Use the free Outlier Detector below: your data never leaves your browser — all work happens on your device, and nothing is uploaded to any server.
Most values in a list sit in a familiar range, and then one value sits far away. It might be a typing error, such as 450 instead of 45, or it might be the most interesting case in the data, such as a genuinely record-breaking sale. Either way it deserves a second look before any summary is trusted, because a single extreme can drag an average far from the typical case. The Outlier Detector flags those values with the IQR method, a rule that is simple, transparent, and widely used.
This tool is built for people checking data, not for statisticians alone. You do not need software, a formula, or an upload. Everything happens in your browser, on your own device. Your numbers never leave the page, and nothing is sent to a server. That matters when the list holds salaries, grades, or business figures.
To use it, paste a number list separated by commas, spaces, or line breaks. The sample already in the box contains a clear outlier, 45 among values near 14, so you can see the method work at once. Press Find outliers and the result shows the quartiles, the IQR, and the two fences, then lists every value outside them. The tool needs at least four numbers, because quartiles from fewer values are not meaningful.
The method needs a plain explanation, because a flagged value is only useful if you understand why it was flagged. Sort the list and split it into quarters. The first quartile, Q1, is the value a quarter of the way through. The third quartile, Q3, is three quarters of the way through. The middle half of the data sits between them, and the IQR, the interquartile range, is Q3 minus Q1: the width of that middle half. The fences sit one and a half IQRs beyond each quartile: the lower fence is Q1 minus 1.5 times the IQR, and the upper fence is Q3 plus 1.5 times the IQR. Anything outside a fence is flagged.
The factor 1.5 is a convention, not a law of nature, and the tool is honest about that. It is the standard choice in box plots and textbooks because it flags genuinely unusual values without flagging every value in a naturally spread-out list. A stricter factor would flag more values, and a looser one fewer, but changing it per dataset to get a preferred answer would be cherry-picking. The tool therefore uses the standard fence and shows its working, so you can judge the flags rather than trust them blindly.
What you do with a flagged value matters more than the flag itself. First, check the source row: is it a typing error, a unit error such as centimetres mixed with metres, or a duplicated entry? Fix real errors and run the check again. If the value is genuine, keep it and note it, because deleting genuine extremes to make an average look tidy is how honest data becomes misleading. Report summaries both with and without the extreme when the decision matters, so readers can see its effect.
The IQR method has limits worth knowing. It works best for lists that are roughly mound-shaped or at least not split into two separate groups. In a list with two clusters, such as two classes with very different scores, the middle half spans both groups and the fences may flag nothing, or flag a whole group. Small lists also give jumpy quartiles, so a flag from four or five values is a hint, not a verdict. For very large lists, a few flags are expected even in clean data, because the fence is a rule of thumb, not a proof of error.
Teachers use this tool to spot marking errors, shops use it to spot data-entry mistakes in prices, and researchers use it as a first check before any average is reported. The flag is the start of an investigation, never the end of one.
How to find outliers with the IQR method
- Paste your numbers, separated by commas, spaces, or line breaks.
- Press Find outliers.
- Read Q1, the median, Q3, the IQR, and the two fences.
- Check each flagged value in the source before changing anything.
Frequently asked questions
What is the IQR? Q3 minus Q1: the width of the middle half of the data.
Why 1.5 times the IQR? It is the standard fence used in box plots. It flags unusual values without flagging ordinary spread.
Should I delete flagged values? Only if they are errors. Genuine extremes should be kept and noted.
Is my data uploaded? No. All calculation happens in your browser.
Paste a list you already trust and run the check anyway. The value it flags is often a typo worth fixing, and when it is genuine, it is often the most interesting number in the list.