PLOS Biology
Published April 9, 2019 | 14:00 EDT (News release from PLOS)
A coin toss has exposed a problem that may affect many biomedicine studies. In many of these studies, small sample sizes and frequent loose thresholds for accepting statistical significance can lead to high rates of false-positive results and overestimation of true effects. Contrary to expectation, replication may not add more evidence compared to what could be gained from tossing a coin.
Canadian co-author: Robert Nadon, McGill University – robert.nadon@mcgill.ca
