N Z Med J. 2026 Mar 13;139(1631):110-115. doi: 10.26635/6965.7313.
ABSTRACT
Achieving equity in health research requires sub-groups to have meaningful, if not equal, explanatory power, ideally through similiar sample sizes. Obtaining equal sample size, though, is often not possible. Small sub-group sizes increase the risk of false conclusions being drawn, which may reinforce inequities if results are misinterpreted (e.g., saying there is a difference between study arms when there is not and, conversely, saying there is no difference when there is). Here we provide examples of common pitfalls and potential considerations to guide researchers, reviewers and editors when analysing and interpreting sub-group data. We propose that researchers focus on presenting effect sizes and confidence intervals rather than statistical significance.
PMID:41818764 | DOI:10.26635/6965.7313
