Field Sobriety Tests in Texas: What the Science Actually Says

The Validation Research: What It Actually Showed

The three standardized field sobriety tests were developed and validated through a series of research studies commissioned by the National Highway Traffic Safety Administration (NHTSA) and conducted by the Southern California Research Institute beginning in 1977. The studies produced accuracy estimates for each test when administered under the specific conditions of the research protocol.

Those conditions included a dry, level, well-lit testing surface; a sober, trained administrator; and test subjects who were not experiencing fatigue, illness, injury, or extreme emotional distress. The accuracy figures most frequently cited (77% for the Walk and Turn, 83% for the One Leg Stand, and 88% for the HGN) are flawed to begin with and specific to those controlled conditions. They are not accuracy figures for roadside administration at 2 a.m. on an uneven highway shoulder in headlight glare.

The defense implication is direct: when the conditions under which a test is administered materially deviate from the conditions under which it was validated, the validation research no longer supports the accuracy claim. Officers are trained to administer these tests in a standardized way precisely because any deviation compromises the reliability of the result. When that standardized administration did not happen, the officer’s opinion of impairment is not supported by the science the prosecution is relying on.

The Three Tests: What Each Requires and Where Each Fails

Horizontal Gaze Nystagmus (HGN)

The HGN test asks the subject to track a stimulus (the officer’s pen, finger, or a light) horizontally with their eyes while the officer observes for nystagmus, the involuntary jerking of the eye. The officer looks for six clues across both eyes: lack of smooth pursuit in each eye, distinct and sustained nystagmus at maximum deviation in each eye, and onset of nystagmus prior to 45 degrees in each eye. Four or more clues is the threshold for a positive result.

What the test requires of the officer. The stimulus must be held 12–15 inches from the subject’s face and moved at a specific pace which is approximately two seconds across the full range of motion for the smooth pursuit passes, and four seconds held at maximum deviation for the nystagmus-at-maximum-deviation evaluation. The officer must be positioned correctly relative to the subject, must account for ambient light conditions, and must be able to clearly observe eye movement. Officers who move the stimulus too quickly, too slowly, or at an angle, or who do not maintain the correct distance, compromise the validity of the clue observations.

Nystagmus has many causes other than alcohol. Nystagmus can be produced by dozens of conditions unrelated to alcohol consumption — from prescription medications to inner ear disorders to the flashing emergency lights present at every traffic stop. An officer who identifies nystagmus without ruling out these alternative causes is offering an opinion that exceeds what the test supports. We cover the full list in the dedicated section below.

The HGN is not admissible to prove a specific BAC. Texas courts have held that HGN evidence is admissible to support an officer’s opinion of impairment but not to establish a specific blood or breath alcohol concentration. Prosecutors who imply otherwise are misrepresenting what the test shows.

Walk and Turn (WAT)

The Walk and Turn requires the subject to stand heel-to-toe in a defined starting position while receiving instructions, then walk nine heel-to-toe steps along a real or imaginary line, turn in a specific manner, and walk nine steps back. The officer watches for eight clues: cannot keep balance during instructions, starts before instructions are finished, stops while walking, misses heel-to-toe contact, steps off the line, uses arms for balance, turns improperly, or takes the wrong number of steps. Two or more clues is the threshold for a positive result.

Surface and environmental conditions. The Walk and Turn was validated on a dry, level, hard surface. Gravel, grass, asphalt with a crown or slope, wet pavement, or any surface that is not level introduces an independent variable that affects balance performance in ways unrelated to alcohol. An officer who administers the Walk and Turn on a highway shoulder with a drainage slope and then reports that the subject ‘stepped off the line’ is reporting a result that the validation research does not support.

The instruction phase is itself a scored element. An officer who does not give the complete standardized instructions, or who gives them in an abbreviated or non-standard way, affects the subject’s ability to perform the test correctly. The instructions include a specific demonstration that the officer is required to perform. Officers who skip the demonstration or modify the instructions are not administering a standardized test.

Medical and physical conditions. The NHTSA research excluded subjects over 65 years old, subjects with inner ear disorders or back problems, and subjects who were more than 50 pounds overweight from the validation studies for the WAT and One Leg Stand. This means the accuracy figures for those tests were not established for these populations. When an officer administers the WAT to a subject who falls into one of these categories, the accuracy claim does not apply.

One Leg Stand (OLS)

The One Leg Stand requires the subject to stand with one foot raised approximately six inches off the ground, with toes pointed out, and count out loud by thousands until told to stop. The officer times the test for 30 seconds and watches for four clues: swaying, using arms for balance, hopping, and putting the foot down. Two or more clues is the threshold.

Fatigue and the 30-second duration. A 30-second one-leg stand is physically demanding for many people regardless of sobriety. Research on postural sway has consistently demonstrated that the difficulty of standing on one leg increases significantly after the first 10 to 15 seconds as the stabilizing muscles fatigue. An officer who scores swaying or arm use in the final seconds of a 30-second stand without accounting for the role of muscle fatigue is misattributing a normal physiological response to alcohol impairment.

