How Cops Learn to Catch Drunks

How Cops Learn to Catch Drunks

By Jeff Winkler 02/04/13

Say the alphabet backwards. Hop on one foot. How do officers assess whether a driver is over the limit? The Fix visits the Austin Police Training Academy to find out.

Austin police cadets scientifically boozing their
Sobriety test photo via Shutterstock.
This photo courtesy the author.

Witnessing a roadside sobriety test can be a sad—or sadly hilarious—affair. But watching about 60 simultaneous sobriety tests administered by Austin’s soon-to-be-finest in the early evening? Mesmerizing. Particularly when the officers’ tiny flashlights oscillate and bounce in the air, like an army of wizards conjuring an ominous spell, or club ravers dancing in slo-mo.

The cadets of the Austin Police Training Academy aren’t quite that impressive or dangerous. They are simply diligent students finishing up their week-long DWI training session with a “live-fire” scenario: administering Standardized Field Sobriety Tests (SFST) to yellow-vested volunteers operating under varying levels of sobriety—and drunkenness.

“It’s their first time doing it with somebody who’s actually been drinking,” Austin Police Academy Coordinator Sgt. Zac Pruett tells me. “They’ve been practicing on each other sober all week, so now they’re actually seeing the booze.”

In the interest of good journalism, I volunteer to go over the legal limit. During my admirably performed sobriety tests, I am just barely over .08 BAC.

But unlike your average college weekend, the police academy officials don't just hand out shot glasses and demand that volunteers start playing Quarters. In fact, the entire process is very scientific—scales, beakers and plenty of official forms included. Volunteers are weighed, asked if they wish to be over or under the legal limit, and then given Solo party cups marked with specific numbers, to indicate how much hard alcohol each participant will receive during the hour-and-a-half drinking regimen.

And it is a regimen: no eating after 2 pm (class started at 6 pm); “pre-gaming” is forbidden; blow into this breathalyzer before picking up your next drink; hand your cup to the officer, who will carefully measure and pour each shot; go back to your seat; and please have a chaperone escort you wherever you go—even the bathroom.

While the initial atmosphere inside the police academy gymnasium seems about as much fun as an SAT test, the future suspects all make the best of it. Periodically, the instructors call upon the volunteers to have their Solo cups filled with liquor of their choosing. There is decent whiskey, vodka, gin—you name it. Purely in the interests of good journalism, I volunteer to go over the legal limit.

Many of the volunteers have obviously been here before. Some are friends of officers, some are civilian employees of local law enforcement agencies and some consider this a weird weekend adventure. Those not foolish enough to bring fine-print reading material keep themselves occupied with a loud and colorful board game. Even at the smokers’ spot just outside, things are getting festive. One volunteer confidently recites the alphabet backwards, an apocryphal roadside sobriety test. Twice.  

Shortly afterwards, all the drinkers are summoned, given one more breath-test and then escorted outside. There, we fan out on the parking lot moments before a herd of cadets comes forward, politely asking us if we mind being temporarily detained.

For those who are blissfully unfamiliar with the process, the SFST, devised by the National Highway and Transportation Safety Authority (NHTSA), is comprised of three basic tests. There’s the Horizontal Gaze Nystagmu (where cops use their wands to detect involuntary eye-twitching, which increases with intoxication), the “Walk and Turn” and the “One-Leg Stand.”

Easy enough, right? Nevertheless, the NHTSA suggests 24 hours of training to be certified. And, in keeping with the state’s unofficial motto (“Everything’s Bigger in Texas”), the Austin Police Department requires twice as much training, both for roadside sobriety tests and the Austin Police Academy as a whole.

“Yeah, we tend to get more than enough [training],” laughs Cadet Arika Austin, a few days after the live-fire night. Although Cadet Austin was previously a cop in Illinois—where, during academy, she did her sobriety homework by practicing on friends and family—she had to take part in the academy like every other rookie. And it wasn’t necessarily easy.

“The funny thing is, even still, I did have some nervousness because it had been a while,” she said. “So I can only imagine what it would have been like for someone for whom this was brand new. Because it is a lot of pressure as far as making sure we know the signs and also knowing the indicators [of inebriation], and making sure that we apply it right.”

Since the “suspects” are merely volunteers—intoxicated or no—and because academy instructors are prowling around watching for mistakes, it is the police cadets who are noticeably on edge. For at least one volunteer, this table turn of the normal, late-night sobriety interrogation is immensely satisfying.

“Some of the main things you’ll see is [the cadets] learn to interact with people, which is pretty tough in and of itself,” says Sgt. Pruett. “You have to get over that nervousness and just be able talk to a perfect stranger about whatever. It’s going to mess them up because they’ve been practicing their speech ... for some time, but now they’re nervous so they mess it up.”

