What are the four requirements for using a parametric test (assume normal distribution)?

• A. Random selection of samples from population; Data are interval-level; Data are normally distributed; Variances are equal
• B. Random selection; Data are nominal; Data are non-normal; Variances are unequal
• C. Convenience sampling; Data are ordinal; Data are normally distributed; Variances equal
• D. Random selection; Data are ratio; Data are not normally distributed; Variances equal

Answer: (A)

Parametric tests rely on a normal population distribution and a few practical conditions to be valid. The four key requirements are: random sampling from the population (to support generalizability and independence of observations), data measured on an interval or ratio scale (not nominal or ordinal, so means and variances are meaningful), the population being normally distributed (we’re assuming this in the scenario), and equal variances across groups (homogeneity of variance) so the test statistic has a stable distribution across comparisons. The option that lists random selection, interval-level data, normal distribution, and equal variances aligns with these requirements. The other options violate one or more: using nominal or ordinal data, sampling not random (convenience), or explicitly not normally distributed (which contradicts the assumption).