babel tower problem babel tower problem
Diese Seite wurde seit 9 Jahren inhaltlich nicht mehr aktualisiert. Unter Umständen ist sie nicht mehr aktuell.
A Tower of Babel bias can arise when results published in languages other than English are excluded from the analysis (Grégoire et al. 1995). This exclusion can be biasing because there is evidence that non-English speaking authors are reluctant to submit negative or nonsignificant results to English-language journals. The thinking is that if the results are strong, they will be submitted to good international (i.e., English-language) journals, but if the results are unimpressive they will be submitted to local (i.e., non-English-language) journals.Von Paul D. Ellis im Buch The Essential Guide to Effect Sizes (2010) im Text Minimizing bias in meta-analysis auf Seite 120
Evidence for the Tower of Babel bias was provided by Grégoire et al. (1995). These authors reviewed sixteen meta-analyses that had explicitly excluded non-English results. They then searched for non-English results that were relevant to the reviews. They found one paper (written in German and published in a Swiss journal) that, had it been included in the relevant meta-analysis, would have turned a nonsignificant result into a statistically significant conclusion. Grégoire et al. (1995) interpreted this as evidence that linguistic exclusion criteria can lead to biased analyses.Von Paul D. Ellis im Buch The Essential Guide to Effect Sizes (2010) im Text Minimizing bias in meta-analysis auf Seite 120