Advanced search
1 file | 324.01 KB

Estimation of controlled direct effects in the presence of exposure-induced confounding and latent variables

Tom Loeys (UGent) , Beatrijs Moerkerke (UGent) , An Raes, Yves Rosseel (UGent) and Stijn Vansteelandt (UGent)
Author
Organization
Abstract
Estimation of the direct effect of an exposure on an outcome requires adjustment for confounders of the exposure-outcome and mediator-outcome relationships. When some of the latter confounders have been affected by the exposure, then standard regression adjustment is prone to possibly severe bias. The use of inverse probability weighting under so-called marginal structural models has recently been suggested as a solution in the psychological literature. In this article, we show how progress can alternatively be made via G-estimation. We show that this estimation method can be easily embedded within the structural equation modeling framework and could in particular be used for estimating direct effects in the presence of latent variables. Moreover, by avoiding inverse probability weighting, it accommodates the typical problem of unstable weights in the alternative estimation approaches based on marginal structural models. We illustrate the approach both by simulations and by the analysis of a longitudinal study in individiduals who ended a romantic relationship. In this example we explore whether the effect of attachment anxiety during the relationship on mental distress 2 years after the breakup is mediated by rumination or not.
Keywords
SCORES, THINKING, OUTCOMES, VALIDITY, CAUSAL INFERENCE, ATTACHMENT, MEDIATION ANALYSIS, MARGINAL STRUCTURAL MODELS, mediation, Author Keywords:G-estimation, inverse probability weighting

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 324.01 KB

Citation

Please use this url to cite or link to this publication:

Chicago
Loeys, Tom, Beatrijs Moerkerke, An Raes, Yves Rosseel, and Stijn Vansteelandt. 2014. “Estimation of Controlled Direct Effects in the Presence of Exposure-induced Confounding and Latent Variables.” Structural Equation Modeling-a Multidisciplinary Journal 21 (3): 396–407.
APA
Loeys, T., Moerkerke, B., Raes, A., Rosseel, Y., & Vansteelandt, S. (2014). Estimation of controlled direct effects in the presence of exposure-induced confounding and latent variables. STRUCTURAL EQUATION MODELING-A MULTIDISCIPLINARY JOURNAL, 21(3), 396–407.
Vancouver
1.
Loeys T, Moerkerke B, Raes A, Rosseel Y, Vansteelandt S. Estimation of controlled direct effects in the presence of exposure-induced confounding and latent variables. STRUCTURAL EQUATION MODELING-A MULTIDISCIPLINARY JOURNAL. 2014;21(3):396–407.
MLA
Loeys, Tom, Beatrijs Moerkerke, An Raes, et al. “Estimation of Controlled Direct Effects in the Presence of Exposure-induced Confounding and Latent Variables.” STRUCTURAL EQUATION MODELING-A MULTIDISCIPLINARY JOURNAL 21.3 (2014): 396–407. Print.
@article{5756197,
  abstract     = {Estimation of the direct effect of an exposure on an outcome requires adjustment for confounders of the exposure-outcome and mediator-outcome relationships. When some of the latter confounders have been affected by the exposure, then standard regression adjustment is prone to possibly severe bias. The use of inverse probability weighting under so-called marginal structural models has recently been suggested as a solution in the psychological literature. In this article, we show how progress can alternatively be made via G-estimation. We show that this estimation method can be easily embedded within the structural equation modeling framework and could in particular be used for estimating direct effects in the presence of latent variables. Moreover, by avoiding inverse probability weighting, it accommodates the typical problem of unstable weights in the alternative estimation approaches based on marginal structural models. We illustrate the approach both by simulations and by the analysis of a longitudinal study in individiduals who ended a romantic relationship. In this example we explore whether the effect of attachment anxiety during the relationship on mental distress 2 years after the breakup is mediated by rumination or not.},
  author       = {Loeys, Tom and Moerkerke, Beatrijs and Raes, An and Rosseel, Yves and Vansteelandt, Stijn},
  issn         = {1070-5511},
  journal      = {STRUCTURAL EQUATION MODELING-A MULTIDISCIPLINARY JOURNAL},
  language     = {eng},
  number       = {3},
  pages        = {396--407},
  title        = {Estimation of controlled direct effects in the presence of exposure-induced confounding and latent variables},
  url          = {http://dx.doi.org/10.1080/10705511.2014.915372},
  volume       = {21},
  year         = {2014},
}

Altmetric
View in Altmetric
Web of Science
Times cited: