We therefore now completely decoupled the Fisher transformation across the various tabs, which required us to include some additional arguments in the Calculation tab. In particular, if studies did not report sufficient information for this transformation (number of observations and number of predictors), they were discarded in the subsequent analysis based on the selection of the Fisher transformation in the Forest plot tab, irrespective of whether the Fisher transformation was selected for use in that subsequent analysis. We noticed that using the Fisher transformation for the main Forest Plot analysis could trigger exclusion of some studies in subgroup, moderator, and publication bias analysis. We did some more substantial work on Workbook 6 for partial correlations, all related to the Fisher's r-to-z transformation.In the standardized residuals histogram (publication bias tab), the labels of the bins were not always correctly displayed in the plot.We noticed that in some studies that used Meta-Essentials for publication-bias analysis, the method was used incorrectly (even though they appeared in peer-reviewed journals). The trim-and-fill plot now automatically searches on the correct side of the combined effect size for potential asymmetry.
In version 1.4 we addressed several issues:.This does not affect the performance or functionality of any of the workbooks.Some objects were renamed (S1 -> SD1_ S2 -> SD2_).In version 1.5, we fixed a single issue that affected the use of workbooks 3 and 4 in German-language Microsoft Excel.(For Google Chrome use the right mouse button to save the zip-file.) This downloads a zip-file containing the workbooks, designed for Microsoft Excel. Meta-Essentials: Workbooks for meta-analysis (Version 1.5).
0 kommentar(er)
