Hoffmann O
Acta Neurochir (Wien). 1984;71(3-4):307-15. doi: 10.1007/BF01401325.
Out of 395 contributions in the volumes 51 to 64 of Acta Neurochirurgica 141 reported numerical data. The frequency of procedures used for describing distribution, location or variability and the frequency of analytical methods were evaluated. Critical assessment of these articles revealed a considerable rate of fundamental errors in application of statistics. Considering only papers involving analytical methods 52% of these (33 out of 64) contained simple statistical errors which can be avoided without any statistician's help. Some rules for improving statistical quality of medical articles are presented.