以回应错误作为诚信的衡量标准

At Science, we typically retract two to five papers a year. Although the number is relatively small, most retractions are discussed more widely. Unfortunately, the defensiveness of researchers, institutions, and journals is widely discussed, too. code. The results—highlighted in a Science Editor’s blog to kick off 2026—are encouraging. And yet, there’s more work to do. Science will take two new steps in 2026 related to improving data sharing, as outlined in the blog. Like many publishers, we’re also thinking about balancing open- ness of data with safety, where needed. A paper we published in October gave our editors an opportunity to work with authors to de- velop an innovative framework to manage an information hazard. …