How selection bias perpetuates low power in science

In which we study how statistical power and significance thresholding influences reproducibility by relating it to positive and negative predictive values.