Which statement best describes Dixon's Range Statistic compared with Horn's Algorithm in outlier detection?

• A. Dixon's Range Statistic focuses on the single most extreme value and tends to retain data (conservative).
• B. Dixon's Range Statistic identifies multiple outliers at both ends and tends to eliminate data (liberal).
• C. Horn's Algorithm focuses on the single most extreme value.
• D. Horn's Algorithm ignores outliers.

Answer: (B)

Outlier detection methods differ in whether they target more than one extreme value or focus on a single point. Dixon's Range Statistic looks at the extremes from both ends of the dataset and uses the gaps relative to the overall range to decide if those extreme values are outliers. Because it evaluates ends on both sides and can flag multiple values as outliers in small samples, it tends to be more liberal in eliminating data.

Horn's Algorithm, on the other hand, is geared toward identifying only the single most extreme value as an outlier and is more conservative about discarding data, often preserving the rest of the dataset.

So the description that Dixon's Range Statistic identifies multiple outliers at both ends and tends to eliminate data aligns with its more liberal, multi-end approach, while Horn's focuses on a single extreme and is more conservative.