Skip to main content

Table 4 Summary of empirical estimates of log-OR and relative bias (%), power (%) and MC standard error, coverage (%), model standard error and empirical standard error derived from χ2 method and GLMM test with compound symmetric variance-covariance matrix from 1000 simulations with a total sample size N = 776

From: Power difference in a χ2 test vs generalized linear mixed model in the presence of missing data – a simulation study

 

pi1 = 0.5, ρ = 0.7

Log-odds (% bias)

Power (%)c

Coverage (%)c

Model SEc

Empirical SEc

Complete

χ2 a

0.394 (0)

77.2

95.5

0.145

0.145

 GLMMb

0.624 (0)

78.7

97.2

0.234

0.208

MCAR

Equal drop out (30% drop out in each arm)

χ2

0.397 (0.7)

63.3

95.5

0.174

0.175

  GLMM

0.631 (1.1)

70.0

97.6

0.259

0.225

Unequal dropouts 1 (20% drop out in control, 40% drop out in treatment)

χ2

0.396 (0.5)

61.5

95.1

0.176

0.176

  GLMM

0.633 (1.4)

70.7

97.0

0.261

0.229

Unequal dropouts 2 (40% drop out in control, 20% drop out in treatment)

χ2

0.392 (−0.6)

61.3

96.0

0.175

0.170

  GLMM

0.628 (0.7)

71.5

97.7

0.260

0.220

MAR1 (observations more likely to be deleted if prior observation was positive)

Equal drop out (30% drop out in each arm)

χ2

0.453 (14.8)

74.3

93.2

0.173

0.173

  GLMM

0.672 (7.8)

77.5

96.9

0.258

0.226

Unequal dropouts 1 (20% drop out in control, 40% drop out in treatment)

χ2

0.257 (−34.9)

30.1

87.4

0.174

0.173

  GLMM

0.635 (1.7)

72.8

97.4

0.259

0.225

Unequal dropouts 2 (40% drop out in control, 20% drop out in treatment)

χ2

0.649 (64.6)

95.3

71.0

0.177

0.178

  GLMM

0.709 (13.7)

82.1

95.6

0.259

0.227

MAR2 (observations less likely to be deleted if prior observation was positive)

Equal drop out (30% drop out in each arm)

χ2

0.374 (−5.2)

54.7

94.9

0.179

0.177

  GLMM

0.586 (−6.1)

63.6

97.5

0.260

0.224

Unequal dropouts 1 (20% drop out in control, 40% drop out in treatment)

χ2

0.583 (47.9)

89.9

82.5

0.183

0.183

  GLMM

0.632 (1.2)

70.6

97.4

0.262

0.227

Unequal dropouts 2 (40% drop out in control, 20% drop out in treatment)

χ2

0.164 (−58.3)

14.7

75.7

0.180

0.177

  GLMM

0.541 (−13.3)

54.3

96.6

0.261

0.221

  1. aIndependent two-sample χ2 test for test of treatment effect at time point 3
  2. bGeneralized linear mixed model with compound symmetric variance-covariance matrix, contrast used to estimate treatment effect at time point 3
  3. cRange of Monte Carlo SE: Power (0.007–0.016); Coverage (0.005–0.014); Model SE (< 0.0001–0.001); Empirical SE (0.003–0.005)
Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us