Nonparametric methods are recommended for what sample size, and which fractiles define the lower and upper reference limits?

• A. n > 120; The 2.5th and 97.5th fractiles serve as the lower and upper reference limits
• B. n > 50; The 5th and 95th fractiles serve as the limits
• C. n < 40; The 1st and 99th fractiles serve as the limits
• D. n > 120; The 5th and 95th fractiles serve as the limits

Answer: (A)

Nonparametric reference interval estimation uses the data order rather than assuming any distribution shape, so you rely on where the tails fall in the sorted sample. To define a standard central 95% reference interval with good precision, you need a sufficient number of samples, and this is typically achieved with more than 120 individuals. Using the 2.5th and 97.5th fractiles as the lower and upper reference limits captures the central 95% of values without assuming normality. For n = 120, those bounds correspond roughly to the 3rd and 118th values in the ordered data, illustrating how the tails define the interval. The other options either target different tails (leading to nonstandard intervals) or imply a smaller sample size where nonparametric estimates would be less reliable. Hence, the recommended approach is n > 120 with the 2.5th and 97.5th fractiles.