Data Reproducibility
摘要
Reproducibility is an essential element of the scientific enterprise, but the capacity to reproduce findings can be taken to an absurd degree. The meanings of reproducing findings within a single laboratory and between laboratories are contrasted. There are many explanations for the failure to reproduce. Regardless of the explanation, the most effective antidote to irreproducibility is maximizing the quality of the original data. The puritanical moral outrage that is induced by the failure to reproduce may be more to do with the appearance of wastefulness to funders. Instead of the sense that reproducibility failure is to be avoided at all costs, it may prove to provide valuable opportunity to advance science by resolving the cause of the discrepancy. The hypothetical situation with all results reproducible may seem superficially desirable, but I argue that it would be more akin to a dystopia with the tendency to eschew risk-taking.