Are biologically meaningful effect sizes a factor in study design? A systematic review of translational chronic variable stress studies.

The design of in vivo studies using laboratory animals is normally guided by the 3Rs: Replacement, Reduction and Refinement. The concept of Reduction is particularly important in the context of estimating sample size; the selected sample size should allow the detection of a predetermined effect size using appropriate statistical tests, but not at the expense of using too many animals. To explore this, we conducted a systematic review of animal studies using chronic variable stress to ask whether the authors used a biologically meaningful effect size to determine the sample size. Only one article in our sample of 385 reported doing this, and most articles did not report a justification for the sample size used. Determining a biologically meaningful effect size is not always straightforward, but all appropriately powered studies based on a biologically meaningful effect size are useful, including studies where the data do not support the hypothesis. Accordingly, we believe the use of biologically meaningful effect sizes is central to decisions about study design and interpretation, and we discuss reasons and ways to promote its use.