Conditions That Cause Sober People to Fail Field Sobriety Tests

Field sobriety tests are presented to juries as objective evidence of intoxication. They are not. The three standardized tests (Horizontal Gaze Nystagmus (HGN), Walk and Turn, and One Leg Stand) have documented accuracy limitations, known sources of error, and conditions under which their results are scientifically unreliable. Any officer who administers these tests in conditions that deviate from the validated protocol, to a subject with physical or medical characteristics outside the studied population, is generating results that cannot carry the scientific weight the prosecution will claim they carry.

Managing Partner Deandra M. Grant is a trained SFST instructor who administers and grades the SFST component of the ACS-CHAL Forensic Lawyer-Scientist examination and the DUIDLA Board Certification examination. She holds the ACS-CHAL Forensic Lawyer-Scientist designation along with a Master’s Degree in Pharmaceutical Science and a Graduate Certificate in Forensic Toxicology. Partner Douglas E. Huff holds the same ACS-CHAL designation. When we challenge SFST results, we do so from the position of people who teach these tests and know precisely what proper administration requires.

The Clues Do Not Distinguish Causes

Everything on this page follows from one mechanical fact about how these tests work.

The officer is not judging whether you seemed drunk. The officer is counting defined behaviors called clues and comparing the total against a threshold: two of eight on the walk-and-turn, two of four on the one-leg stand, four of six on the eye test. The form has a box for each clue and a place for the total. It has no box for why.

A clue counted is a clue counted. Swaying because the shoulder slopes toward a drainage ditch and swaying because of alcohol produce the same mark on the same form. The scoring system was built to detect impairment efficiently, not to sort out competing explanations, and it does exactly what it was built to do.

Which makes the useful question narrower and more answerable than “was this person drunk.” It is: what else, in this specific stop, could have produced the specific clues this officer wrote down.

What the Training Materials Themselves Say

The strongest version of this argument does not come from a defense expert. It comes from the officer’s own training manual, which names several of these conditions outright.

On the balance tests, the original research indicated that individuals over 65 years of age, and individuals with back, leg, or inner ear problems, had difficulty performing them. For the one-leg stand, the research added people who are overweight by 50 or more pounds. The materials instruct that persons wearing heels more than two inches high should be given the opportunity to remove their shoes. And both balance tests call for a reasonably dry, hard, level, non-slippery surface, with the one-leg stand materials conceding that standardizing the test for every road condition is unrealistic.

Two admissions in that same research matter as much as the conditions themselves. Fewer than 1.5% of the subjects were over 65, and the final report contained no data at all about what the subjects weighed. So on two of the factors the materials flag, the underlying research has almost nothing to say about how those groups actually perform.

That research is public. It is Development and Field Test of Psychophysical Tests for DWI Arrests, and the later field validation studies conducted in Colorado, Florida, and San Diego are the source of every accuracy figure a prosecutor will quote. How those studies were designed and what they actually measured is on A Critical History of the SFSTs.

Age

Affects: walk-and-turn and one-leg stand. The training materials name people over 65 as having difficulty with the balance tests, and fewer than 1.5% of the original research subjects were in that group. Balance genuinely declines with age through reduced vestibular function, loss of muscle strength, and changes in proprioception. An older driver who sways or reaches out an arm is doing something ordinary for their age, on a test whose accuracy figures were never really established for people like them.

Weight and Body Type

Affects: one-leg stand primarily. The original research flagged people carrying 50 or more pounds of extra weight as likely to have difficulty holding the one-leg position, and the final report recorded no data about subject weight. Body mass changes where your center of gravity sits and how much work the standing leg has to do over 30 seconds. This is not a comment about fitness. It is a statement about physics and about a research population that was never described.

Back, Leg, Knee, and Ankle Injuries

Affects: walk-and-turn and one-leg stand. Back, leg, and related problems are named in the training materials as conditions that produce difficulty. A bad knee, a healed ankle fracture, a chronic back condition, or a hip problem changes how you hold a heel-to-toe stance and how long you can stand on one foot, and none of that has anything to do with what you drank.

The officer is supposed to ask. Before administering the balance tests, an officer is trained to find out whether the person is a suitable candidate, with back, leg, and inner ear issues named specifically, and the training gives options including alternative tasks. What happens instead is that the question gets skipped, gets asked too quickly to answer, or gets asked and ignored, and the person is scored as though they never mentioned the injury. Whether it was asked, and what the answer was, is on the video.

Inner Ear and Balance Disorders

Affects: all three tests, including the eye test. The vestibular system in the inner ear is the body’s primary balance mechanism, and it is the one condition on this list that can produce clues on all three tests, because vestibular problems can generate nystagmus as well as unsteadiness. Conditions like benign paroxysmal positional vertigo, labyrinthitis, vestibular neuritis, and Meniere’s disease are not visible to an officer and are rarely volunteered during a roadside stop. Peer-reviewed research has documented nystagmus in subjects who had consumed no alcohol at all, and the full list of non-alcohol causes is on The HGN Eye Test.

A diagnosed condition is a different and stronger argument than an ordinary characteristic, because it is provable with medical records rather than inferred from footage. That work belongs on Medical Conditions Defense.

The Surface

Affects: walk-and-turn and one-leg stand. Both balance tests call for a reasonably dry, hard, level, non-slippery surface. Almost no roadside meets that description. Highway shoulders are crowned or sloped for drainage, which means the two feet in a heel-to-toe stance are not at the same height. Gravel shifts. Wet or icy pavement is not non-slippery. Grass is neither hard nor level.

