Meta-analysis, power analysis, and sensitivity analysis are widespread statistical techniques, which can be correctly performed only if variability statistics, such as standard deviation, are available; however, standard deviations are often missing in published articles. This work illustrates the functionality and the versatility of a newly developed Excel (c) tool for the standard deviation extraction from ANOVA and Multiple Comparison Test (MCT) results. The tool implements four methods, which can be alternatively applied according to the available statistics usually reported in ANOVA and/or MCT tables and graphs: 1) least significant difference (LSD), 2) significance level (p(F)), 3) letters for means separation assigned by MCT, 4) a range of significance level, indicated by "stars". The tool can be applied in one, two and three-way factorial experiments arranged in complete randomization, randomized block, split-plot or split-block. The performances of the different methods were tested in a case study about meta-analysis database preparation.
Acutis, M., Tadiello, T., Perego, A., Di Guardo, A., Schillaci, C., Valkama, E. (2022). EX-TRACT: An excel tool for the estimation of standard deviations from published articles. ENVIRONMENTAL MODELLING & SOFTWARE, 147(January 2022) [10.1016/j.envsoft.2021.105236].
EX-TRACT: An excel tool for the estimation of standard deviations from published articles
Di Guardo A.;
2022
Abstract
Meta-analysis, power analysis, and sensitivity analysis are widespread statistical techniques, which can be correctly performed only if variability statistics, such as standard deviation, are available; however, standard deviations are often missing in published articles. This work illustrates the functionality and the versatility of a newly developed Excel (c) tool for the standard deviation extraction from ANOVA and Multiple Comparison Test (MCT) results. The tool implements four methods, which can be alternatively applied according to the available statistics usually reported in ANOVA and/or MCT tables and graphs: 1) least significant difference (LSD), 2) significance level (p(F)), 3) letters for means separation assigned by MCT, 4) a range of significance level, indicated by "stars". The tool can be applied in one, two and three-way factorial experiments arranged in complete randomization, randomized block, split-plot or split-block. The performances of the different methods were tested in a case study about meta-analysis database preparation.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
