Biases in the estimates of τ₂, and biases and SEs of the pooled estimates of <i>β</i>₂ from multivariate vs. univariate approaches by whether or not <i>ρ</i>_<i>b</i> is estimated at parameter boundary in 5000 replications in complete summary data scenario.

<p>Scenario: <math><mi>N</mi><mi> </mi><mo>=</mo><mi> </mi><mn>10000</mn><mo>,</mo><mi> </mi><mi>m</mi><mi> </mi><mo>=</mo><mi> </mi><mn>10</mn><mo>,</mo><mi> </mi><mi>M</mi><mi>A</mi><mi>F</mi><mi> </mi><mo>=</mo><mi> </mi><mn>0.20</mn><mo>,</mo><mi> </mi><msub><mrow><mi>β</mi></mrow><mrow><mn>1</mn></mrow></msub><mi> </mi><mo>=</mo><mi> </mi><mn>0.3</mn><mo>,</mo><mi> </mi><msub><mrow><mi>β</mi></mrow><mrow><mn>2</mn></mrow></msub><mi> </mi><mo>=</mo><mi> </mi><mn>0.4</mn><mo>,</mo><mi> </mi><msubsup><mrow><mi>τ</mi></mrow><mrow><mn>1</mn></mrow><mrow><mn>2</mn></mrow></msubsup><mi> </mi><mo>=</mo><mi> </mi><msubsup><mrow><mi>τ</mi></mrow><mrow><mn>2</mn></mrow><mrow><mn>2</mn></mrow></msubsup><mi> </mi><mo>=</mo><mi> </mi><mn>0.0033</mn><mo>;</mo><mi> </mi><mrow><mi>I</mi></mrow><mrow><mn>2</mn></mrow><mi> </mi><mo>=</mo><mi> </mi><mn>50</mn><mi>%</mi></math>, <i>ρ</i>_<i>b</i> = 0.75, <i>ρ</i>_<i>w</i> = 0.5. Symbols and abbreviations: <i>N</i>, total subjects; <i>m</i>, number of studies, <i>β</i>₂ and <i>τ</i>₂, average effect and between-study standard deviation of true study-wise effects for end point 2, respectively; <i>I</i>² = degree of between-study heterogeneity; <i>ρ</i>_<i>b</i> and <i>ρ</i>_<i>w</i>, true between-and within-study correlations, respectively; MAF, minor allele frequency; SE, standard error; MV, multivariate approach; UV, univariate approach.