Skip to main content

Table 2 Simulation results with interval-censored data, including the estimated bias (Bias), the sample standard error (SSE) of the estimates, the average of the standard error estimates (SEE), and the 95% empirical coverage probability (CP)

From: A pairwise pseudo-likelihood approach for regression analysis of left-truncated failure time data with various types of censoring

   

Proposed method

 

CL method

 

Ignoring truncation

n

Par

True

Bias

SSE

SEE

CP

 

Bias

SSE

SEE

CP

 

Bias

SSE

SEE

CP

\(A^{*}\) follows the uniform distribution

                

100

\(\beta _1\)

1

0.057

0.253

0.258

96.1

 

0.100

0.302

0.262

89.5

 

0.212

0.290

0.323

93.3

 

\(\beta _2\)

1

0.079

0.408

0.411

95.2

 

0.114

0.510

0.427

87.4

 

0.214

0.471

0.451

88.2

 

\(\Lambda (0.4)\)

0.16

0.013

0.103

0.104

94.0

 

0.009

0.103

 

-0.076

0.052

 

\(\Lambda (0.8)\)

0.64

-0.053

0.186

0.184

93.9

 

-0.039

0.185

 

-0.236

0.113

 

\(\Lambda (1.2)\)

1.44

-0.033

0.306

0.301

92.1

 

-0.065

0.308

 

-0.431

0.230

300

\(\beta _1\)

1

0.025

0.134

0.132

93.9

 

0.042

0.155

0.150

93.1

 

0.154

0.149

0.209

90.3

 

\(\beta _2\)

1

0.016

0.212

0.214

95.3

 

0.032

0.249

0.244

93.2

 

0.169

0.250

0.240

83.4

 

\(\Lambda (0.4)\)

0.16

0.023

0.075

0.073

96.5

 

0.022

0.076

 

-0.068

0.034

 

\(\Lambda (0.8)\)

0.64

-0.006

0.131

0.132

94.2

 

0.002

0.127

 

-0.224

0.069

 

\(\Lambda (1.2)\)

1.44

-0.021

0.229

0.222

95.2

 

-0.017

0.225

 

-0.427

0.139

500

\(\beta _1\)

1

0.005

0.100

0.100

96.1

 

0.017

0.121

0.115

94.3

 

0.132

0.107

0.189

86.6

 

\(\beta _2\)

1

0.022

0.165

0.163

93.6

 

0.036

0.202

0.189

91.8

 

0.148

0.188

0.179

75.4

 

\(\Lambda (0.4)\)

0.16

0.019

0.059

0.055

93.9

 

0.018

0.063

 

-0.066

0.028

 

\(\Lambda (0.8)\)

0.64

-0.014

0.098

0.101

95.1

 

-0.008

0.097

 

-0.222

0.057

 

\(\Lambda (1.2)\)

1.44

-0.022

0.186

0.182

94.6

 

-0.019

0.189

 

-0.429

0.107

\(A^{*}\) follows the exponential distribution

                

100

\(\beta _1\)

1

0.084

0.250

0.266

96.8

 

0.115

0.292

0.259

88.9

 

0.164

0.273

0.309

94.2

 

\(\beta _2\)

1

0.084

0.411

0.428

96.3

 

0.125

0.484

0.424

88.9

 

0.162

0.449

0.437

90.9

 

\(\Lambda (0.4)\)

0.16

0.008

0.096

0.101

96.2

 

0.007

0.097

 

-0.062

0.056

 

\(\Lambda (0.8)\)

0.64

-0.046

0.178

0.174

93.6

 

-0.05

0.179

 

-0.169

0.134

 

\(\Lambda (1.2)\)

1.44

-0.043

0.289

0.303

93.5

 

-0.037

0.290

 

-0.267

0.272

300

\(\beta _1\)

1

0.030

0.136

0.135

94.3

 

0.046

0.152

0.145

92.5

 

0.126

0.145

0.179

91.4

 

\(\beta _2\)

1

0.019

0.223

0.221

94.9

 

0.034

0.254

0.239

93.0

 

0.136

0.248

0.218

82.1

 

\(\Lambda (0.4)\)

0.16

0.021

0.066

0.068

95.9

 

0.023

0.067

 

-0.056

0.035

 

\(\Lambda (0.8)\)

0.64

-0.012

0.117

0.116

94.4

 

-0.005

0.119

 

-0.169

0.079

 

\(\Lambda (1.2)\)

1.44

-0.027

0.206

0.211

95.1

 

-0.235

0.205

 

-0.288

0.161

500

\(\beta _1\)

1

0.014

0.105

0.102

94.5

 

0.024

0.116

0.111

93.5

 

0.100

0.110

0.187

90.4

 

\(\beta _2\)

1

0.017

0.168

0.168

95.0

 

0.027

0.191

0.185

94.2

 

0.103

0.190

0.185

82.7

 

\(\Lambda (0.4)\)

0.16

0.019

0.054

0.052

97.7

 

0.017

0.054

 

-0.055

0.027

 

\(\Lambda (0.8)\)

0.64

-0.006

0.096

0.099

96.5

 

-0.003

0.094

 

-0.169

0.063

 

\(\Lambda (1.2)\)

1.44

-0.002

0.184

0.184

95.1

 

-0.200

0.186

 

-0.288

0.131

  1. Note: “Proposed method” denotes the proposed pairwise pseudo-likelihood method, “CL method” denotes the conditional likelihood method, and “Ignoring truncation” denotes the NPMLE approach that ignores the existence of left truncation