Reproducibility in science is a prominent topic in both lay and scientific press. But a new facet of this discussion has arisen in a recent comparison of two pharmacogenomic studies, and it calls for an evaluation of how we interpret science in the face of discrepant results.