London knife crime statistics: what the figures count and what they leave out

Start with the definition, not the headline

A published knife-crime total commonly refers to offences recorded by police in which a knife or sharp instrument was used or threatened. The underlying offences can include robbery, assault, sexual offences or homicide. Because it is a cross-cutting classification, it cannot be compared safely with a broad offence group unless the definitions and counting rules match.

Different datasets answer different questions

Met and Home Office police data describes offences recorded by police. NHS admissions describe injuries treated under a clinical classification. Homicide records count deaths after investigation. The Crime Survey estimates experiences among the population it samples. Agreement can strengthen an interpretation; disagreement can reveal reporting, severity or coverage differences. Adding the counts together produces no meaningful total.

How to test a claimed rise

Check whether the comparison uses the same number of months, the same London geography and the same knife-enabled definition. Compare with the same period a year earlier to reduce seasonal distortion. Read the absolute count alongside the percentage: a large percentage based on a small number can still represent only a few cases.

Why a borough ranking needs context

Raw borough totals do not account for resident population, commuters, nightlife or transport interchanges. A resident-population rate is useful but can also distort central boroughs whose daytime population is much larger. State the denominator and avoid converting an area total into a claim about one person’s chance of victimisation.