The one-leg stand materials acknowledge the problem directly, and where conditions are unsuitable the original research recommended moving the test elsewhere or administering only the eye test. Officers almost never do either. The test happens where the car stopped, and the resulting clues get recorded as if the ground were flat.

Footwear

Affects: walk-and-turn and one-leg stand. The instruction is narrower than people assume, and its exact shape matters. The training materials say a person wearing heels more than two inches high should be given the opportunity to remove their shoes. That is an offer the officer is supposed to make, not a rule that shoes come off, and it is written around heel height specifically.

So two failures are possible. The officer never makes the offer to someone in heels, which is a straightforward protocol deviation. Or the person is in cowboy boots, work boots, wedges, flip-flops, or slick-soled dress shoes, which the instruction does not cover at all but which make a heel-to-toe line just as hard. What you happened to be wearing is not evidence about your blood.

Weather and Wind

Affects: walk-and-turn and one-leg stand, and the eye test indirectly. Cold stiffens muscles and degrades fine motor control. Wind pushes against a person holding a narrow stance, and a gust during a 30-second one-leg stand will move almost anybody. Rain wets the surface and pulls attention off the task. Older training materials also warned that wind and dust irritating the eyes can interfere with the eye test.

Later editions replaced that list of interfering environmental conditions with a statement that field validation studies show environmental conditions do not affect performance. The materials do not identify those studies, and no new validation studies have been conducted since the 1990s.

Lighting and Traffic

Affects: all three tests. The eye test requires the officer to actually see your eyes, which is hard in the dark, harder in oncoming headlights, and harder still with a light bar cycling behind you. Older training materials directed officers to face the subject away from rotating lights, strobe lights, and traffic passing in close proximity, precisely because those conditions interfere.

Traffic does its own work on the balance tests. Standing on one foot a few feet from vehicles at highway speed is a different task than standing on one foot in a quiet room, and the divided attention being measured is already going to not getting hit.

Fatigue, Nerves, and Time of Night

Affects: all three tests. This is the most common condition on the list and the least documented, which is exactly why it gets dismissed.

Sleep deprivation degrades coordination, slows reaction time, and impairs divided attention. Those are the same faculties the balance tests are designed to probe, which means an exhausted sober person and an impaired person are being measured on the same axis. Most DWI stops happen late at night, on people at the end of a long day, a double shift, or a long drive.

Nerves compound it. Being stopped by an armed officer, in the dark, with a camera running and your freedom plainly at stake, produces trembling, difficulty concentrating, and worse performance on unfamiliar coordinated tasks. It would be easier to stand on one foot in a courtroom than on the shoulder of a freeway at 1 a.m., and anyone who thinks about it for ten seconds knows that.

Language and Instruction Comprehension

Affects: walk-and-turn and one-leg stand. Part of your score measures how well you followed spoken instructions, which makes comprehension part of the score.

Two of the eight walk-and-turn clues are recorded during the instruction stage, before the walking starts, and one of them is starting too soon. Someone who did not fully catch a rapid set of instructions delivered once, in a second language, over road noise, can produce a failing score before taking a step. The same applies to anyone with a hearing difficulty, and to anyone who simply missed part of a fourteen-element instruction the first time through. The training materials tell the officer to give the full instructions, confirm understanding, and demonstrate. Whether that happened is on the video.

What This Does and Does Not Prove

This page would be worth less if it overstated its case, so here is the honest boundary.

None of these factors proves you were sober. They do not disprove intoxication and they do not by themselves win a case. What they do is supply an alternative explanation for specific clues, and that matters because the State’s argument runs the other way: the officer counted clues, clues indicate impairment, therefore impairment.

That chain has a weak link, and it is the middle one. Clues indicate impairment only when the test was given to a suitable candidate, in the prescribed manner, under conditions resembling the ones it was validated in. Where that is not true, the clue count is still a clue count, but it stops carrying the meaning the State attaches to it. NHTSA’s own materials say as much: change an element of the standardized administration and the validity is compromised.

So the realistic use of this page is not “I was sober because I have a bad knee.” It is “the officer scored two clues on a sloped gravel shoulder, in boots, from someone who told him about a bad knee, and those two clues are what the arrest rests on.”

Proving the Conditions Later

Almost everything on this page is provable at the time and nearly unprovable afterward. The slope of the shoulder, the shoes, the wind, the position of the light bar, the exact words of the instructions, and whether the officer asked about injuries all live in one place: the footage. It is not permanent and it is not produced automatically, and getting it preserved early is the difference between an argument and an assertion. That is covered on Video and Evidence Preservation.

The other half is your own account, written down while you still remember it. What you were wearing, when you last slept, what you told the officer about your knee, and what the ground felt like are things only you know and will not recall precisely in six months. How to do that is on What You Can Do to Help Your Own Case.

Where This Goes in a Case

These factors are not a standalone defense. They are raw material for a specific cross-examination about a specific test, and they get their force from being paired with the protocol: what the officer was required to do, what the video shows the officer did, and which recorded clues survive that comparison. How the three tests work and where each breaks is on Field Sobriety Tests.

Doing that well takes someone who knows the protocol at the level of the training manual rather than at the level of general skepticism, which is worth asking about when you decide who to hire. Choosing a DWI Lawyer covers what to look for.