Which approach identifies multiple outliers at both ends simultaneously?

• A. Horn's Algorithm
• B. Dixon's Range Statistic
• C. Box-plots
• D. Dot plots

Answer: (A)

Identifying multiple outliers at both ends requires a procedure that tests extremes and then updates the data as outliers are removed, so it can uncover several extremes on either side in one analysis. Horn's algorithm does exactly that: it repeatedly tests the most extreme value against a cutoff derived from the current data, removes it if it's an outlier, and continues with the remaining observations. This stepwise, dual-tail approach allows detection of multiple outliers on both ends in a single pass.

Dixon's range test is designed to flag a single outlier at one end, so it isn’t suited for identifying multiple extremes across both tails. Box-plots visually flag outliers beyond the 1.5 IQR threshold and can show multiple outliers on either end, but they’re identified via a rule rather than a formal, iterative, multi-outlier testing process. Dot plots are simply visual representations, not a formal multi-outlier detection method.