Table 1, Summary of Results With Linear Outcome Modelsa - Developing and Testing New Methods for Estimating Treatment Effectiveness in Observational Studies Using High-Dimension Data

Warning: The NCBI web site requires JavaScript to function. more...

Table 1Summary of Results With Linear Outcome Models^a

	(C1) cor PS, cor OR			(C2) cor PS, mis OR			(C3) mis PS, cor OR
	RML	RML2	RCAL	RML	RML2	RCAL	RML	RML2	RCAL
	n = 800 and p = 200
Bias	0.019	0.002	0.026	−0.038	−0.009	0.012	0.010	0.002	0.016
$\sqrt{V a r}$	0.070	0.079	0.070	0.072	0.092	0.071	0.070	0.076	0.070
$\sqrt{E V a r}$	0.068	0.076	0.068	0.072	0.094	0.069	0.069	0.075	0.068
Cov90	0.864	0.880	0.854	0.859	0.906	0.889	0.875	0.886	0.871
Cov95	0.922	0.947	0.914	0.914	0.950	0.938	0.941	0.950	0.942
	n = 800 and p = 1000
Bias	0.031	−0.002	0.033	−0.038	−0.001	0.015	0.026	0.009	0.028
$\sqrt{V a r}$	0.068	0.109	0.070	0.070	0.394	0.071	0.069	0.092	0.070
$\sqrt{E V a r}$	0.067	0.112	0.067	0.070	0.388	0.068	0.068	0.089	0.068
Cov90	0.851	0.911	0.836	0.842	0.896	0.877	0.867	0.892	0.854
Cov95	0.922	0.949	0.915	0.912	0.944	0.929	0.924	0.933	0.923

: Abbreviations: cor, correct; Cov90, coverage proportion of the 90% CI; Cov95, coverage proportion of the 95% CI; mis, misspecified; OR, outcome regression; PS, propensity score; Var, variance.
: Reproduced from Tan Z. Ann Statist. 2020;48:811-837.
: Reprinted with permission from Annals of Statistics (Copyright ©2020). Institute of Mathematical Statistics. All Rights Reserved.
: ^aRML denotes ${\hat{μ}}^{1} ({\hat{m}}_{RWL}^{1}, {\hat{π}}_{RML}^{1})$ , and RML2 denotes the variant with ${\hat{m}}_{RWL}^{1}$ and ${\hat{π}}_{RML}^{1}$ replaced by the fitted values obtained by refitting OR and PS models only including the variables selected from the corresponding Lasso estimation. RCAL denotes ${\hat{μ}}^{1} ({\hat{m}}_{RWL}^{1}, {\hat{π}}_{RCAL}^{1})$ . Bias and Var are the Monte Carlo bias and variance of the points estimates. EVar is the mean of the variance estimates, and hence $\sqrt{EVar}$ also measures the L₂-average of lengths of CIs.

From: Developing and Testing New Methods for Estimating Treatment Effectiveness in Observational Studies Using High-Dimension Data

Cover of Developing and Testing New Methods for Estimating Treatment Effectiveness in Observational Studies Using High-Dimension Data

Developing and Testing New Methods for Estimating Treatment Effectiveness in Observational Studies Using High-Dimension Data [Internet].

Tan Z, Gerhard T, Sun B.

Washington (DC): Patient-Centered Outcomes Research Institute (PCORI); 2021 Nov.

This book is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License which permits noncommercial use and distribution provided the original author(s) and source are credited. (See https://creativecommons.org/licenses/by-nc-nd/4.0/

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Table 1Summary of Results With Linear Outcome Modelsa

Table 1Summary of Results With Linear Outcome Models^a