There is a chart that shows up in nearly every American city's public-safety dashboard, and it looks reassuring at first glance. A line, or a bar, or a shaded band, tracks the average minutes and seconds between a 911 call and the arrival of the first uniformed officer. The line trends downward. The department, the mayor's office, and often the local paper describe this as an improvement. It is not always clear what it is an improvement in.

Response time was adopted as a headline metric in the 1970s for a reasonable reason: departments needed something to measure, and time is easy to measure. The trouble is that the underlying research, once you read it, has almost nothing to say about what happens after the officer arrives. Studies from the Kansas City experiment onward found that shaving minutes off patrol response had a small effect on arrests for in-progress crimes and no measurable effect on the far larger category of calls that are already over by the time anyone dials.

What a city actually wants to know is different, and harder. Did the call get resolved? Did the person in crisis end up in a hospital bed, a jail cell, or their own kitchen? Was force used, and if so, was it proportionate? Did the neighborhood report the next call, or did it stop calling? None of those questions have a stopwatch answer. All of them are more predictive of long-run safety than the number on the dashboard.

A handful of departments have started publishing what they call outcome dashboards alongside the traditional response-time chart. Longmont, Colorado, publishes a monthly breakdown of mental-health calls by disposition. Camden County, New Jersey, tracks use-of-force incidents per thousand contacts rather than per officer. These are modest tools. But they represent the beginning of an honest answer to the question a stopwatch cannot ask: was anyone made safer.

The political economy of the response-time metric is what keeps it alive. It is legible to voters. It is friendly to overtime budgets. It rewards the parts of the department most likely to speak to the press. Replacing it will require the same coalition that entrenched it — mayors, chiefs, budget offices, unions — to agree that a better measurement is worth the discomfort of a worse-looking chart in the first year of the switch.

None of this is an argument for slower response. It is an argument for a public conversation about what a police department is for, and whether the number we have been staring at for fifty years is the number that answers the question. A four-minute arrival at the wrong call is not safety. It is a stopwatch that measures the wrong thing, faster.