Run automated statistics and image checks on every cancer manuscript before review
Software can already spot impossible statistics, mismatched p-values and duplicated images in a paper. Journals should run these checks on every submission, as spell-check runs on every document.
Tools such as statcheck (recomputing p-values from reported test statistics), GRIM and SPRITE (checking whether means are possible given sample sizes) and image-duplication detectors identify errors and manipulation at scale. Some journals run image checks; almost none run statistical checks routinely. Integrating a pipeline at submission, with results shown to authors and reviewers, would catch a substantial fraction of errors before publication and deter fabrication.
- Preclinical results do not reproduce · Fewer than half of landmark cancer biology findings reproduce when someone else tries.
Pages like this
not linked directly; found by shared links- IdeaRandom audits of published, funded research, like tax audits
Shares Deposit the raw blots, gels and microscopy images behind every figure, Preclinical results do not reproduce.
- IdeaTime-stamped electronic lab notebooks submitted with the paper
Shares Deposit the raw blots, gels and microscopy images behind every figure, Preclinical results do not reproduce.
- IdeaJournals check that data links actually work, and flag papers whose data vanish
Shares Deposit the raw blots, gels and microscopy images behind every figure, Preclinical results do not reproduce.
- IdeaIndependent replication of the key experiment before first-in-human academic trials
Shares Paid independent statistical review for preclinical papers that inform trials, Preclinical results do not reproduce.
- CollectionPubMed & Europe PMC
Shares Automatically detect when a trial changes its outcomes after the fact, Preclinical results do not reproduce.