Outliers need a closer look — Data, 11–13 years
An outlier is a value far from the rest. It may be a mistake, a rare event, or a real clue, so good data work investigates it instead of deleting it automatically.
What an outlier is
An outlier is a data value that sits unusually far from the main group. In a list of test scores, 18, 19, 20, 21 and 95 would make 95 stand out. Standing apart does not prove the value is wrong.
Why investigate outliers?
One unusual value can change a mean, a range or the story a chart seems to tell. Checking it can reveal a typing mistake, a broken measuring device, or something genuinely unusual that deserves attention.
Checking one unusual result
A sensor records 10, 11, 10, 12 and 110 degrees. The 110 is an outlier, so first check the original reading and the sensor. If it was meant to be 11, correct the record; if it was real, keep it and explain why it is unusual.
Do not delete it automatically
People often remove an outlier because it makes the average look untidy. That impulse is understandable: a clean group is easier to summarise. But deleting a real value can hide the event we most need to understand, while keeping a measurement error can mislead us.
Outliers in real life
Hospitals, factories and websites watch for unusual readings. A sudden high temperature in a machine may signal danger, while an unusually long delivery time may reveal a problem. The unusual value is a clue, not a conclusion.
Keep exploring
Other languages
Loading MyLeoNes™…