. Author manuscript; available in PMC: 2019 Jul 19.

Published in final edited form as: Stat Sci. 2018 May 3;33(2):261–276. doi: 10.1214/17-STS638

Table 2.

Simulation summary statistics of $\hat{β}$ , ${\hat{β}}_{PPL}$ and ${\hat{β}}_{PL}$

		$\hat{β}$				${\hat{β}}_{PPL}$		${\hat{β}}_{PL}$
n	Cen	Bias	SE	SEE	CP	Bias	SE	Bias	SE
					Scenario I
100	25%	(2,1)	(38,25)	(37,23)	(95,93)	(3,3)	(40,26)	(3,3)	(45,28)
	50%	(11,4)	(41,28)	(40,26)	(94,92)	(3,4)	(50,31)	(2,4)	(55,33)
400	25%	(2,0.3)	(18,12)	(18,11)	(95,94)	(1,1)	(19,12)	(0.5,1)	(21,13)
	50%	(6,2)	(21,14)	(21,13)	(93,93)	(1,1)	(23,14)	(0.3,1)	(25,16)
					Scenario II
100	25%	(0.3,2)	(35,23)	(33,22)	(94,94)	(5,4)	(39,25)	(5,3)	(46,29)
	50%	(6,0.4)	(37,26)	(35,23)	(92,93)	(9,8)	(53,32)	(5,3)	(55,35)
400	25%	(0.3,0.1)	(16,11)	(16,11)	(95,95)	(3,2)	(18,11)	(1,1)	(21,14)
	50%	(3,1)	(18,12)	(18,11)	(94,95)	(5,4)	(24,15)	(1,0.4)	(26,16)

Note: Cen is the censoring rate; Bias is the empirical bias (×100); SE is the empirical standard error (×100); SEE is the empirical mean of the standard error estimates; CP is the empirical coverage probability (×100) of the 95% confidence interval. $\hat{β}$ is the proposed estimator; ${\hat{β}}_{PPL}$ is the pseudo partial likelihood estimator; ${\hat{β}}_{PL}$ is the partial likelihood estimator.