Improving translational oncology with better statistics
EXCLI Journal · 2026-10-21 · 25 · 1583-1587
DOI: 10.17179/excli2026-9843 ↗Abstract
In most empirical studies, each research question is assessed using statistical hypothesis testing, with a P value below the pre-specified significance level (α) providing evidence against the null hypothesis. Whilst reporting P values is common practice, they may only indicate the presence or absence of evidence for a difference. Of the 129 preclinical anticancer studies published in 2025 that were reviewed, 95 performed null hypothesis significance testing, including analysis of variance and/or t tests. Of these, only two reported effect size measures. In particular, several aspects of statistical reporting could be improved in future research to facilitate evidence synthesis and translational oncology.
EUVIMED