How to Use Outlier Detector

About this tool

Flag outliers with the IQR rule, z-scores or modified z-scores without deleting data.

How to use Outlier Detector

1. Upload or paste your data and pick the column to check.
2. Choose the IQR rule, z-scores or modified z-scores and a threshold.
3. Review the flagged rows and download them to check against the source.

Worked example

Glucose values between 86 and 141 plus one value of 410 give Tukey fences of 52 to 164, so the 410 is flagged as a high outlier.

Which method should I use?

The IQR rule and modified z-score are robust to the outliers themselves. Plain z-scores suit roughly normal data and larger samples.

Should I delete outliers?

Not automatically. First check for data-entry errors, then decide based on your analysis plan and report what you did.

Put this guide into practice

• Outlier Detector — Flag outliers with the IQR rule, z-scores or modified z-scores without deleting data.