Table 2 

Methods available in MetaAnalyst 

Fixed 
Random 
Bayes 



IV* 
MH 
Peto 
DL 
EM^{†} 



Binary outcomes 

Odds ratio (OR) 
√ 
√ 
√ 
√ 
√ 
√^{‡} 
Risk ratio (RR) 
√ 
√ 
 
√ 
√ 
√^{‡} 
Risk difference (RD) 
√ 
√ 
 
√ 
√ 
√^{‡} 
Proportion** 
√ 
 
 
√ 
√ 
√^{‡} 
Continuous outcomes 
 
 



WMD 
√ 
 
 
√ 
√ 
√^{‡} 


Hedge's g 
√ 
 
 
√ 
√ 
√^{‡} 


Cohen's d 
√ 
 
 
√ 
√ 
√^{‡} 


Glass' δ 
√ 
 
 
√ 
√ 
√^{‡} 


Mean** 
√ 
 
 
√ 
√ 
√^{‡} 


Diagnostic test data 
 
 



Specificity 
√ 
 
 
√ 
√ 
√^{‡} 


Sensitivity 
√ 
 
 
√ 
√ 
√^{‡} 


Accuracy 
√ 
 
 
√ 
√ 
√^{‡} 


Positive predictive value (PPV) 
√ 
 
 
√ 
√ 
√^{‡} 


Negative predictive value (NPV) 
√ 
 
 
√ 
√ 
√^{‡} 


Positive likelihood ratio 
√ 
√ 
 
√ 
√ 
√^{‡} 


Positive likelihood ratio 
√ 
√ 
 
√ 
√ 
√^{‡} 


Diagnostic odds ratio 
√ 
√ 
 
√ 
√ 
√^{‡} 


Summary ROC curve 
[weighted, unweighted] 
[weighted] 



Bivariate 
 
 
√ 



Hierarchical SROC 
 
 
√ 



*Fixed effects metaregression using weighted least squares is available here if there is at least one numerical covariate. ^{†}Random effects metaregression using an expectation maximization approach is available here if there is at least one numerical covariate. ^{‡}Control rate metaregression (linear or quadratic) is possible here (with or without adjusting for additional covariates, as deemed appropriate). ** e.g., for the metaanalysis of data from single arm studies.  = not applicable, √ = available, DL: DerSimonian and Laird model, EM = Expectationmaximization, IV = inverse variance, MH: MantelHaenszel method, ROC = Receiver operating characteristic curve 

Wallace et al. BMC Medical Research Methodology 2009 9:80 doi:10.1186/14712288980 