Language Bias in Neuroscience – Is the Tower of Babel located in Germany?

Objective: Language bias has been demonstrated in general medicine in the sense that German authors tend to publish randomised controlled trials (RCTs) with non statistically significant findings in German journals and RCTs with significant findings in international English-language journals. Method: To assess whether this is also a problem in the field of neuroscience we hand searched five German psychiatric journals for RCTs and matched them with English publications of the same key authors found on Medline. Results: Our main result was the number of pairs of studies showing discordance in the level of significance. In addition to the p value, we tested for differences in the quality of report, the design of the trial and the sample size. We found a total of 21 pairs for the analysis involving 12 discordant pairs. Trials with significant results show a trend towards publication in English but the findings did not achieve statistical significance. Conclusion: The reason for this inconclusive finding might have been a relatively low number of RCTs published by German neuroscientists. None of the other factors analyzed functioned as a predictor for publication in either language.