Statistical errors in clinical studies.

Abstract

Statistics are an integral part of any scientific paper. Unfortunately, statistical errors are common, which can falsely legitimize data. As Song et al1 noted “The inappropriate use of statistical analysis can lead to incorrect conclusions”. Paul Manske2 in his editorial entitled: Lies, Damn Lies and Statistics (quote by British Prime Minister Benjamin Disraeli, 1804–1881) noted that statistical analysis in scientific papers had acquired a shady reputation in part because it had been used to manipulate facts, rather than evaluate them which is embodied by saying that “ Statistics will prove anything, even truth (author unknown). Szabo3 noted that statistics is all about data analysis. Application of best statistical methodology to poor-quality data is analogous to claiming the operation was a success but patient died. Knowledge of appropriate statistical test to apply in any given situation is important, but so is ability to recognize common statistical errors.