Common statistical mistakes in published
articles in biomedical journals
Excerpts from the presentation by Dr. Farrokh Habibzadeh MD*
It is not uncommon to see unnecessary precision in published articles from the Middle East. Sometimes the use of words is inappropriate, results reported are contradictory and p value reported is wrong. Points worth consideration are:
* Do not imply greater precision than the measurement instrument.
* Use only one decimal place more than the basic unit of measurement when reporting statistics.
* Never start a new sentence with figures but with words.
For example: 15, 67.7% of patients who developed ES presented between 0-6 moths after surgery, 9, 34.9% presented between 6-12 months after surgery and 2, 7.7% presented 12-18 months postoperatively.
Instead it should be written as: Fifteen (67.7%) of patients who developed ES presented between 0-6 months, nine (34.9%) presented between 6-12 months after surgery and two (7.7%) presented 12-18 months postoperatively.
* Do not use average instead of “mean” and markedly instead of “significantly”
* It is always helpful if the services of a qualified statistician are acquired before finalizing the manuscript and its submission for publication.
* Director at large,
World Association of Medical Editors (WAME)
Chairman, Editorial Committee
Eastern Mediterranean Association of Medical Editors (EMAME)
Editorial Consultant, The Lancet
Director, Shiraz NIOC Medical Education and Research Center,
E mail: email@example.com