84876015325&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84876015325&partnerID=8YFLogxK

U2 - 10.1111/j.2044-8317.2012.02051.x

DO - 10.1111/j.2044-8317.2012.02051.x

M3 - Article

C2 - 22594884

AN - SCOPUS:84876015325

VL - 66

SP - 290

EP - 307

JO - British Journal of Mathematical and Statistical Psychology

JF - British Journal of Mathematical and Statistical Psychology

SN - 0007-1102

IS - 2

ER -