Footwear. Subjects wearing heels, boots, sandals, or other footwear that does not provide a stable base are at a significant mechanical disadvantage on the One Leg Stand independent of any alcohol effect. The NHTSA protocol permits subjects to remove footwear before performing the test, but officers are not always consistent in offering this option.

The Negative Scoring System

A fundamental feature of the standardized field sobriety tests that is rarely explained to subjects or juries: the tests use a negative scoring system. A subject receives no credit for doing things correctly. Clues are counted only when something goes wrong. This means a person who walks eight of nine steps perfectly and steps off the line once receives the same score as a person who stumbled on every step. The tests are not designed to demonstrate sobriety. They are designed to identify clues that support a finding of impairment.

Factors That Produce “Clues” Without Alcohol

Because the tests score only failure, and because that failure can come from many ordinary sources, a sober driver can produce the same “clues” an impaired driver would. Any officer who administers these tests 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.

Age

The NHTSA validation studies for the Walk and Turn and One Leg Stand were conducted primarily on subjects between the ages of 18 and 65. Balance degrades naturally with age due to reduced vestibular function, diminished muscle strength, and changes in proprioception. The NHTSA manual acknowledges that individuals over 65 may have difficulty with balance tests for reasons unrelated to alcohol. A driver in this age range who loses balance or uses arms for steadiness is exhibiting behavior that may be entirely typical of their age group.

Weight

The One Leg Stand has an explicit validation limitation: the test was not studied on individuals who are 50 or more pounds overweight. The NHTSA manual states this directly. An officer who administers the One Leg Stand to an overweight individual and scores balance difficulties as clues of impairment is applying a test to a subject outside its validated population.

Physical Injuries and Musculoskeletal Conditions

Injuries or conditions affecting the legs, knees, ankles, feet, or back directly impair performance on the Walk and Turn and One Leg Stand for reasons unrelated to intoxication. A driver with a recent knee injury, a chronic hip condition, or foot pain cannot maintain the required stance and movement pattern without discomfort that affects performance. Officers are required to ask whether the subject has any physical conditions that might affect their ability to perform the tests but this inquiry is frequently perfunctory or omitted entirely.

Inner Ear Disorders and Vestibular Conditions

The vestibular system, housed in the inner ear, is the primary mechanism of balance. Conditions that affect it (benign paroxysmal positional vertigo (BPPV), labyrinthitis, vestibular neuritis, and Meniere’s disease) cause balance disturbances and can independently produce nystagmus. A driver with an active inner ear condition may fail the Walk and Turn, One Leg Stand, and HGN tests for reasons entirely unrelated to alcohol. Because these conditions are not always visible or disclosed during a brief roadside encounter, they are frequently missed in the officer’s assessment.

Fatigue

Significant fatigue produces physical and cognitive effects that closely resemble alcohol impairment: reduced coordination, impaired divided attention, difficulty following multi-step instructions, and slowed reaction time. A driver coming off a double shift, returning from a long drive, or running on minimal sleep may perform poorly on all three standardized tests for reasons that have nothing to do with alcohol consumption. The divided attention tasks that the SFSTs are designed to reveal as impaired are also the tasks most affected by sleep deprivation.

Footwear

The NHTSA manual specifies that a subject wearing heels more than two inches high should be given the opportunity to remove their shoes before performing balance tests. Officers who fail to make this offer (or who administer the tests to a driver in boots, wedges, or elevated footwear without documentation of why the offer was declined) have deviated from the protocol. Balance difficulties attributable to footwear cannot be scored as evidence of intoxication.

Nervousness and Anxiety

A roadside stop at night, with flashing lights and a uniformed officer, is a stressful situation for most sober drivers. Anxiety produces physical effects (trembling, difficulty concentrating, altered gait, and performance anxiety on coordinated tasks) that are indistinguishable from mild intoxication in SFST scoring. An officer who attributes every physical manifestation of nervousness to alcohol without investigating whether the subject was already anxious before any alcohol-related observations is conflating cause and effect.

HGN: Nystagmus Has More Than 47 Causes Other Than Alcohol

Horizontal Gaze Nystagmus is presented to juries as the most reliable of the three standardized tests because it is involuntary and not subject to conscious control. What is not always explained is that HGN has more than 47 documented causes that have nothing to do with alcohol.

Non-alcohol causes of HGN include: inner ear disorders (BPPV, labyrinthitis, Meniere’s disease, vestibular neuritis); neurological conditions including multiple sclerosis, brain stem lesions, cerebellar dysfunction, and traumatic brain injury; naturally occurring pathological nystagmus that is present at baseline in a percentage of the population; certain prescription medications including anticonvulsants, sedatives, and some antihistamines; fatigue; and gaze-evoked nystagmus that is present at extreme lateral gaze in many people regardless of impairment.

An officer who observes HGN clues but does not know (and does not ask) whether the driver has any inner ear condition, neurological history, or relevant medication use cannot reliably attribute the observed nystagmus to alcohol. Cross-examination on the officer’s knowledge of alternative HGN causes and whether they made any inquiry before attributing the nystagmus to intoxication is a standard component of an effective SFST challenge.

