Skip to main content

Table 6 The bias and RMSE for the population average treatment effect using various balancing methods on the IHDP datasets. The quasi-rerandomization (QReR) reconstructs the rerandomization under the acceptance probability \(p_a=0.1\)

From: Quasi-rerandomization for observational studies

Method

Bias

RMSE

IPW

-0.853

1.143

PSM

0.065

0.231

FM

-0.101

0.295

EBAL

-0.028

0.184

SBW

-0.030

0.180

EBCW

-0.028

0.184

\(\textrm{QReR}_{\textrm{M}} (p_a=0.1)\)

-0.021

0.191