What the Deterrence Research Actually Found

Four findings, each with its caveat attached, because the caveats are the lesson.

Alarms. Interviews with 422 incarcerated burglars, led by Joseph Kuhns at UNC Charlotte, found 83% would check whether an alarm was present, and of those who found one about half would abandon the attempt with a further third sometimes retreating. A separate study by Seungmug Lee at Rutgers, of five years of Newark police data, found burglaries declined as registered alarms increased.

Both were funded by the alarm industry's own research foundation. They are real university work, not marketing — and they are not independent corroboration of each other, and no comparably rigorous independently funded study was located. Naming the funder is not an accusation; it is how a reader weighs evidence.

Cameras in public areas. The systematic reviews by Brandon Welsh and David Farrington identified 94 evaluations with 44 rigorous enough to include and found a 51% crime decrease in parking lots — with statistically non-significant effects in city centers, public housing and public transport.

The effect concentrates where the space is bounded and single-purpose and the crime is against stationary property. The parking-lot number is the one quoted in camera marketing, and generalizing it to a residential street is exactly the error the review structure exists to prevent.

Doorbell cameras. No systematic review comparable to the public CCTV work exists.

A manufacturer's claimed 55% reduction in a Los Angeles pilot has not been independently verified, the police department's own analysis found a smaller effect, and area burglary later rose. An independent look in one Utah city found homes without the devices no worse off.

On whether footage improves clearance rates, no rigorous research was found either way. Buy one for the package theft and the record, not for a deterrence figure nobody has demonstrated.

Lighting, which is the interesting one. Welsh and Farrington's systematic review found a significant crime reduction, then tested why by comparing daytime and nighttime effects.

If lighting worked by letting people see threats, night crime should have fallen more. It did not — crime fell in daylight too, pointing away from visibility and toward place investment and the informal social control that follows from a street being cared for and used. A well-maintained, occupied street appears to be doing the work, not the lumens.

And the standard objection, answered with numbers: a review of 102 evaluations by Rob Guerette and Kate Bowers found displacement in about 26% of examined instances and diffusion of benefits — nearby unprotected places getting safer too — in about 27%. Hardening your home is at least as likely to help the house next door as to hurt it.