Table 4. Relative mean bias percentage, RMSE and coverage probability when N = 30000, m = 30, β ₁ = 0.1, β ₂ = 0.2, ρ _b = 0.6, ρ _w = 0.3.

	Summary		Effects						Heterogeneity				Correlation
	data		% Bias		RMSE b		Coverage		% Bias		RMSE b		% Bias	%(${\widehat{\mathit{\rho }}}_{\mathit{b}})$
Method	scenario	I 2, a	${\widehat{\mathit{\beta }}}_1$	${\widehat{\mathit{\beta }}}_2$	${\widehat{\mathit{\beta }}}_1$	${\widehat{\mathit{\beta }}}_2$	${\widehat{\mathit{\beta }}}_1$	${\widehat{\mathit{\beta }}}_2$	${\widehat{\mathit{\tau }}}_1^2$	${\widehat{\mathit{\tau }}}_2^2$	${\widehat{\mathit{\tau }}}_1^2$	${\widehat{\mathit{\tau }}}_2^2$	(${\widehat{\mathit{\rho }}}_{\mathit{b}}$)	-1	+1	|1|
MV	COM	50%	0	0	0	0	95.1	95.2	0	3	0	0	1	0.7	9.4	10.1
UV	COM		0	0	0	0	95.1	95.3	0	3	0	0	—	—	—	—
MV	COM c		0	0	0	0	95.3	95.4	6	6	0	0	40	0.1	39.5	39.6
MV	MAR		0	0	0	-2	95.1	94.8	0	3	0	0	-2	1.9	16.3	18.2
UV	MAR		0	0	0	0	95.1	95.3	0	3	0	0	—	—	—	—
MV	MIF		0	27	0	-12	95.2	57.8	0	-31	0	-4	-9	6.4	26.8	33.2
UV	MIF		0	32	0	0	95.1	49.7	0	-33	0	0	—	—	—	—
MV	COM	75%	0	0	0	0	94.8	94.5	1	0	0	0	-2	0	0.1	0.1
UV	COM		0	0	0	0	94.9	94.4	1	0	0	0	—	—	—	—
MV	COM c		0	0	0	0	94.9	94.4	1	1	0	0	15	0	1.7	1.7
MV	MAR		0	0	0	-3	94.8	94.1	1	1	0	2	-2	0	0.7	0.7
UV	MAR		0	0	0	0	94.9	94.5	1	1	0	0	—	—	—	—
MV	MIF		0	26	0	-16	94.8	76.1	1	6	0	-4	-1	0.1	1	1.1
UV	MIF		0	36	0	0	94.9	69.9	1	6	0	0	—	—	—	—

Abbreviation: RMSE, root mean square error; CI, confidence interval; MV, multivariate meta-analysis; UV, Univariate meta-analysis; COM, complete data scenario; MAR, end point 2 missing at random for 30% studies; MAR, end point 2 missing informatively for 30% studies.

^aThe between-study variances for both end-points ${\tau }_j^2, j = \mathrm{1,2}$ are: ${\tau }_j^2$ = 0.0036 for I ² = 50%, ${\tau }_j^2$ = 0.0109 for I ² = 75%.

^bRMSE of estimates by MV method are expressed as % smaller (-) or larger (+) of corresponding estimates by UV method.

^{c${\widehat{\rho }}_w\mathrm{ }$} ignored.