Lectures and seminars Webinar: Reduction by design; three statistical principles for minimising animal use
The aim of these CPD webinars (Continuing professional development) is to follow the legal requirements for maintenance and demonstration of competence in laboratory animal science, and to facilitate the implementations of the 3R’s in routine animal work.
Penny S. Reynolds, Ph.D., Assistant Professor, Department of Anesthesiology, College of Medicine, University of Florida, Gainesville, FL, USA.
Penny S. Reynolds is an Assistant Professor of Anesthesiology, a wildlife biologist, a statistics expert, a 3Rs advocate, and an author. She has received the UK LASA-Animals in Science Education Trust 3Rs Prize, and the AAALAC Global 3Rs Award. She was one of the co-authors of the ARRIVE 2.0 Reporting Guidelines and Explanation and Elaboration document. Her book “A Guide to Sample Size for Animal-based Studies” is scheduled for release in November 2023.
Register here (by November 30, 16.00 CEST)
Reduction of animal numbers is one of the three pillars of the 3Rs, but – paradoxically - is often perceived as conflicting with the goal of maximizing statistical power. However, a large sample size does not guarantee adequate power, and high power does not ensure that results are informative.
Based on Ronald Fisher’s concept of statistical information, I outline three statistically-based but non-technical strategies for maximizing experimental signal and reducing noise. These include a clear statement of the research question, statistically-structured study design, and variance minimisation. These statistical process basics were developed almost a century ago, and were clearly described by Russell and Burch in 1959.
However, most researchers seem to be completely unaware of their existence, and research quality is still poor. Too-small, noisy, and biased studies are uninformative and waste animals. It is our shared responsibility to promote well-established best practice standards, and emphasise appropriate design of experiments over statistical significance.