Summary Data Meta-analysis


Menu location: Analysis_Meta-Analysis_Summary.


This function provides a substitute for a proper meta-analysis when only summary statistics (odds ratio, relative, risk difference - with confidence intervals or standard errors) are known from a group of related studies.


Please do not use this function if the raw data from original studies are available, as the pooled estimates will be much more precise when using the original raw data. Better alternatives to this function are odds ratio, relative risk or risk difference meta-analysis based on counts from the included studies.


A generalised meta-analysis is calculated for the fixed effects model and the random effects model. Stratum weights are calculated as the inverse of the variance for the summary statistic (Y) supplied. If Y is specified as a ratio then calculations are performed on a natural log transformation of Y. The pooled estimate of Y is calculated as a weighted mean, i.e. the sum of weighted Y or ln(Y) for each stratum divided by the sum of the weights, transformed back to a natural scale if necessary.


The inconsistency of results across studies is summarised in the I² statistic, which is the percentage of variation across studies that is due to heterogeneity rather than chance – see the heterogeneity section for more information.


Please consult with a Statistician before using this method.



Copyright © 2000-2016 StatsDirect Limited, all rights reserved. Download a free trial here.