Cortex | 2019
The red thread in the maze
Abstract
Doing science can feel like running in a maze. If we base our research on false assumptions, we can easily go astray – just ask the phrenologists. Scientists must therefore validate their claims through direct replication of previous findings. But certainly, this mustn’t come at the expense of actually moving the field forward. Nobody wants to err aimlessly through a maze, but you also won’t find your way out, if you’re too scared to walk at all. Replication should be the red thread we unroll behind us as we traverse this maze, lest those who follow us waste endless time and resources getting stuck in dead-ends. In this issue of Cortex, Huber et al. argue that the balance in cognitive neuroscience is still tipped too far towards novelty and narrative at the expense of scientific reliability. To encourage replication, and thus improve the reliability of published research, they suggest we must raise the bar for accepting novel results by mandating internal replication and preregistered designs. While I entirely agree with promoting replication in general, I nevertheless feel their proposals are somewhat unclear.