How this calculator works
The default rule places fences 1.5 IQR below Q1 and above Q3, using the selected quartile algorithm. The alternative rule places boundaries a chosen number of sample standard deviations from the mean. Values strictly beyond a boundary are flagged; equality stays inside. Standard-deviation screening is sensitive to the very extremes it aims to detect, and neither rule determines whether a value is erroneous.
Formula
IQR fences: lower = Q1 - k(Q3 - Q1), upper = Q3 + k(Q3 - Q1). SD rule: mean ± threshold × sample SD.
Worked example
For 1, 2, 3, 4, 5, 100, linear quartiles are 2.25 and 4.75, so IQR = 2.5. The default fences are -1.5 and 8.5, flagging 100.
Advertisement
Frequently asked questions
Should I delete every flagged value?
No. Check measurement, entry and sampling context first. Legitimate rare observations can be central to the analysis.
What if IQR is zero?
Both fences meet at the shared quartile, so other values can be flagged. This often happens with heavily tied data.
Why is no chart shown?
The table identifies original positions and boundaries. Use a graph alongside the results to inspect clusters, trends and skew.