Outlier Calculator
Paste your numbers to find the outliers. Use Tukey's IQR fences (1.5 × or 3 × IQR), a z-score cutoff, or the robust modified z-score based on the median absolute deviation. You get the fences, the list of outliers, a box plot and a cleaned copy of your data.
Separate numbers with commas, spaces or new lines.
How to use the Outlier Calculator
- Enter your data, separated by commas, spaces or new lines.
- Choose a method: IQR fences are the standard choice; the modified z-score is best for small or messy data.
- Adjust the cutoff (fence multiplier or z threshold) if needed.
- Read the outliers and the fences, and check them on the box plot.
- Copy the results to get the outlier list and the cleaned data.
Outlier rules at a glance
| Method | Outlier if | Best for |
|---|---|---|
| IQR fences (Tukey) | x < Q1 − 1.5·IQR or x > Q3 + 1.5·IQR | Most data, any shape; the box-plot standard |
| Extreme IQR fences | x < Q1 − 3·IQR or x > Q3 + 3·IQR | Flagging only far-out values |
| Z-score | |x − x̄| ÷ s > 3 (or 2, 2.5) | Large, roughly normal samples |
| Modified z-score | |0.6745 (x − median) ÷ MAD| > 3.5 | Small samples and data with several outliers |
Worked example
Data: 1, 51, 52, 53, 53, 54, 55, 56, 57, 58, 59, 98 (n = 12, sorted). With the exclusive quartile method, Q1 = (52 + 53) ÷ 2 = 52.5 and Q3 = (57 + 58) ÷ 2 = 57.5, so IQR = 5.
- Lower fence = 52.5 − 1.5 × 5 = 45
- Upper fence = 57.5 + 1.5 × 5 = 65
- Outliers: 1 (below 45) and 98 (above 65). Both are also beyond the 3 × IQR fences (37.5 and 72.5), so they are extreme outliers.
The same data with the z-score rule at |z| > 3 finds no outliers at all, because the two extreme values inflate the standard deviation to about 20.9. This masking effect is why the IQR and modified z-score methods are usually preferred.
Frequently asked questions
How do you find outliers with the IQR method?
Find Q1 and Q3 and compute IQR = Q3 − Q1. The lower fence is Q1 − 1.5 × IQR and the upper fence is Q3 + 1.5 × IQR. Any value below the lower fence or above the upper fence is an outlier. Using 3 × IQR instead flags only extreme outliers.
What is the difference between mild and extreme outliers?
In Tukey's scheme, values beyond the 1.5 × IQR fences but inside the 3 × IQR fences are mild outliers, and values beyond the 3 × IQR fences are extreme outliers.
When should I use the z-score method?
The z-score rule (|z| > 3, or sometimes 2 or 2.5) suits large, roughly normal data sets. In small samples it works poorly: the largest possible |z| is (n − 1) ÷ √n, so with 10 values no point can exceed 2.85. Outliers also inflate the standard deviation, which can hide them.
What is the modified z-score?
The modified z-score, M = 0.6745 × (x − median) ÷ MAD, replaces the mean and standard deviation with the median and the median absolute deviation, which are not distorted by the outliers themselves. Iglewicz and Hoaglin recommend flagging values with |M| > 3.5.
Should I delete outliers from my data?
Not automatically. First check whether an outlier is a data-entry or measurement error, in which case correct or remove it. Genuine extreme values may be the most interesting part of the data. If you remove any, report it and consider showing results with and without them.
Why does my textbook give different fences?
Quartiles are calculated differently by different textbooks, calculators and spreadsheets, especially for small data sets. Choose the matching quartile method in the calculator (TI-84, Tukey hinges, Excel QUARTILE.INC or QUARTILE.EXC).