What scares burglars away? People, mostly

The UNC Charlotte study (Kuhns et al., 2013) put structured surveys in front of 422 incarcerated burglars and asked how they chose targets, what they checked first, and what made them quit. It was funded by the Alarm Industry Research and Educational Foundation, so read its alarm findings with that in mind. The Oregon material is journalism, not peer review: in 2017, reporter Kyle Iboshi sent questionnaires to inmates convicted of burglary and published answers from 86 of them.

Neither is a randomized trial, and self-reports from prison have an obvious limit — these are, by definition, the burglars who got caught. But the two samples were collected independently, years apart, by people with different motives, and they converge on the same core finding. That convergence is why this article exists.

Essentially every Oregon burglar said they knocked or rang the doorbell before trying a door, with a cover story ready in case anyone answered. A car in the driveway was among their strongest reasons to skip a house entirely. In the UNC Charlotte survey, the proximity of other people ranked among the top target-selection factors, ahead of any product you can buy.

The corollary is uncomfortable: the Bureau of Justice Statistics found a household member was present in roughly 28 percent of burglaries. The occupancy check fails often enough that "they'll leave if we're home" is not a plan — but the effort burglars put into that check tells you how badly they want to avoid you.

Sound inside the house ranked with the car in the driveway among the Oregon burglars' strongest avoid signals — a TV, a radio, anything suggesting the knock might be answered. Big, loud dogs sat in the same tier. Not dogs generally: big and loud, the kind that announce an intruder to the whole street and cannot be predicted, distracted, or reasoned with.

Notice what the dog is actually doing in the burglar's mental model. It is not the bite; injury barely comes up in either study. It is exposure — a dog converts a quiet entry into a loud public event and sharply raises the odds that a person shows up to see what the noise is.

Burglary works as a crime because it is anonymous. The UNC Charlotte respondents weighed escape routes and the nearness of other people — witnesses — as heavily as anything else. The Oregon burglars treated cameras with ambivalence, calling them a deterrent but also a possible sign of valuables, precisely because a camera is an uncertain witness: it might identify them later, or it might record an unrecognizable hood.

What burglars describe fearing is not being recorded in general but being seen in particular: a neighbor who gets a good look, a homeowner who speaks to them, anyone who can attach a face and a description to them while they still have to walk back down the same street they came in on.

Run the standard checklist through that lens and it stops being a list of gadgets. An alarm's threat is not the sound itself; it is that the sound summons attention — roughly 60 percent of the UNC Charlotte burglars said they would pick another target if an alarm was present. A lit yard means a neighbor might see. A yard sign claims a system, and a system claims a response. A dog is a resident that always answers the door.

That is also why the weakest deterrents are the ones that fake the proxy without the payoff — a sticker with no system behind it, a dummy camera, a light on a timer that never varies. Burglars who case houses are testing whether the signal is connected to a person. The knock on the door is exactly that test, and most houses fail it silently.

This is the finding a property-defense system is designed around. Monitoring systems — cameras, alarms, app alerts — record what happens and notify someone far away; the burglar's bet is that no person arrives in time to matter. PerimeterEye, an Automatic Property Defense System, is built to lose that bet for them: it verifies an unknown person, locks a motorized searchlight onto them and physically follows them across the property, and addresses them with live, targeted AI speech — "You, in the dark hoodie — you are not authorized to be on this property" — before escalating to a 130 dB siren and priority alerts with live video.

Every one of those steps is a confrontation signal straight from the survey research: you have been seen, individually; noise is now guaranteed; attention is on you. Because the system knows the household, family arrivals and mail carriers never trigger any of it, and you can watch, talk live, or take over at any point. The honest disclaimer belongs here too: no security system can guarantee prevention of every incident — the evidence simply says confrontation is the thing burglars leave to avoid.