An 8-year effort to test the rigor of 50 influential cancer lab experiments found that half of the scientific claims in those studies didn’t hold up, AP reports.
- The study, an effort of the Reproducibility Project, focused on experiments from cancer biology papers—not well-established treatments. Their findings, according to Science: Out of 23 papers, results from only 5 papers could be fully reproduced.
- Overall, only 46% of 112 reported experimental effects met at least 3 of 5 criteria for replication, such as a change in the same direction—increased or decreased cancer cell growth.
- Researchers were stymied by lack of detail provided in original studies and a lack of cooperation from scientists who did the original work.
Takeaway: Funders and journals need to require authors to share their methods and materials in more detail.
Related: Is cancer biology research reproducible? After a decade of work, the answer still isn’t clear – STAT