Limits...
The use of bootstrapping when using propensity-score matching without replacement: a simulation study.

Austin PC, Small DS - Stat Med (2014)

Bottom Line: An important issue when using propensity-score matching is how to estimate the standard error of the estimated treatment effect.The second method involved drawing bootstrap samples from the original sample and estimating the propensity score separately in each bootstrap sample and creating a matched sample within each of these bootstrap samples.The former approach was found to result in estimates of the standard error that were closer to the empirical standard deviation of the sampling distribution of estimated effects.

View Article: PubMed Central - PubMed

Affiliation: Institute for Clinical Evaluative Sciences, Toronto, Canada; Institute of Health Management, Policy and Evaluation, University of Toronto, Toronto, Canada; Schulich Heart Research Program, Sunnybrook Research Institute, Toronto, Canada.

Show MeSH

Related in: MedlinePlus

Empirical versus estimated standard errors (true propensity score).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4260115&req=5

fig03: Empirical versus estimated standard errors (true propensity score).

Mentions: The corresponding results when we matched on the true propensity score are reported in Figures 3 and 4. In comparing Figures 1 and 3, one notes that there were minor differences in results when matching on the true propensity score compared with matching on the estimated propensity score. Several observations merit being highlighted. First, the naïve matched estimator of the standard error tended to result in the greatest overestimate of the standard deviation of the empirical sampling distribution of the estimated effect compared with the three other competing methods. Second, the other three variance estimation methods (the matched parametric estimator and the two bootstrap methods) tended to result in estimates of the sampling variability that were almost indistinguishable from one another. Third, in most scenarios, these three variance estimation methods (the matched parametric estimator and the two bootstrap methods) tended to more closely approximate the sampling distribution compared with what was observed when matching on the estimated propensity score was used.


The use of bootstrapping when using propensity-score matching without replacement: a simulation study.

Austin PC, Small DS - Stat Med (2014)

Empirical versus estimated standard errors (true propensity score).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4260115&req=5

fig03: Empirical versus estimated standard errors (true propensity score).
Mentions: The corresponding results when we matched on the true propensity score are reported in Figures 3 and 4. In comparing Figures 1 and 3, one notes that there were minor differences in results when matching on the true propensity score compared with matching on the estimated propensity score. Several observations merit being highlighted. First, the naïve matched estimator of the standard error tended to result in the greatest overestimate of the standard deviation of the empirical sampling distribution of the estimated effect compared with the three other competing methods. Second, the other three variance estimation methods (the matched parametric estimator and the two bootstrap methods) tended to result in estimates of the sampling variability that were almost indistinguishable from one another. Third, in most scenarios, these three variance estimation methods (the matched parametric estimator and the two bootstrap methods) tended to more closely approximate the sampling distribution compared with what was observed when matching on the estimated propensity score was used.

Bottom Line: An important issue when using propensity-score matching is how to estimate the standard error of the estimated treatment effect.The second method involved drawing bootstrap samples from the original sample and estimating the propensity score separately in each bootstrap sample and creating a matched sample within each of these bootstrap samples.The former approach was found to result in estimates of the standard error that were closer to the empirical standard deviation of the sampling distribution of estimated effects.

View Article: PubMed Central - PubMed

Affiliation: Institute for Clinical Evaluative Sciences, Toronto, Canada; Institute of Health Management, Policy and Evaluation, University of Toronto, Toronto, Canada; Schulich Heart Research Program, Sunnybrook Research Institute, Toronto, Canada.

Show MeSH
Related in: MedlinePlus