Skip to main content

Table 3 Confidence interval width of the Q-profile and Jackson’s Generalised Q statistic approaches under equal, optimal and 0.01:0.04 split strategies; k=15

From: Confidence intervals for the between-study variance in random-effects meta-analysis using generalised heterogeneity statistics: should we use unequal tails?

τ 2

I 2

CI width

CI width ratio (wrt equal α)

  

Equal α

W-opt

0.01:0.04 split

  

Q-profile

Gen-Q

Q-profile

Gen-Q

Q-profile

Gen-Q

0.05

28.49

0.305

0.296

0.777

0.778

0.844

0.846

0.09

39.93

0.393

0.385

0.811

0.811

0.865

0.867

0.13

48.82

0.478

0.470

0.834

0.834

0.878

0.879

0.17

55.47

0.558

0.551

0.849

0.849

0.886

0.887

0.21

60.85

0.638

0.633

0.860

0.861

0.891

0.892

0.24

65.21

0.718

0.714

0.868

0.868

0.894

0.895

0.28

68.60

0.796

0.793

0.873

0.874

0.896

0.897

0.32

71.40

0.873

0.871

0.877

0.878

0.898

0.899

0.36

73.77

0.949

0.949

0.879

0.880

0.899

0.900

0.40

75.74

1.030

1.030

0.881

0.882

0.900

0.900