Bivariate random-effects meta-analysis and the estimation of between-study correlation

Riley, Richard D; Abrams, Keith R; Sutton, Alexander J; Lambert, Paul C; Thompson, John R

doi:10.1186/1471-2288-7-3

BMC Medical Research Methodology

Table 2 Scenarios used in the simulations based on equation (1)

From: Bivariate random-effects meta-analysis and the estimation of between-study correlation

	Pooled values		Between-study variances		Within- and between-study correlation		Within-study variation
Scenario	β ₁	β ₂	$τ_{1}^{2}$	$τ_{2}^{2}$	ρ _Wi	ρ _B	Median value of the $s_{i j}^{2}$ s:		Description
Complete data							n = 50	n = 5
(i)	0	2	0.25	0.25	0	0	0.254	0.147	Zero correlation; within-study variation similar to between-study variation
(ii)	0	2	0.25	0.25	0	0.8	0.254	0.147	No within-study correlation but high between-study correlation; within-study variation similar to between-study variation
(iii)	0	2	0.25	0.25	0.8	0	0.254	0.147	High within-study correlation but no between-study correlation; within-study variation similar to between-study variation
(iv)	0	2	0.25	0.25	0.8	0.8	0.254	0.147	High within- and between-study correlation; within-study variation similar to between-study variation
(v)	0	2	0.0025	0.0025	0.8	0.8	0.254	0.147	High within- and between-study correlation; within-study variation large relative to between-study variation
(vi)	0	2	0.0025	1.5	0.8	0.8	0.254	0.147	High within- and between-study correlation; within-study variation large (for endpoint 1) and small (for endpoint 2) relative to between-study variation
(vii)	0	2	1.5	1.5	0.8	0.8	0.254	0.147	High within- and between-study correlation; within-study variation small relative to between-study variation
Missing data							n = 50	n = 10
(viii)	0	2	1.5	1.5	0	0.8	0.244	0.183	No within-study correlation but high between-study correlation; within-study variance small relative to between-study variance
(ix)	0	2	0.25	0.25	0	0.8	0.244	0.183	No within-study correlation but high between-study correlation; within-study variance similar to between-study variance
(x)	0	2	1.5	1.5	0.8	0.8	0.244	0.183	High within- and between-study correlation; within-study variance small relative to between-study variance
(xi)	0	2	0.25	0.25	0.8	0.8	0.244	0.183	High within- and between-study correlation; within-study variance similar to between-study variance

Back to article page

ISSN: 1471-2288

Contact us

General enquiries: journalsubmissions@springernature.com