Difficulties and Practical Issues in Statistical Analysis in Research with Small Sample Size: A Review

This research stems from the high use of small samples in various disciplines, which potentially reduces the validity of statistical inference. This study aims to identify the main challenges and methodological strategies to improve the quality of analysis under these conditions. The method used is a PRISMA-based systematic literature review of 39 articles (2020–2025) from Scopus.

The results show that small samples trigger several major issues, such as low statistical power, estimation instability, assumption violations, increased type I and II errors, as well as overfitting and low generalization. Various approaches such as bootstrap, Bayesian methods, penalized regression, and robust techniques have proven to help reduce bias, although they do not completely overcome data limitations. Implicatively, this research emphasizes the importance of transparency, design planning, and the appropriate selection of analysis methods to maintain the credibility of research results, especially in the context of evidence-based policy making.