The median is sensitive to outliers only when they change the middle position; it is usually much less sensitive than the mean. The mean uses every value's magnitude, while the median uses the middle position after sorting. Reading both is usually more informative than treating either number as a universal summary.
The mean uses every magnitude
For 2, 3, 3, 4 and 38, the sum is 50 and the arithmetic mean is 10. Every value contributes directly to that total, so changing 38 to 138 adds 100 to the sum and 20 to the mean. This sensitivity is not an error: it is exactly what equal sharing of the total means. It matters when the total itself has meaning, but it can make a skewed list look more typical than it is.
Is the median sensitive to outliers?
Yes, but usually far less than the mean. Sort the same five values as 2, 3, 3, 4, 38: the middle value and median are 3. Replacing 38 with 138 moves the mean from 10 to 30, while the median stays 3 because the middle position does not change. With an even count, the median is the arithmetic mean of the two central values. The median is resistant, not immune: it changes when enough observations move across the center, and it does not show how far an endpoint lies beyond that center.
Compare summaries with the sorted evidence
A gap between mean and median is a prompt to inspect the distribution, not a diagnosis by itself. Check the count, minimum, maximum, range and sorted values; confirm units and decide whether repeated observations are intentional. SetStats performs neutral descriptive arithmetic only. It does not decide which statistic is appropriate for a report, estimate sampling uncertainty, remove outliers or validate the source data. Those decisions remain part of the analysis around the calculation.