Table 2.

Effect estimates (posterior medians and 95% credibility intervals) for the examples from Section 3 using different priors for the heterogeneity τ

Heterogeneity prior	Crins et al.	Miller et al.	Mozobil	Romiplostim	Krystexxa
$\tau \sim \phantom{}$half‐Normal(1.0)	0.16 (0.04, 0.78)	0.59 (0.18, 1.74)	5.34 (1.80, 15.95)	0.19 (0.04, 0.80)	7.14 (1.04, 49.15)
$\tau \sim \phantom{}$half‐Normal(0.5)	0.16 (0.05, 0.49)	0.60 (0.29, 1.14)	5.34 (2.73, 10.48)	0.19 (0.07, 0.53)	7.14 (1.39, 36.70)
$\tau \sim \phantom{}$log‐Normal(‐1.048,1.046)	0.16 (0.05, 0.58)	0.60 (0.27, 1.23)	5.34 (2.56, 11.19)	0.19 (0.06, 0.60)	7.14 (1.25, 40.80)
$\tau \sim \phantom{}$log‐Normal(‐1.741,1.046)	0.16 (0.06, 0.45)	0.60 (0.35, 0.99)	5.34 (3.18, 8.98)	0.19 (0.07, 0.49)	7.14 (1.46, 35.01)
${\tau }^2\sim \phantom{}$inv‐Gamma(0.5,0.0076)	0.16 (0.06, 0.47)	0.60 (0.35, 1.00)	5.34 (3.14, 9.08)	0.19 (0.07, 0.51)	7.14 (1.39, 36.57)
${\tau }^2\sim \phantom{}$inv‐Gamma(0.5,0.0019)	0.16 (0.06, 0.42)	0.61 (0.41, 0.90)	5.33 (3.45, 8.25)	0.19 (0.08, 0.46)	7.14 (1.49, 34.08)