TO STUDY THE EFFECT OF MONOSODIUM GLUTAMATE ON THE URINARY BLADDER

Abstract

Objective: This study evaluates the appropriateness of statistical tests (Z-test, t-test, and Mann–Whitney U test) for small-sample analyses, addressing common misconceptions and providing evidence-based guidelines for test selection in low-data scenarios. Methods: A comparative framework was developed to assess the performance of parametric and non-parametric tests under varying sample sizes (n<10 to n=30) and distributional assumptions (normal vs. skewed). Normality was tested through Shapiro–Wilk tests, and simulated datasets with controlled variance and outliers were analyzed. Results: Z-tests produced inflated type I errors for n<30 due to reliance on known population variance, a rarely feasible assumption in small samples. T-tests maintained robustness for n≥15 with normal distributions, whereas Mann–Whitney U tests outperformed parametric alternatives for n<15 or non-normal data (skewness >2). Conclusion: Researchers should default to t-tests for small samples with approximate normality and use Mann–Whitney U tests for highly skewed data or n<15. Transparent reporting of sample sizes, normality checks, and test rationale is critical to ensure methodological validity.