How Persons Arrested for DWI Are Identified — The Science Behind Standardized Field Sobriety Tests
When a police officer pulls over a driver they suspect has been drinking, the next few minutes often follow a very specific, carefully designed script. The driver walks a straight line, balances on one leg, or follows a moving object with their eyes. These roadside exercises are not arbitrary — they are the product of years of funded research. The landmark 1981 report, Development and Field Test of Psychophysical Tests for DWI Arrest, by V. Tharp, M. Burns, and H. Moskowitz of the Southern California Research Institute, laid the scientific foundation for the three-test battery that law enforcement agencies across the United States still rely on today.
The Hidden Scale of the Drunk Driving Problem
To understand why this research mattered so much, it helps to grasp just how rarely impaired drivers were actually caught before standardized testing existed. Estimates from that era placed the combined probability of a drunk driver having an accident or being arrested during any single trip at just 0.00089 — less than one in a thousand.
Part of this detection failure came down to a striking imbalance in who actually got stopped and charged. Nationally there are approximately three times as many drivers on the road with a blood alcohol concentration (BAC) between 0.10% and 0.14% as those with BACs in the 0.15% to 0.19% range. Yet among the persons arrested for DWI in that era, this ratio was effectively reversed — those with moderate BACs were only half as likely to be arrested as the more heavily intoxicated group.
Heavily impaired drivers swerved, ran red lights, or made obvious errors that caught police attention. Moderately impaired drivers were harder to spot and harder for officers to feel confident about arresting without a reliable method to assess impairment. Closing that gap was the whole point of building a standardized test battery.
From a Long List Down to Three: Building the Battery
The 1981 report was the second phase of a research effort that began in 1977. In that earlier study, Burns and Moskowitz had police officers administer over ten different physical and cognitive tests to 238 participants with varying alcohol levels in their systems. The tests ranged from finger-to-nose coordination and backward counting to letter cancellation and balance exercises.
All of the tests showed some sensitivity to alcohol, but a statistical analysis identified the three that performed best as a combined set:
- Walk-and-Turn
- One-Leg Stand
- Alcohol Gaze Nystagmus (an eye-movement test)
Together, this three-test battery correctly classified more than 83% of the 1977 participants as being above or below the 0.10% BAC threshold. The 1981 project set out to refine those tests further — standardizing exactly how they were administered and scored — and then validate the standardized versions both in a controlled laboratory and in real field conditions for the first time.
The Three Tests Explained
Walk-and-Turn
The walk-and-turn is a divided attention task, which matters because alcohol is particularly effective at disrupting the brain’s ability to handle two demands simultaneously. The suspect is asked to stand in a heel-to-toe position on a line — arms at the sides — while listening to instructions. They then take nine heel-to-toe steps forward, turn around keeping one foot on the line, and return nine steps.
Holding the heel-to-toe stance while listening to instructions dramatically improved the test’s sensitivity to alcohol. Someone who has been drinking will typically either ignore the balance requirement to listen, or lose their footing trying to do both — a difficulty sober individuals rarely encounter. The test is not appropriate for people over 65, those with back, leg, or inner-ear problems, or anyone wearing heels higher than two inches; for these groups, the nystagmus test alone is recommended.
One-Leg Stand
The suspect stands with heels together, raises one leg approximately six inches off the ground with both legs straight, and counts aloud from 1001 to 1030 to mark a 30-second window. A key finding: a person at 0.10% BAC may hold their balance for the first 20 to 25 seconds but will typically falter beyond that — which is why the full count matters. Like the walk-and-turn, this test is inappropriate for people over 65, those with leg, back, or inner-ear conditions, or anyone overweight by 50 or more pounds.
Horizontal Gaze Nystagmus
This is the eye-movement test officers perform by moving an object slowly in front of a driver’s face and watching how the eyes track it. Nystagmus refers to the involuntary jerking of the eyes — a normal event at extreme lateral gaze, but one that begins at far less extreme angles when the brain is under the influence of alcohol or certain depressants.
Understanding this test requires a brief note about vision. When tracking a moving object at comfortable angles, the eye uses foveal vision — the sharp, central zone of the visual field responsible for precise detail. As gaze moves toward the lateral extremes and foveal vision can no longer lead the tracking process, any alcohol-related disruption to the neural control of eye movement becomes visible as nystagmus. Officers are trained to watch for exactly where this breakdown begins.
The research revealed a strong and consistent correlation between a person’s BAC and the angle at which this jerking starts: the higher the BAC, the earlier the nystagmus begins. For the left eye, the correlation between BAC and angle of onset was -0.78; for the right eye, it was -0.74. At a BAC of 0.10%, nystagmus onset was found to occur at approximately 41 degrees of lateral deviation.
Officers were trained to watch for three specific signs:
- Onset of nystagmus before the eye reaches 45 degrees of lateral deviation
- Pronounced, distinct nystagmus at the maximum lateral position
- The breakdown of smooth pursuit — the inability of the eye to follow a moving stimulus without jerking
One important caveat: roughly 3% of the general population shows early-onset nystagmus with no alcohol in their system at all, due to certain medications, neurological conditions, or brain damage. Officers were trained to account for this.
