Comparison of logistic regression versus propensity score when the number of events is low and there are multiple confounders.

By M. Cepeda, R. Boston, John T. Farrar, Brian L. Strom. American Journal of Epidemiology, 2003

855 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/3aa6363a161dedf15c65c1af109415271e90889b

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Economics and Econometrics

In Statistics and Probability

Citations in the corpus

Assessing the Effectiveness of a College Freshman Seminar Using Propensity Score Adjustments (2011, background): “Furthermore, there is evidence that suggests that using aggregate covariates, such as propensity scores, are more effective than single or multiple covariates in reducing selection bias ( Cepeda et al. 2003; Grunwald and Mayhew 2008; Peterson et al. 2003).”
A review of uses of health care utilization databases for epidemiologic research on therapeutics. (2005, positive): “Because exposure propensity scores model the relation of covariates and their interactions with the exposure and not directly with the study outcomes, such study configurations are ideal for applying EPS to avoid the risk of over-fitting if there were fewer than 10 outcomes per variable in a traditi”
Robust Tree-based Causal Inference for Complex Ad Effectiveness Analysis (2015, methodology): “In addition, when the outcome is binary, prior research suggested that at least 8 events should be observed for every covariate in the regression model [9], which prevents its application in practice [2].”
Propensity score estimation: machine learning and classification methods as alternatives to logistic regression (2010, background): “A propensity score model helps achieve this goal by estimating the probability of treatment, given individual covariates, such that conditioning on this probability (the propensity score) ensures that the treatment is independent of covariate patterns [2] and, in particular, by achieving balance on ”
Activated injectable vitamin D and hemodialysis survival: a historical cohort study. (2005, background): “propensity score analysis could not be performed (28).”
Inverse Probability Weighting with Missing Predictors of Treatment Assignment or Missingness (2014, positive): “Cepeda et al. (2003) and Stürmer et al. (2005) discuss the relation between, and relative advantages of, these two approaches.”
Shining a spotlight on small rural businesses: How does their performance compare with urban? (2019, methodology): “Simulations also indicate that PSM is both more robust and precise and has greater power than logistic regression (Cepeda et al., 2003; Glynn et al., 2006).”
Sarcopenia, intramuscular fat deposition, and visceral adiposity independently predict the outcomes of hepatocellular carcinoma. (2015, methodology): “Propensity scoring was used to control of selection bias and potential confounding [26].”
Performance of technology transfer offices in Germany (2011, positive): “As it has been shown by Cepeda et al. (2003) propensity scores are especially useful in correcting biases in small samples with relatively many confounders.”
Evaluating uses of data mining techniques in propensity score estimation: a simulation study (2008, result): “models.(21,22) In our stimulation study, although the magnitude of bias in LR (misspecified model for Scenario B–G) was not large, LR tended to produce more biased estimates than NN in Scenario B–G.”