What is the recommended action if data are non-Gaussian and normality cannot be established after transformation?

• A. Parametric methods cannot be used
• B. Nonparametric methods cannot be used
• C. Robust methods can be used only
• D. Bootstrap methods are required

Answer: (A)

When data remain non-Gaussian and normality cannot be established after transformation, the assumptions that underlie parametric analyses are violated. Parametric methods rely on normality (or near-normality) of the data or residuals, so you cannot trust their results here. The appropriate path is to avoid parametric approaches and use alternatives that do not require strict normal distributions. Nonparametric tests are a natural choice because they do not assume a specific distribution. Robust methods can help with outliers or deviations but do not fix the fundamental normality requirement for all parametric tests, and bootstrap methods offer a resampling approach to inference without relying on normality, though they are not strictly required in every case.