And the cadets do mess up. While performing the HGN test, one trainee is gently chided for moving his wand so high in the air that the suspect’s eyes roll back into his head. The same cadet is corrected shortly thereafter when, during the Walk and Turn, he follows the suspect’s footsteps with his flashlight rather than pointing the beam a few feet down the line, as per protocol.

Apart from the instructors looking over cadets’ shoulders, there are other reasons for nervousness. For one, says Sgt. Pruett, cadets have two chances to pass the live-scenario test. And failing it means failing out of the academy entirely.

The SFST themselves also are somewhat arbitrary—although cops nationwide strongly deny that. The SFST are, after all, the No. 1 way police successfully get dangerous drunks off the road. “The basis for all this is ... science, and there’s studies out there that are done on percentages,” says Sgt. Pruett. “So [the cadets] go over the scientific cases.”

Discuss DWIs with police officers long enough, and inevitably one “percentage” will come up: the supposed 91% accuracy of the SFST themselves. That figure comes primarily from an NHTSA test that is more than 30 years old.

“You don’t want to falsely arrest someone," says Cadet Arika Austin. "But we have a job to protect the public, and [to ensure] that people who are impaired don’t continue driving.”

In the original 1975 study, about 200 unemployed men were used as guinea pigs. The men were scientifically boozed (to an antiquated .10 BAC) in front of 10 police officers, who tested six different sobriety tests on them. The top-performing tests, i.e. the ones that seemed most indicative of inebriation, were the ones we have today. At the time, scientists claimed the tests had an 81% chance of correctly identifying intoxication.

With the publication of a seminal report in 1981, the three tests became the benchmark. As noted in an entertaining 2005 Washington Post story, police officers nationwide now had a scientifically sound sobriety test. No more would they have to resort to past tricks of the trade: “Some threw coins on the ground and ordered that only nickels or quarters be picked up. They would have a driver lean back and touch one finger to his nose. Count backward from 100 by threes. Trace a paper maze. Rapidly tap a telegraph key. [Or give] tongue twisters such as ‘Methodist, Episcopal, sophisticated statistics.’”

It wasn’t until states were forced to lower the DWI threshold to .08 BAC, in 1998, that the study was addressed again. Perhaps unsurprisingly, the chances of guessing “intoxication” increased to 91% where .08 BAC was the dividing line, and 94% at .04 BAC. (The NHTSA attributes this to officers’ increased “operational experience” with the field tests, rather than a lower BAC threshold.)

As scientifically appealing as the test was, it still had its skeptics. Most famous is a 1994 study during which 21 people were videotaped taking six field sobriety tests. Fourteen officers were then shown the tape and instructed to find-the-drunkard. Collectively, the cops said 46% of the subjects were under the influence. They were off by 46%. Every subject was stone-cold sober.

More recently, a 2005 study found that when a suspect isn’t, you know, obviously drunk, the probability of accurately guessing whether he or she is above or below .08 BAC using the SFST range from 30% to 60%. Even for the twitching eyes of HGN, said to be the most accurate test because it measures a person’s unconscious physiological reaction, there are “literally hundreds of known causes of nystagmus, most of which have nothing to do with intoxication.”

None of the above is to suggest that police are wrong in using the SFST as part of their toolkit—only that it must be hell on the cadets. Imagine being graded on a Rorschach test. “It’s almost an art as much as a science,” says Cadet Austin—albeit art with much higher stakes.

“It can go either way, and you don’t want to falsely arrest someone,” Austin adds, noting that she still finds borderline suspects to be particularly tricky to evaluate. “But also, we have a job to protect the public, and [to ensure] that people who are impaired don’t continue driving if they have been drinking.”

After the academy class spends about 45 minutes out in the parking lot, putting all their knowledge and nerves to the test, time is up. Pencils and flashlights down!

Ten minutes later, the cadets convene in a classroom for the big reveal. Who had been drunk and who had been sober?

Sitting attentively, the class watches as each volunteer is brought in. Sgt. Pruett asks the appropriate cadets if they would have arrested the individual for a DWI. Each time, there is at least some disagreement or hesitation among the cadets.

When Sgt. Pruett details the volunteer’s BAC during the various incremental periods (including just before being presented to the class), the officer-in-training who had done the testing either drops his or her head in failure, makes that embarrassed “oh no” face—or smiles slightly with relief. The students who had tested me hem and haw when asked if they would have made an arrest. (As it turned out, I was just slightly over the legal limit, at .083 BAC.)

“I think some people took it kind of hard,” chuckles Cadet Austin, after the fact. “You know, nobody wants to make a mistake. But I think we look at it for the most part as a learning experience.”

Jeff Winkler has written for VICE magazine, The Awl and The New Republic, among other publications, and can be reached by email at He also wrote about evangelical rehabs, "No Refusal" DWI nights, and a real-life breathalyzer tech for The Fix.