|
Meta
|
k
|
I
2
|
Equal- α
|
W-optimal
|
\(\alpha ^{*}_{2}\)
|
Width
|
|---|
|
Analysis
| | |
CI (τ)
|
interval (τ)
| |
Ratio
|
|---|
|
CERVIX3
|
5
|
56 %
|
(0, 1.287)
|
(0, 1.048)
|
0.050
|
0.814
|
|
NSCLC4
|
11
|
75 %
|
(0.227, 0.887)
|
(0.193, 0.824)
|
0.040
|
0.954
|
|
NSCLC1
|
17
|
45 %
|
(0.013, 0.426)
|
(0.028, 0.436)
|
0.021
|
0.986
|
|
CERVIX1
|
18
|
62 %
|
(0.201, 0.707)
|
(0.182, 0.678)
|
0.035
|
0.982
|
- I2 is the heterogeneity statistic of Higgins and Thompson [28]. In each case we show the equal tailed (α1=α2=0.025) 95 % confidence interval for τ, the W-optimal interval for τ, the value of \(\alpha ^{*}_{2}\) that provides the W-optimal interval (also for τ) and the ratio of the width of the W-optimal interval and the equal tailed confidence interval. In each case we see that there is reduction in the interval width by adopting α2>>α1
