Including both sexes in in vivo research does not necessitate an increase in sample size: a key role for factorial analysis methods

Abstract

In recent years, there has been a strong drive to improve the inclusion of animals of both sex during in vivo research, driven by a need to improve sex representation in fundamental biology and drug development. This has resulted in inclusion mandates by funding bodies and journals, alongside numerous published manuscripts highlighting the issue and providing guidance to scientists. However, progress is slow and blockers to the routine use of both sexes remain. From a statistical and experimental design perspective, concerns include difficulty selecting and conducting an appropriate analysis and the perceived need for a higher sample size to achieve an equivalent level of statistical power. When both sexes are included, analysis errors are frequent, including inappropriate pooling or sex-disaggregation of the data. These mistakes result in a failure to properly account for the variation in the data that arises from sex, and subsequently lead to poor inference regarding the biological impact of sex. The purpose of this manuscript is to address frequently cited blockers and analysis errors, thus providing a practical guide to support scientists in the design of in vivo studies which include both sexes. Primarily, we demonstrate that there is no loss of power to detect treatment effects when splitting the sample size across sexes in most common biological scenarios, providing that the data are analysed appropriately. In the rare situations where power is lost, the benefit of understanding the role of sex outweighs the power considerations. When estimating a generalisable translatable effect, where exploring sex differences are not the primary scientific objective, we recommend splitting the sample size across male and female mice as a standard strategy. We also demonstrate an optimal analysis pipeline for analysing data gathered using both sexes which is designed to help address common analysis errors.