Non-parametric tests are particularly useful in which scenarios?

• A. Large samples with normal distributions
• B. Small, homogeneous samples or non-normal distributions
• C. When population parameters are precisely known
• D. When data are measured on interval scales only

Answer: (B)

Non-parametric tests don't rely on a specific population distribution or known parameters, and they often use ranks rather than raw values. This makes them especially useful when the sample size is small or the data aren't normally distributed (for example, ordinal data or data with outliers). In such cases, they remain valid without the strong assumptions that parametric tests require. With large samples that do meet normality, parametric tests typically have more statistical power, which is why non-parametric methods are less preferred there. It's not about data being strictly interval-scale; non-parametric tests can handle ordinal data as well, so they’re not limited to interval measurements.