What the Laboratory Testing Found
The core of the 1981 research was a controlled laboratory evaluation. Ten police officers — from the Los Angeles Police Department, the Los Angeles County Sheriff’s Department, and the California Highway Patrol — administered the standardized battery to 297 volunteers with BACs ranging from 0% (placebo) to 0.18%. The group included 202 men and 95 women; 145 returned for a second session to allow reliability testing.
Using the standardized battery, officers correctly classified 81% of participants with respect to whether their BAC was above or below the 0.10% legal threshold. Their BAC estimates also proved impressive: on average, they differed from the true measured figure by just 0.03%.
| Metric | Result |
|---|---|
| Volunteers tested | 297 |
| BAC range tested | 0% to 0.18% |
| Police officers administering tests | 10 |
| Correct classification rate (above/below 0.10%) | 81% |
| Average BAC estimate error | ±0.03% |
| Interrater and test-retest reliability range | 0.60 – 0.80 |
| HGN correlation with BAC (left eye) | -0.78 |
| HGN correlation with BAC (right eye) | -0.74 |
| BAC at which nystagmus onset occurs (~41°) | 0.10% |
Taking It to the Streets: The Field Evaluation
Laboratory results are one thing — real traffic stops are another. To assess whether trained officers could apply the battery effectively during actual police work, a field phase was conducted at multiple agencies over a three-month period. Officers filled out 3,128 data forms, each representing a stopped driver.
The field component was designed as a before-and-after comparison. Officers first worked as they normally would, with no special training. Then they received instruction in the standardized battery — its administration and scoring procedures — before returning to patrol.
After training, officers’ arrest rates increased, suggesting they were more willing and able to act on moderate impairment rather than only pursuing obvious cases. They also became more accurate at estimating the BAC levels of stopped drivers.
One of the more revealing findings came from anonymous breath testing of drivers who had been stopped but released without arrest. The data showed that many drinking drivers were stopped but never given a sobriety test — and were sent back onto the road. This pointed to a gap not in the tests themselves, but in officer motivation and willingness to apply them consistently.
Who Should — and Should Not — Take These Tests
The researchers carefully documented that the physical tests have real limitations with certain populations. Expecting everyone to perform walk-and-turn and one-leg stand at the same baseline level is unrealistic, and the report was explicit about this.
| Group | Walk-and-Turn | One-Leg Stand | Gaze Nystagmus |
|---|---|---|---|
| People over age 65 | Not recommended | Not recommended | Appropriate |
| Back, leg, or inner-ear conditions | Not recommended | Not recommended | Appropriate |
| Overweight by 50+ pounds | Caution advised | Not recommended | Appropriate |
| Wearing heels over 2 inches | Remove shoes first | Remove shoes first | Appropriate |
| Hard contact lens wearers | Not affected | Not affected | May interfere with eye movement |
Surface conditions matter too: both the walk-and-turn and one-leg stand require a hard, dry, level, non-slippery surface. When roadside conditions do not allow for this, the researchers recommended relocating the test or relying solely on the gaze nystagmus examination.
How This Research Changed Law Enforcement
Before this work, sobriety testing in the United States was largely informal and inconsistent. Different officers in different jurisdictions used different methods, making it difficult for courts to weigh the evidence. The research gave both law enforcement and the legal system something more solid: a battery of tests with documented validity and reliability, administered the same way every time.
The three tests became the foundation of the Standardized Field Sobriety Test (SFST) program later endorsed by the International Association of Chiefs of Police and rolled out nationwide. In 1986, the IACP’s Advisory Committee on Highway Safety formally recommended that law enforcement agencies adopt the SFST program built on this research foundation.
The shift toward documentation and standardization extended into broader police procedure as well. A question commonly raised in law enforcement training is: which of the following is a way police have changed their interrogation techniques to lower the risk of false memory syndrome? Police have increasingly adopted structured, non-leading interview formats and mandatory recording of interactions — the same underlying philosophy of consistency and verifiability that gives the standardized field sobriety battery its legal credibility. When every step is documented, both the officer’s conduct and the suspect’s performance can be independently reviewed.
The officer background data from the laboratory phase adds useful context: the ten test officers had between 1 and 19 years of experience and had encountered anywhere from 5 to 10,000 DWI suspects over their careers. Standardized training brought them all to a consistent performance level regardless of background — which was precisely the point.
A question that often appears in driver safety courses is: if you are stopped by traffic police and you have a bac of .82, which of the following is true? At that level — more than eight times the standard legal limit — an officer has immediate and overwhelming grounds for arrest, and the situation raises serious medical concerns. The SFST battery was calibrated around the 0.10% legal threshold, but its real value lies in detecting the moderate drinker at 0.10% to 0.14% who might otherwise appear functional enough to escape notice — precisely the drivers who were being missed before this research changed enforcement practice.
Conclusion
The next time you see a driver asked to follow a moving object with their eyes or walk a straight line at the roadside, you are watching nearly a decade of scientific work in action. The 1981 NHTSA-funded research didn’t just produce a list of physical exercises — it built a rigorous, standardized system designed to close the gap between how many impaired drivers were on the road and how few were actually being identified.
An 81% correct classification rate, BAC estimates accurate to within 0.03%, and field data showing measurable improvements after training: these numbers convinced a generation of law enforcement agencies to change how they approached drunk driving enforcement. The detection gap was always the real problem — and this research was a serious, evidence-based attempt to solve it with tools that could actually work at a roadside stop.