Environmental Conditions That Affect Reliability

Testing Surface

Walk and Turn and One Leg Stand require a dry, hard, level, non-slippery surface. Gravel shoulders, cracked or uneven pavement, wet surfaces, sloped roadways, and grass do not meet this standard. Maintaining the required positions on an unstable or uneven surface is physically more difficult than on the ideal surface, and difficulty that arises from the surface, not from the subject, should not be scored as a clue of impairment.

Lighting Conditions

HGN observation requires that the officer clearly see the subject’s eyes tracking the stimulus throughout the test. Poor ambient lighting, flashing emergency lights in the peripheral field, and glare from oncoming headlights all compromise observation quality. The NHTSA protocol requires adequate lighting, and tests conducted under poor lighting conditions produce observations of reduced reliability.

Weather and Temperature

Cold temperatures cause muscle stiffness and reduce fine motor control. High winds create physical challenges for maintaining static balance positions. Rain and adverse weather affect both the testing surface and the subject’s physical comfort and concentration. None of these conditions were present in the controlled environments used for the NHTSA validation studies, and the accuracy figures from those studies do not apply when these conditions are present.

Traffic and Distractions

Performing divided attention tasks alongside active traffic (with vehicles passing at speed, horns, emergency vehicle sounds, and the physical stress of standing near a travel lane) imposes external cognitive demands not present in the validation environment. Distraction from traffic is an independent factor that can affect performance on multi-step balance tests.

Protocol Deviations: When Officers Administer the Tests Incorrectly

SFST accuracy depends entirely on protocol compliance. The NHTSA manual specifies the exact sequence of instructions, the demonstration requirements, the stimulus movement for HGN, the clue observation criteria, and the scoring standards. Deviations from any element of this protocol undermine the scientific foundation of the results. Common protocol deviations the defense examines in every SFST case include:

• HGN stimulus movement speed. The stimulus must move at a specified pace across the visual field. Too fast or too slow affects whether maximum deviation nystagmus can be accurately observed.
• HGN holding time at maximum deviation. The NHTSA protocol requires holding the stimulus at maximum lateral deviation for a minimum of four seconds per eye to observe sustained nystagmus. Officers who do not hold the stimulus long enough cannot accurately score this clue.
• Walk and Turn instruction phase. The subject must maintain the heel-to-toe stance throughout the entire instruction phase, before the walking begins. Officers who begin scoring before the walking phase starts (or who do not observe the instruction phase stance) are not following the protocol.
• Number of steps. The Walk and Turn requires nine steps in each direction. An officer who instructs a different number, or who does not count steps accurately, is not administering the test as validated.
• One Leg Stand counting method. The subject must count out loud by thousands while maintaining the raised-foot position. Officers who do not specify the counting method or who deviate from this requirement are not following the validated protocol.
• Failure to ask about physical conditions. The NHTSA manual requires the officer to ask whether the subject has any physical conditions that might affect their ability to perform the tests before administering them. An officer who skips this inquiry has failed a required step.

Can You Refuse Field Sobriety Tests in Texas?

Yes. Field sobriety tests are not mandatory under Texas law. Unlike the chemical test (breath or blood) which carries implied consent consequences for refusal, there is no statutory penalty for declining to perform the standardized field sobriety tests. An officer can still arrest you based on other observations and can still obtain a warrant for a blood draw.

Whether to perform the tests is a decision with real consequences in either direction. A person who performs the tests and provides clues gives the prosecution additional evidence. A person who declines forces the prosecution to rely on the officer’s observations alone, without the test results. The right answer depends on the specific facts of the stop. What is always true is that the tests are harder to do correctly than they appear, and that officers are trained to identify clues that support arrest. If you are facing DWI charges and have questions about how your FST performance is being characterized, contact Deandra Grant Law for an evaluation.

How We Use This in Your Defense

In every case involving SFST evidence, Deandra Grant Law obtains the officer’s dashcam and bodycam video, the officer’s training records and SFST certification documentation, and any available information about the subject’s relevant physical and medical history. The video is reviewed against the NHTSA protocol to identify specific deviations in instruction, demonstration, stimulus movement, holding time, clue observation, and scoring.

This review carries more weight because of who performs it. Managing Partner Deandra Grant is a trained SFST instructor — she has completed 24 hours of NHTSA standardized field sobriety test training and the NHTSA Field Sobriety Testing Instructor Course, and she administers and grades the SFST component of the ACS-CHAL Forensic Lawyer-Scientist examination and the DUIDLA Board Certification examination. Partner Douglas E. Huff holds the same ACS-CHAL Forensic Lawyer-Scientist designation. When an officer testifies that a subject ‘stepped off the line’ or ‘failed to maintain balance,’ the cross-examination is not about the general reliability of field sobriety tests. It is about whether that specific officer, on that specific night, administered that specific test the way the NHTSA protocol requires — and whether what the officer observed is attributable to alcohol or to the conditions, the subject’s physical characteristics, or the officer’s own administration errors. Protocol deviations that would be invisible to an attorney without that training become the basis for specific, grounded cross-examination a jury can understand and credit.

If your DWI case involves field sobriety test results, call (214) 225-7117 for a free consultation, or schedule online at texasdwisite.com.