- By ICR Secretariat
- Posted Wednesday, June 7, 2017
Dozens of recent clinical trials contain false data
Dozens of clinical trial reports likely contain inaccurate data. The information comes from John Carlisle, who used statistical tests to determine where sampling may have not been random. Essentially, Carlisle’s method identiﬁes papers in which the baseline characteristics (e.g. age, weight) exhibit either too narrow or too wide a distribution than expected by chance, resulting in an excess of p values close to either one or zero. This may indicate data not actually drawn from patients, but rather fabricated by the papers authors. Setting a p-value of 0.0001 for a 'find', Carlisle found that 2% of the 5000 papers he analysed were hits.
You can read the complete study here.