Advanced search
1 file | 431.80 KB

On model selection and model misspecification in causal inference

Author
Organization
Abstract
Standard variable selection procedures, primarily developed for the construction of outcome prediction models, are routinely applied when assessing exposure effects in observational studies. We argue that this tradition is sub-optimal and prone to yield bias in exposure effect estimators as well as their corresponding uncertainty estimators. We weigh the pros and cons of confounder-selection procedures and propose a procedure directly targeting the quality of the exposure effect estimator. We further demonstrate that certain strategies for inferring causal effects have the desirable features (a) of producing (approximately) valid confidence intervals, even when the confounder-selection process is ignored, and (b) of being robust against certain forms of misspecification of the association of confounders with both exposure and outcome.
Keywords
double robustness, confounder-selection, influential weights, model selection, model uncertainty, propensity score, causal inference, MARGINAL STRUCTURAL MODELS, FOCUSED INFORMATION CRITERION, VARIABLE SELECTION, CONFOUNDER-SELECTION, PROPENSITY SCORE, INCOMPLETE DATA, EPIDEMIOLOGIC RESEARCH, SEMIPARAMETRIC MODELS, RANDOMIZED-TRIALS, BIAS

Downloads

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

Citation

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

Chicago
Vansteelandt, Stijn, Maarten Bekaert, and Gerda Claeskens. 2012. “On Model Selection and Model Misspecification in Causal Inference.” Statistical Methods in Medical Research 21 (1): 7–30.
APA
Vansteelandt, S., Bekaert, M., & Claeskens, G. (2012). On model selection and model misspecification in causal inference. STATISTICAL METHODS IN MEDICAL RESEARCH, 21(1), 7–30.
Vancouver
1.
Vansteelandt S, Bekaert M, Claeskens G. On model selection and model misspecification in causal inference. STATISTICAL METHODS IN MEDICAL RESEARCH. 2012;21(1):7–30.
MLA
Vansteelandt, Stijn, Maarten Bekaert, and Gerda Claeskens. “On Model Selection and Model Misspecification in Causal Inference.” STATISTICAL METHODS IN MEDICAL RESEARCH 21.1 (2012): 7–30. Print.
@article{1078625,
  abstract     = {Standard variable selection procedures, primarily developed for the construction of outcome prediction models, are routinely applied when assessing exposure effects in observational studies. We argue that this tradition is sub-optimal and prone to yield bias in exposure effect estimators as well as their corresponding uncertainty estimators. We weigh the pros and cons of confounder-selection procedures and propose a procedure directly targeting the quality of the exposure effect estimator. We further demonstrate that certain strategies for inferring causal effects have the desirable features (a) of producing (approximately) valid confidence intervals, even when the confounder-selection process is ignored, and (b) of being robust against certain forms of misspecification of the association of confounders with both exposure and outcome.},
  author       = {Vansteelandt, Stijn and Bekaert, Maarten and Claeskens, Gerda},
  issn         = {0962-2802},
  journal      = {STATISTICAL METHODS IN MEDICAL RESEARCH},
  keyword      = {double robustness,confounder-selection,influential weights,model selection,model uncertainty,propensity score,causal inference,MARGINAL STRUCTURAL MODELS,FOCUSED INFORMATION CRITERION,VARIABLE SELECTION,CONFOUNDER-SELECTION,PROPENSITY SCORE,INCOMPLETE DATA,EPIDEMIOLOGIC RESEARCH,SEMIPARAMETRIC MODELS,RANDOMIZED-TRIALS,BIAS},
  language     = {eng},
  number       = {1},
  pages        = {7--30},
  title        = {On model selection and model misspecification in causal inference},
  url          = {http://dx.doi.org/10.1177/0962280210387717},
  volume       = {21},
  year         = {2012},
}

Altmetric
View in Altmetric
Web of Science
Times cited: