Crime taxonomy · Data selection

Outlier reference point

Measuring change from an abnormal year, like the 2020 lockdowns.

What it looks like

Growth or decline is computed against a period that was anything but normal: a pandemic trough, a one-off boom, a record year, or an index set to 100 at an extreme point.

Example

OUTLIER BASELINE

FIXED

Why it is bad

• Recovering from a crash looks like record growth if the crash is the starting point.
• A slight dip from a record year can be sold as a collapse.
• Percent-change headlines rarely show the baseline year, so readers cannot tell it was unusual.

How to fix it

1. Compare against a normal period (e.g. 2019 for pandemic-era data) or a multi-year average.
2. Show the full series so the anomaly is visible, and label it.
3. State both numbers: change since the anomaly and change since the last normal year.