Before using a parametric test, you need to perform assumption tests. Key assumptions include: Normality: Test using a Shapiro-Wilk or Kolmogorov-Smirnov test. Equal variances (homoscedasticity): Test using Levene's Test for Equality of Variances. Sample size: Ensure the sample size is large enough, typically n > 30. If your data fails these tests, limitations such as biased results and reduced statistical power may arise. For instance, if the variances are unequal, the reliability of a t-test decreases, leading to the use of non-parametric tests like the Mann-Whitney U test. Non-parametric alternatives: t-test: Mann-Whitney U test ANOVA: Kruskal-Wallis test Pearson's r: Spearman's rank correlation Excel's Analysis ToolPak does support some non-parametric tests (like the rank correlation), but other tests may require additional tools or manual calculations