In this episode we are in conversation with biologist Dr Tim Errington, Director of Metascience for the Center for Open Science (COS), which is dedicated to the replication of scientific research. First we discuss the Reproducibility Project: Psychology, which found that of the 97% of significant results published in high-ranking psychological science journals only 36% of the replications were significant. We discuss some of the post-hoc explanations provided by researchers to explain this failure of replication, as well as the funding incentive system within science which discourages reproduction of research. We follow on from this by discussing the Reproducibility Project: Cancer Biology, which aimed to reproduce medical studies that were published in top-tier journals such as Nature, Science and Cell. So far, 3 of the 5 studies they have attempted to reproduce had strikingly different results from the originals. We discuss this in terms of the extremely high failure rate of phase 2 clinical trials, and how these drugs may not make it into human trials if the pre-clinical trials were more statistically and methodologically robust. We also talk about the general statistical illiteracy in the field as well as the extreme pressure to find positive results in science which is dominated by a culture of 'publish or perish'.

Image courtesy of www.news.virginia.edu