Statistics and animals in biomedical research

Abstract

A UK government-sponsored review1 has found that out of 271 randomly chosen papers reporting experiments on laboratory animals only 59% stated the objective of the study and the number of animals used. Most did not use randomisation or blinding, and only 70% of the publications that used statistical methods gave their methods and a measure of error. Michael Festing, one of the authors of the review, wants better statistics in animal experiments. © 2010 The Royal Statistical Society.