Research

The chases nobody counted

When federal researchers asked 737 American police agencies for their pursuit records in 1994, more than a third could not supply them — because they had never collected any.

The survey that mostly failed

Researchers at the University of South Carolina, funded by the National Institute of Justice, contacted 737 police agencies and asked for their pursuit and use-of-force data.

The report does not soften it. It calls the 38 percent figure discouraging and alarming, reflecting the dismal state of record keeping around pursuit driving — and notes, with some relief, that of the agencies which did hold the data, only four percent refused to hand it over.

The problem was never secrecy. It was that more than a third of American police agencies were conducting high-speed pursuits and writing nothing down.

And almost nobody was required to

Only 11 percent of responding agencies reported that their pursuit data collection was mandated by the state. Everyone else who kept records did so voluntarily.

The detail underneath that is the one worth reading twice. Agencies in ten states said they operated under a state requirement — but in only three of them did the responding agencies agree with each other.

In Minnesota, eight departments described a state-mandated pursuit reporting programme. Three other Minnesota departments were apparently unaware such a requirement existed. In six further states, only a single agency reported being under a mandate at all.

Against that, the policy findings look almost healthy: 91 percent of responding agencies had a written pursuit policy, and 72 percent had theirs reviewed by a legal authority before adoption. Agencies were taking advice on what the rules should say. They were not recording what happened under them.

What the few real datasets showed

In place of national figures, the report assembles the handful of departments that had studied themselves — and the numbers are worth setting side by side, with the caution the authors attach.

Somewhere between 29 and 38 percent of pursuits ended in a collision. Not a near miss — a crash. That is the base rate the 1990s were working from, and it is startlingly high.

The authors are careful about why these cannot simply be averaged. The CHP data is mostly freeway. Metro-Dade is a single large county agency. Baltimore County is another. The Minnesota and Illinois figures they also cite come from a mix of urban, rural and suburban departments of every size. Different reporting procedures, different definitions — some agencies record voluntary termination as an outcome category and others do not.

What strikes the authors is not the variation but the opposite: that despite all of that, the rates of accidents, injuries and deaths come out remarkably similar across time and location.

The Metro-Dade breakdown

One dataset is detailed enough to show what pursuits were actually for, and it complicates every argument about them.

More than half of all pursuits began with a traffic offence. Two percent began with the reckless or impaired driving that the public most readily accepts as a reason to chase somebody.

And then the line that both sides of this argument use. Of 952 pursuits, 341 produced an arrest for something that had nothing to do with why the chase started — mostly drugs and weapons. Read one way, that is a third of pursuits taking armed or drug-carrying offenders off the street who would otherwise have driven away. Read the other, it means the officer did not know what they were chasing when they decided to chase it.

Only 4 percent were voluntarily broken off.

Training does not always do what you expect

The full report contains a finding the published summary flattens, and it matters.

Recruits at four locations were surveyed before training and again afterwards. At three of them, training reduced willingness to pursue for minor offences — sharply. Asked about a traffic violation in low-risk conditions, St. Petersburg recruits fell from 58 percent to 24, and Miami-Dade from 71 to 46.

At the fourth, it went the other way.

South Carolina Academy recruits came out of training more willing to pursue than they went in — on traffic violations, on property felonies, on violent felonies. After the course, 96 percent of them would chase a property felon in low-risk conditions, at least 29 points above any other group.

So the honest statement is not that training makes officers more cautious. It is that training moves recruits’ attitudes substantially, in whichever direction the academy points them. Three of these four academies were teaching restraint. One was teaching something else, and it worked just as well.

That is a stronger argument for training standards than the simple version, and a considerably less comfortable one.

Why the gap mattered

Put the pieces together and the shape of the problem in 1995 is clear. Nine in ten agencies had a pursuit policy. Seven in ten had it lawyer-checked. And nearly four in ten kept no record of what happened when officers acted under it.

Departments could not audit their own decisions because there was nothing to audit. Researchers could not compare agencies because the definitions differed. Nobody could say whether a policy change had worked, except where — as in Omaha — somebody happened to be counting.

It took until the 2013 federal survey for a national count to exist at all. Eighteen years after this report said the record keeping was alarming.